Abuse-Deterrent Formulations of Opioids: Effectiveness and Value Final Evidence Report August 8, 2017 Prepared for: ©Institute for Clinical and Economic Review, 2017 ICER Staff/Consultants Reiner Banken, MD, MSc Senior Fellow Institute for Clinical and Economic Review Foluso Agboola, MBBS, MPH Research Scientist Institute for Clinical and Economic Review Patricia Synnott, MALD, MS Senior Research Associate Institute for Clinical and Economic Review Margaret Webb, BA Research Assistant Institute for Clinical and Economic Review Varun Kumar, MBBS, MPH, MSc Health Economist Institute for Clinical and Economic Review Rick Chapman, PhD, MS Director of Health Economics Institute for Clinical and Economic Review Celia Segel, MPP Program Manager, New England CEPAC Institute for Clinical and Economic Review Daniel A. Ollendorf, PhD Chief Scientific Officer Institute for Clinical and Economic Review Steven D. Pearson, MD, MSc President Institute for Clinical and Economic Review DATE OF PUBLICATION: August 8, 2017 We would also like to thank Anne Loos, Noah Mwandha, and Erin Lawler for their contributions to this report. ©Institute for Clinical and Economic Review, 2017 Page i Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value About ICER The Institute for Clinical and Economic Review (ICER) is an independent non-profit research organization that evaluates medical evidence and convenes public deliberative bodies to help stakeholders interpret and apply evidence to improve patient outcomes and control costs. ICER receives funding from government grants, non-profit foundations, health plans, provider groups, and health industry manufacturers. For a complete list of funders, visit http://www.icer- review.org/about/support/. Through all its work, ICER seeks to help create a future in which collaborative efforts to move evidence into action provide the foundation for a more effective, efficient, and just health care system. More information about ICER is available at http://www.icer- review.org About New England CEPAC The New England Comparative Effectiveness Public Advisory Council (New England CEPAC) – a core program of ICER – provides a public venue in which the evidence on the effectiveness and value of health care services can be discussed with the input of all stakeholders. New England CEPAC seeks to help patients, clinicians, insurers, and policymakers interpret and use evidence to improve the quality and value of health care. The New England CEPAC is an independent committee of medical evidence experts from across New England, with a mix of practicing clinicians, methodologists, and leaders in patient engagement and advocacy. All Council members meet strict conflict of interest guidelines and are convened to discuss the evidence summarized in ICER reports and vote on the comparative clinical effectiveness and value of medical interventions. More information about New England CEPAC is available at http://icer-review.org/programs/new-england-cepac/. ©Institute for Clinical and Economic Review, 2017 Page ii Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Expert Review and Acknowledgements Data described in the State-Specific Analysis were developed by the Commonwealth of Massachusetts Health Policy Commission (HPC). In the development of this report, ICER’s researchers consulted with clinical experts, patients, manufacturers and other stakeholders. The following experts provided input and data that helped guide the ICER team as we shaped our scope and draft report. None of these individuals is responsible for the final contents of this report or should be assumed to support any part of this report, which is solely the work of the ICER team and its affiliated researchers. Conflict of Interest disclosures are included in Appendix H of the report. For a complete list of stakeholders from whom we requested input, please visit: https://icer- review.org/material/adf-stakeholder-list/ Clinical Reviewers Lewis S. Nelson, MD Professor and Chair Department of Emergency Medicine Director, Division of Medical Toxicology Rutgers New Jersey Medical School Chief of Service, Emergency Department University Hospital of Newark New Jersey Poison Information & Education System Richard C. Dart, MD, PhD Director, Rocky Mountain Poison and Drug Center Denver Health and Hospital Authority Professor, University of Colorado School of Medicine Economic Reviewer Alan G. White Managing Principal Analysis Group, Inc. Patient/Advocacy Reviewer Paul Gileno President US Pain Foundation ©Institute for Clinical and Economic Review, 2017 Page iii Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Table of Contents Executive Summary ............................................................................................................................. ES1 1. Background ......................................................................................................................................... 1 2. The Topic in Context ............................................................................................................................ 6 3. Summary of Coverage Policies ........................................................................................................... 21 4. Comparative Clinical Effectiveness ..................................................................................................... 24 4.1 Overview ........................................................................................................................................... 24 4.2 Methods ............................................................................................................................................ 25 4.3 Results ............................................................................................................................................... 27 5. Other Benefits or Disadvantages........................................................................................................ 47 6. Cost-Benefit and Potential Budget Impact of Abuse-Deterrent Opioid Formulations .......................... 49 6.1 Overview ........................................................................................................................................... 49 6.2 Cost-Benefit Model ........................................................................................................................... 50 6.3 Prior Published Evidence, Model Validation..................................................................................... 64 6.4 State-Specific Policy Analysis for Massachusetts.............................................................................. 66 6.5 Summary and Comment ................................................................................................................... 68 7. Summary of the Votes and Considerations for Policy ......................................................................... 71 7.1 About the New England CEPAC Process ........................................................................................... 71 7.2 Clinical Effectiveness Voting Results ................................................................................................. 72 7.3 Policy Voting Results ......................................................................................................................... 74 7.4 Roundtable Discussion and Key Policy Implications ......................................................................... 77 References ............................................................................................................................................ 84 Appendix A. Search Strategies and Results............................................................................................. 98 Appendix B. Public and Representative Private Insurer Coverage Policies ............................................ 103 Appendix C. Previous Systematic Reviews and Technology Assessments ............................................. 111 Appendix D. Ongoing Studies............................................................................................................... 112 Appendix E. Comparative Clinical Effectiveness Supplemental Information.......................................... 113 Appendix F. Evidence Summary Tables ................................................................................................ 115 Appendix G. Cost-Benefit and Budget Impact Supplemental Information............................................. 162 Appendix H. Conflict of Interest Disclosures for Expert Reviewers ....................................................... 167 Appendix I. Oral Public Comments ....................................................................................................... 168 ©Institute for Clinical and Economic Review, 2017 Page iv Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value List of Acronyms Used in this Report ADF Abuse-deterrent Opioid Formulation with an FDA label AHRQ Agency for Healthcare Research and Quality CI Confidence interval CDC Centers for Disease Control and Prevention CMS Centers for Medicare and Medicaid Services DEA Drug Enforcement Administration DSM Diagnostic and Statistical Manual ER Extended release ER/LA Extended release/Long acting FDA Food and Drug Administration HCV Hepatitis C virus HIV Human immunodeficiency virus ICD International Classification of Diseases IR Immediate release LA Long acting MAT Medication Assisted Treatment MED Morphine equivalent dose NAVIPPRO National Addictions Vigilance Intervention and Prevention Program NIH National Institutes of Health NPDS The National Poison Data System ND No data NR Not reported NSDUH National Surveys on Drug Use and Health OC OxyContin ORF Reformulated OxyContin PDMP Prescription Drug Monitoring Program PRISMA Preferred Reporting Items for Systematic Reviews and Meta-Analyses RADARS Researched Abuse, Diversion, and Addiction-Related Surveillance System REMS Risk Evaluation and Mitigation Strategy RCT Randomized controlled trial SE Single entity SKIP Survey of Key Informants’ Patients Program SUD Substance use disorder TR Tamper-Resistant VAS Visual Analog Scale ©Institute for Clinical and Economic Review, 2017 Page v Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Executive Summary Background Opioids are used to treat acute and chronic pain that arises from a variety of causes, ranging from trauma to advanced illness. Every year, 100 million people in the U.S. suffer from pain, with 9-12% of these individuals experiencing pain that is considered chronic—lasting longer than three months.1 Opioid therapy is an essential component of chronic pain management for many patients, but the addictive and euphoric properties of these drugs make patients vulnerable to misuse, abuse, addiction, and possible death by overdose. Since 1999, the number of deaths from prescription opioids in the U.S. has increased nearly fourfold, rising in parallel with the volume of dispensed prescriptions.2 In addition to the societal impact of opioid-related deaths, the level of abuse and misuse of these agents also has significant consequences for health care utilization. For every one death from prescription opioids, it is estimated that there are 10 treatment admissions for abuse, 32 emergency room visits for misuse or abuse, 130 people who become dependent on opioids, and 825 people who report non-medical use of these drugs.3 A variety of measures have been implemented to attempt to mitigate opioid abuse, one of which is the introduction of abuse-deterrent formulations (ADFs) of these drugs, an increasing number of which have reached the market during the last few years.4 As described further below, abuse deterrence is based on different advances in technology, including physical and chemical barriers, agonist/antagonist combinations, aversive agents, and prodrugs. However, the abuse-deterrent technology does not change the addictive properties of the opioid itself, and while ADFs deter abuse, they are not abuse-proof.5 In online forums for abusing opioids, there are many instructions on how to circumvent certain abuse-deterrent technologies.6 This report focuses on the effectiveness, safety, and economic impact of ADFs relative to non-ADF opioid treatment, and considers the evidence and potential cost-benefit of different strategies to replace non-ADF formulations with ADFs in specific populations. Topic in Context In 2010, the FDA approved Purdue Pharma’s reformulated OxyContin® (extended release oxycodone) with a harder-to-crush exterior to reduce the potential for abuse by snorting or dissolving in order to inject.7 The reformulated opioid was approved as the first abuse-deterrent formulation, and now captures over 90% of the ADF market.8 ©Institute for Clinical and Economic Review, 2017 Page ES1 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents ADFs of extended-release (ER) opioids aim to prevent what is known as “dose-dumping,” the rapid release through tampering of drug contained in an ER dosage form, which is at much higher concentrations than that found in IR opioids. For example, taking a 12 mg Palladone capsule, an ER hydromorphone no longer on the market, together with 8 ounces of 40% ethanol increased the average blood hydromorphone concentration by 6 times compared with co-ingestion with water.9 Increasing the blood opioid concentration can also be achieved by altering the route of administration. The oral bioavailability of hydromorphone is about 10%, which means that an intravenous injection of the same substance increases the bioavailability tenfold in addition to the more rapid onset and rise through the dose-dumping route.10,11 In April 2015, the FDA issued non-binding recommendations encouraging manufacturers to produce abuse-deterrent formulations (ADFs) of opioids. Between 2016-2017, the FDA approved five new ADFs; today, nine ER opioids and one immediate release (IR) opioid have FDA approved labeling describing a variety of abuse-deterrent properties (Table ES1). Only ER ADF formulations were available on the market as of June 2017. ER opioids represent about 10% of all opioids prescribed.12,13 Nine ADF products are in the late-stage pipeline (Stage III or FDA submission).14 ADFs are relatively new, branded therapies for treating pain, and are generally more expensive than both their non-ADF branded equivalents and generic versions. The only generic ADFs on the market are “authorized” generics of OxyContin.15,16 The pills are identical to the original OxyContin, and the price is higher than that of extended release generics approved through the traditional process.17 This absence of true generic competition contributes to the higher prices of currently available ADFs. For example, in 2016, the VA spent an amount of approximately $100 million overall on opioids. If all opioids were to be replaced with ADFs and the costs would be increased 10-fold on the average, this “would result in approximately $1 billion yearly for these products and could represent as much as 20 percent of the VA pharmacy budget”.18 Policymakers are challenged on how to structure conversion to ADFs in a responsible and economically feasible manner.19 ©Institute for Clinical and Economic Review, 2017 Page ES2 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table ES1. Opioid Products with FDA-Approved Abuse-Deterrent Labeling Brand Name Type of Year of Reported Abuse-Deterrence Mechanism Commercially Opioid Approval Available± OxyContin® Oxycodone 2010 When dissolved, forms a viscous gel that is difficult to inject through a hypodermic Yes (reformulated) needle. Embeda® Morphine 2014 Capsules of ER morphine pellets that contain a sequestered core of naltrexone; if the Yes pellets are swallowed, the morphine is gradually released and absorbed, while the naltrexone core passes through the gut intact. If the pellets are crushed, chewed, or dissolved, the naltrexone is released, blocking morphine-induced euphoria. Targiniq® ER Oxycodone 2014 Combination pill containing extended-release (ER) oxycodone and naloxone; if the No formulation is crushed and administered intravenously or intranasally, high naloxone concentrations block opiate-induced euphoria and can induce withdrawal symptoms. Hysingla® ER Hydrocodone 2015 When dissolved, forms a viscous gel that is difficult to inject through a hypodermic Yes needle MorphaBond® Morphine 2015 Formulated with inactive ingredients that make the tablet harder to adulterate while No maintaining ER characteristics if the tablet is subjected to physical manipulation or chemical extraction. Xtampza® ER Oxycodone 2016 Capsules containing microspheres formulated with oxycodone base and inactive Yes ingredients that make the formulation harder to manipulate. Troxyca® ER Oxycodone 2016 Contains pellets that consist of oxycodone that surround sequestered naltrexone. When No taken orally, the naltrexone is intended to remain sequestered and patients receive ER oxycodone. When the pellets are crushed, the naltrexone is released and counteracts the effects of oxycodone. Arymo® ER Morphine 2017 A polymer matrix tablet technology with controlled-release properties as well as physical Yes and chemical barriers that resist manipulation. The technology results in a viscous hydrogel on contact with liquid, making the product very difficult to draw into a syringe. Vantrela® ER Hydrocodone 2017 Incorporates abuse-deterrent technology designed to resist drug extraction through the No most common routes: oral, intranasal, and intravenous. **RoxyBond® Oxycodone 2017 Includes inactive ingredients that make the tablets harder to misuse by physical No manipulation, chemical extraction, or both; in vitro data suggest physicochemical properties that are expected to make abuse through injection difficult. 20 *Modified from Becker, 2017. **Only ADF approved as immediate-release. ±As of June 28, 2017. ©Institute for Clinical and Economic Review, 2017 Page ES3 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents FDA Designation for Abuse-Deterrent Formulations of Opioids In this report, the term ADF is only used for drugs with abuse-deterrent properties as recognized by the FDA. In order to be defined as “abuse-deterrent” in the drug label, there must be sufficient evidence of abuse deterrence according to FDA standards, which are based on premarket studies and mandatory real world studies after drug approval.5 The clinical component of premarket studies, also known as category 3 studies, does not involve pain patients, but healthy, non- dependent recreational drug users between the ages of 18 and 55 years. These studies provide important information on the possible impact of ADFs, but have not been validated regarding their ability to predict real-world abuse.21 Postmarket studies (i.e., following regulatory approval) are also required by the FDA and are designed to measure the real-world impact of ADFs on patterns of abuse and misuse. However, studies of prescription drug abuse differ from traditional pharmacoepidemiologic investigations, as exposure occurs mostly outside the health system in individuals who did not receive prescriptions for these drugs (i.e., “diversion”), and information is not available in clinical information systems used for other drugs.22 The FDA aims to improve postmarket studies by convening a multi-day meeting in July 2017 to understand how to better leverage existing data sources, identify potential new data sources, and highlight new methods and study designs.23 ADFs and their non-ADF counterparts are considered bioequivalent, producing the same analgesic benefits, and have the same profile of adverse effects when used as prescribed.24 However, when abused, the ADFs may present particular safety issues, such as precipitated severe withdrawal symptoms, infections through needle sharing,71,72 thrombotic microangiopathy,25 or other risks26 caused by intravenous exposure of substances produced by the tampering of the excipients used in ADF technology. ADFs may deter chewing, intranasal, and intravenous routes of abuse.27 However, swallowing pills whole is the most common form of abuse and is not deterred by ADFs.28 Opioid Abuse, Diversion, and Shifts in Opioid Use The progression from medical use to non-medical use, to abuse and addiction, has not been very well studied. It is generally believed that chewing an ER opioid is an important step towards addiction, followed by intranasal and intravenous routes of abuse.27 However, even among patients entering drug rehabilitation programs, oral abuse of the IR formulation or the manipulated ER formulation remains the major route, with the exception of morphine, abused through the intravenous route in 66% of patients entering drug rehabilitation.29 Understanding the characteristics and pathways of individuals at higher risk of abuse is quite challenging. For example, we do have some information on the routes of abuse for patients entering drug rehabilitation programs, but our understanding is limited concerning the progression from misuse to abuse, ©Institute for Clinical and Economic Review, 2017 Page ES4 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents including recreational abuse, and finally to addiction. Furthermore, these pathways to abuse and addiction probably differ among different age groups. While greater understanding of abuse risk is an area of active research, there are currently no validated tools for predicting increased risk for abuse30,31 and abuse pathways. Many individuals who abuse opioids do not receive a prescription from a prescriber. According to national surveys, about 50% of people who misused prescription opioids got them from a friend or relative for free, while 22% got them from a doctor and only 4% bought them from a drug dealer.32,33 This is known as diversion, or the transfer of opioid analgesics from a lawful to an unlawful channel for distribution or use.34 The volume of prescription opioids diverted annually for non-medical use is extremely difficult to estimate. However, street prices of specific opioids can be a good indicator of drug availability, demand, and abuse potential.35 It is important to understand how the introduction of ADFs impact diversion and the availability of illicit opioids in order to capture their true impact on overall abuse, including for abusers who obtain opioids through diversion and not through a prescription. We summarize the available evidence on diversion in the Comparative Clinical Effectiveness section in the main report. As ADFs enter the market, it is also critical to understand trends in abuse, since persons already abusing specific opioids may shift to other opioids or routes of administration if a specific opioid is replaced with an abuse-deterrent formulation. Since 2011, the continuing rise in opioid deaths are no longer attributable solely to prescription opioids but also to illicit opioids, mainly heroin and illegally manufactured fentanyl.2 In Section 4, we summarize the available literature that correlate the introduction of ADFs with alternative opioid abuse patterns or methods of administration. Understanding trends in abuse, diversion, and potential shifts in drug use are key pieces of evidence in understanding the impact of ADFs on the overall opioid epidemic. They are summarized in the Comparative Clinical Effectiveness section in the main report. Policy Interventions: Clinical Guidelines and State Policies The context for understanding the potential benefits of ADFs is complex, as these technologies are often part of a multipronged strategy to combat the public health epidemic of prescription opioid deaths. This strategy often includes educating clinicians to reduce initiation of opioid use, shortening the duration of prescriptions, monitoring of prescriptions, and in some states, mandatory substitution of opioid prescriptions with ADFs. In 2016, the Centers for Disease Control and Prevention (CDC) released the CDC Guideline for Prescribing Opioids for Chronic Pain for patients 18 and older in primary care settings. This new guideline constitutes the most recent professional reference for treatment decisions for chronic ©Institute for Clinical and Economic Review, 2017 Page ES5 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents pain (outside of active cancer treatment, palliative care, and end-of-life care). The primary recommendation prioritizes nonpharmacologic and non-opioid therapy for chronic pain.36 The CDC guidelines recommend a universal approach of urine testing to be performed at least annually for all patients receiving an opioid for chronic pain;36 and judged that the evidence on clinical tools for identifying patients at higher risk for abuse was insufficient or absent.30 None of the 12 recommendations of the CDC guideline meets a high standard of evidence, but they are judged to reduce harm and likely improve chronic pain control in the U.S.37 The guidelines do not currently mention ADFs for treating patients with pain. State governments have also stepped up efforts to address the opioid epidemic, with executive led taskforces, physician education, and legislation to establish prescription monitoring programs, restrict the duration and/or quantity available in an opioid prescription, and allocate more funding for abuse treatment options. In August 2014, Massachusetts became the first state to pass legislation requiring pharmacies to automatically substitute ADFs for chemically equivalent non-ADF opioid prescriptions, and requiring insurance carriers to cover these ADFs with no additional cost burden to patients. Maryland, Florida, and West Virginia passed similar legislation, and bills have been introduced in more than 20 other states relating to ADF coverage. At this point, however, data on the impact of these state policy and systems-level interventions are limited and inconsistent.38 Since ADFs are substantially more expensive than their non-ADF equivalents, policymakers throughout the country are wrestling with how best to spend their resources to address the opioid epidemic.19 The present report will examine the specific value of using ADFs as a strategy to influence abuse of prescription opioids and the epidemic of deaths from prescription opioids. As outlined in the analytic framework in the full report, the systematic review covers all impacts of ADFs, but does not compare the value of ADFs to other strategies to curb the opioid epidemic. Insights Gained from Discussions with Patients and Patient Groups As part of our review, we spoke with patient organizations focused on chronic pain and addiction. Patient organizations focused on chronic pain stressed the need for continued, affordable patient access to opioid therapy for daily function while also recognizing the need to curb opioid misuse and addiction. It was felt that the different policy initiatives for reducing the overall use of opioids contributed to increasing difficulties in obtaining prescriptions for long term opioid therapy. Patients with chronic pain were nervous that higher co-payments for ADFs compared to non-ADF ER opioids could act as a potential barrier to accessing needed opioid therapy. Some patients with chronic pain saw the ADF designation as potentially smoothing access to necessary medication, as it might reduce the typical level of stigma associated with controlled substances. The importance of ©Institute for Clinical and Economic Review, 2017 Page ES6 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents assessing the total clinical, economic, and social value of ADFs was widely recognized by the different stakeholders as an essential step for their rational use. Overall, organizations representing patients with chronic pain reported patients’ difficulties accessing specialized multidisciplinary pain care. Some patients believe that access to integrated pain management, including medications and complementary approaches such as acupuncture, physical therapy, and mind–body practices would contribute to diminishing the need for prescription opioids. Patient advocates who worked with patients struggling with addiction helped to illustrate how patients progress to opioid addiction, often beginning with the recreational oral abuse of opioids. One advocate who worked with teenagers described how her young patients abused pills orally and recreationally before getting addicted and entering her treatment program. She also described the stigma for young users in injecting opioids intravenously. These patient advocates saw potential in ADFs to prevent the progression of abuse from oral use to snorting and injecting opioids. However, they also cautioned that individuals who are unable to abuse a particular opioid may substitute an easier-to-abuse option. Comparative Clinical Effectiveness To evaluate the clinical effectiveness of the ten abuse-deterrent formulations (ADFs) with FDA labels, we abstracted evidence from available clinical and observational studies, whether in published, unpublished, or abstract form. The primary comparators examined included non-abuse- deterrent formulations of specific opioids as appropriate. Studies on opioids with abuse-deterrent properties but without an FDA label recognizing these properties were not included in the assessment. We sought evidence on the effects of ADFs on abuse potential endpoints (e.g., VAS measures of drug liking, take drug again), as well as real world outcomes (e.g., abuse and misuse, addiction, overdose, drug diversion). We did not include studies that focused exclusively on the analgesic properties of ADFs without reporting on any abuse-related endpoints. In total, we included 41 references, of which 15 were premarket RCTS that evaluated abuse potential endpoints, and 26 were postmarket observational studies that primarily evaluated the real-world impact of ADFs on levels of abuse and misuse. Data on all outcomes were summarized in evidence tables (Appendix F) and analyzed in descriptive fashion only. ©Institute for Clinical and Economic Review, 2017 Page ES7 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Results Studies that Evaluated Abuse Potential Only We reviewed 15 premarket studies that evaluated the abuse potential of ADFs. These studies were randomized, double-blind, active- and placebo-controlled crossover trials of healthy, non- dependent recreational drug users between the ages of 18 and 55 years.a The trials were broadly divided into two categories: those that assessed oral abuse potential and those that assessed intranasal abuse potential (see Tables ES2 and ES3). Key measures of abuse potential included maximum levels of “drug liking” (“at this moment, my liking for this drug is…”), which was a primary endpoint in the studies of focus, as well as secondary endpoints of “overall drug liking” (typically measured at 12 and 24 hours post-dose), and “take drug again” (“I would take this drug again” measured at 12 and 24 hours post-dose). Drug liking endpoints were measured using a bipolar 0 to 100mm Visual Analog Scale (VAS), in which 0 represents “strong disliking”, 50 represents a neutral response, and 100 represents “strong liking”. Response to whether the subject would take the study drug again was measured on a unipolar scale of 0 to 100, where 0 represents “definitely would not take drug again” and 100 represents “definitely would take drug again.” Relative to non-ADF comparators, both crushed and intact forms of each extended-release ADF produced statistically-significantly lower scores for drug liking. Drug liking in oral abuse potential studies ranged from a 7-point difference between crushed Arymo ER and crushed morphine sulfate ER to a 25-point difference between Hysingla ER and hydrocodone IR solution.39,40 Similarly, the incremental difference in drug liking varied across intranasal abuse potential studies, ranging from seven points (crushed Vantrela ER vs. hydrocodone powder) to 36 points (crushed Targiniq ER vs. oxycodone IR powder).41,42 Crushed versions of each ADF generally produced higher drug liking scores than intact oral versions, but both remained lower than non-ADF comparators. Similar trends were observed for responses to questions regarding the likelihood of participants to take the drug again. Of note, there is no established threshold for what constitutes a clinically-important difference in any of these endpoints, so the clinical significance of these findings remains unclear even if statistical differences were noted. a One study of Targiniq ER was conducted among dependent opioid users ©Institute for Clinical and Economic Review, 2017 Page ES8 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table ES2. Premarket Studies Evaluating the Oral Abuse Potential of ADFs ADF Dose Intact & crushed ADFs & active VAS score, Emax (n) comparators¥ Drug Take drug Overall drug liking again liking Extended-release (ER) OxyContin -- No oral abuse potential study 43 Xtampza ER 40mg Xtampza ER- intact 68.8* 70.2* 69.4* (n=38) Xtampza ER- crushed 73.4* 73.7* 74.2* IR oxycodone- crushed 81.8 75.4 76.2 Troxyca ER44 60mg Troxyca ER- intact 59.3* 48.7* 53.3* (n=41) Troxyca ER- crushed 74.5* 72.5 74.3 IR oxycodone- crushed 89.8 81.5 81.8 ǂ45 Targiniq ER Targiniq ER-intact 54.7 38.5 NR (n=29) -- Targiniq ER-chewed 54.6 32.6 NR Oxycodone IR solution 77.9 61.4 NR Hysingla ER39 60mg Hysingla ER- intact 63.3† 32.6† 54.9† (n=35) Hysingla ER- crushed 69† 43† 56.8† Hydrocodone IR solution 94 86.7 84.1 46 † † Vantrela ER 45mg Vantrela ER- intact 53.9 46.4 49.2† (n=41) Vantrela ER- crushed 66.9† 58.7† 59† Hydrocodone IR 85.2 75.2 75 Embeda47 120mg Embeda- crushed 65.2† 57.7† 58.6† (n=33) Morphine sulfate ER- crushed 80.8 70.7 69.8 48 † Embeda 120mg Embeda- intact 67.6 NR NR (n=32) Embeda- crushed 68.1 † NR NR Morphine solution 89.5 NR NR Morphabond ER -- No oral abuse potential study 40 Arymo ER 60mg Arymo ER- intact 62† 56† 57† (n=38) Arymo ER- crushed 67* 61.5* 63.5 Morphine sulfate ER- crushed 74 68 67.5 Immediate-release (IR) RoxyBond IR -- No oral abuse potential study ¥: Placebo arms not included in table, non-ADF comparator arms indicated by bold font; *p≤0.05 vs. active comparator; †p≤0.001 vs. active comparator; ǂ study conducted in opioid-dependent population ©Institute for Clinical and Economic Review, 2017 Page ES9 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table ES3. Premarket Studies Evaluating the Intranasal Abuse Potential of ADFs ADF Dose Crushed ADFs & active VAS score, Emax (n) comparators¥ Drug liking Take drug Overall drug again liking Extended-release (ER) OxyContin49 30mg OxyContin- crushed NR 64* 69.7* (n=30) Original OxyContin- crushed NR 89.6 87.4 Oxycodone IR powder NR 86.6 84.8 50 † † Xtampza ER 40mg Xtampza ER- crushed NR 47.8 48.2† (n=39) Oxycodone IR- crushed NR 71.3 71.8 51 † * Troxyca ER 30mg Troxyca ER- crushed 60.5 58.9 60.2* (n=28) Oxycodone IR- crushed 92.8 88.4 85.4 ‡42 Targiniq ER 40mg Targiniq ER-Crushed 59.1 42.6 NR (n=23) Oxycodone IR powder 94.8 93.6 NR Hysingla ER52 60mg Hysingla ER- crushed 66.8† 34.6† NR (n=25) Hydrocodone powder 90.4 83.9 83.4 Vantrela ER41 45mg Vantrela ER- crushed 72.8* NR 68.5* (n=45) Hydrocodone powder 80.2 NR 77.1 Zohydro 83.2 NR 79.8 Embeda53 30mg Embeda- crushed 69.6† 60.6† 60.8† (n=33) Morphine sulfate ER- crushed 87.6 84.9 83.8 54 * * Morphabond ER 60mg Morphabond ER- crushed 71.1 NR NR† (n=25) Morphine sulfate ER- crushed 84.8 NR NR Arymo ER55 60mg Arymo ER- crushed 52.5† 50† 50.5† (n=46) Morphine sulfate ER- crushed 77.5 73 71 Immediate-release (IR) RoxyBond IR‡56 30mg RoxyBond IR - crushed 71.1† 62.2 NR (n=29) Oxycodone IR - crushed 82.9† 82.1 NR ¥: Placebo arms not included in table, non-ADF comparator arms indicated by bold font;‡: Data from Targiniq FDA label *p≤0.05 vs. active comparator; †p≤0.001 vs. active comparator Studies that Evaluated Real-World Evidence of Abuse and Misuse We identified 26 postmarket studies that evaluated real-world evidence on the impact of ADFs on abuse and misuse and health system related outcomes; all were non-randomized studies focusing exclusively on OxyContin and comparators. Comparators were either prescription opioids (e.g. IR oxycodone, ER morphine) or illicit drugs (e.g. heroin). There were no prospective studies conducted in inception cohorts of newly prescribed patients that measured real-world incidence of abuse ©Institute for Clinical and Economic Review, 2017 Page ES10 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents among ADF and non-ADF users. Instead, the current evidence of real-world impact is limited to time series that compared aggregate periods before and after the introduction of reformulated OxyContin. Data for these analyses were obtained from a variety of sources, as listed below: • Poison control calls or visits o National Poison Data System (NPDS) o The Researched, Abuse, Diversion, and Addiction (RADARS) Poison Center Program • Individuals entering substance use disorder programs o The National Addictions Vigilance Intervention and Prevention Program (NAVIPPRO) o RADARS Opioid Treatment Program (OTP) and the Survey of Key Informants’ Patients Program (SKIP) • Population-based surveys o National Survey on Drug Use and Health (NSDUH) • Electronic health data and medical claims databases o IMS LRx database o Truven MarketScan commercial database • Other data sources o RADARS Drug Diversion Program o RADARS StreetRx Program Major outcomes examined in these studies included overdose and fatalities, abuse rates, routes of abuse and drug diversion. None of the studies reported addiction as an outcome. Abuse We identified 17 studies that presented evidence on the impact of OxyContin on abuse in different populations. Most of the studies focused on the changes in the rates of abuse of OxyContin (and comparators), presented as the prevalent proportion of the study population that report or identified as abusing OxyContin and other comparator opioids during the specified time period. Examples of populations covered by these studies include patients entering substance use disorder programs (e.g., NAVIPPRO and RADARS SKIP studies), total U.S. population covered by a set of poison control centers (e.g., the RADARS poison center based studies), or commercially insured patients on OxyContin (e.g., the claims-based studies). Evidence on the impact of reformulated OxyContin on opioid abuse from these studies was mixed. The majority of studies found that after ©Institute for Clinical and Economic Review, 2017 Page ES11 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents the abuse-deterrent formulation of OxyContin was introduced, there was a decline in the rate of OxyContin abuse ranging from 12% to 75%, in different study populations and at different post- reformulation time points. However, the non-oral route of abuse declined at a significantly greater rate compared with the oral route of abuse,57,58 suggesting there may have been a shift from non- oral routes to the oral route of abuse. Many of the studies also found a contemporaneous increase in the rate of abuse of other prescription opioids (ER oxymorphone, ER morphine, IR oxycodone) and heroin during the same periods examined (Table ES4). ©Institute for Clinical and Economic Review, 2017 Page ES12 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table ES4. Changes in Abuse Patterns of OxyContin and Comparators Data source Timeframe compared Change in abuse pattern of OxyContin‡ % change of comparators Prior to Post- Outcome (population) % change HeroinPrescription opioids (excludes reformulation reformulation OxyContin) RADARS Poison center59 4Q08 - 3Q10 4Q10 - 1Q12 Quarterly rates of cases at poison control -38* NM All other opioids: NS centers (U.S. population) RADARS Poison center57 3Q09 - 2Q10 1Q11 - 2Q15 Quarterly rates of cases at poison control -75* NM All other opioids: -33* centers (U.S. population) RADARS Poison center60 3Q09 - 2Q10 1Q11- 4Q13 Quarterly rates of cases at poison control -55* NM All other opioids: -7* centers (U.S. population) RADARS SKIP61,62 1Q09 - 2Q10 1Q11 – 2Q14 Past month prevalence (Patients with -42* +100 ER oxymorphone: +38* primary diagnosis of opioid abuse) RADARS SKIP60 3Q09 - 2Q10 1Q11- 4Q13 Past month prevalence (Patients with -30* NM All other opioids: +16* primary diagnosis of opioid abuse) RADARS SKIP63 4Q09 – 3Q10 4Q10 – 1Q12 Past month prevalence (Patients with -37 +78¥ All other opioids: +5¥ primary diagnosis of opioid abuse) NAVIPPRO58 2Q09 – 3Q10 3Q10 – 2Q12 Past month prevalence (Patients entering -41* NM ER oxymorphone: +246* substance use disorder treatment) ER morphine: NS NAVIPPRO64 1Q08 – 3Q10 3Q10 – 4Q11 Past month prevalence (Patients entering -22* -11* ER oxymorphone: +191* substance use disorder treatment) ER morphine: NS NAVIPPRO60 3Q09 - 2Q10 1Q11- 4Q13 NC -48* NM All other opioids: -3* NSDUH65 1Q09 – 4Q09 1Q13 – 4Q13 Past year prevalence (US household survey- -28¥(NS) NM -- 12 years and older) NSDUH66 1Q09 – 4Q09 1Q13 – 4Q13 Past year initiation rate (US household -28¥† NM -- survey-12 years and older) NPDS67 3Q09 – 2Q10 3Q10 – 3Q12 Quarterly rates of calls to poison control -36* +42* Other single entity oxycodone centers (U.S. population) +20* Claims data68 3Q09 – 3Q10 4Q10 – 4Q13 Diagnosed rate (Patients on OxyContin and -35* NM ER oxymorphone: +236* comparator opioids) ER morphine: +44* IR oxycodone: +36* Kentucky cohort69 Pre-3Q10 4Q10 – 1Q11 Past month prevalence (recreational users) -55† NM IR oxycodone: +23 Canada cohort70 1 year prior 3Q12-4Q12 Positive urine drug screen (recreational -12* NM ER morphine: NS users) Australia cohort71 1Q14-1Q14 2Q14 – 3Q14 Past month prevalence (recreational users) -57* NM Other opioids: NS *p<0.01; † value not reported; ¥estimated; NM-not measured; NC-not clear; NS-Not significant; ‡There were some differences in the operational definition of abuse across sources (Table 10). ©Institute for Clinical and Economic Review, 2017 Page ES13 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents An additional study interviewed 153 recreational users with a history of long-term abuse of original OxyContin regarding the impact of introduction of ADF OxyContin on their choice of drug used for recreational purpose. Thirty-three percent of participants indicated that the reformulation had no effect on them and they continued to abuse OxyContin, another 33% indicated that they replaced OxyContin with other drugs as a result of the ADF, and only 3% indicated that the ADF influenced their decision to stop abusing drugs.61 The remaining 30% did not use OxyContin enough to change actions. Among those changing to other drugs (n=51), 70% indicated they switched to heroin, 29% to other prescription opioids, and one participant (2%) changed to cocaine.61 Overdoses and Fatalities Limited evidence indicates that rates of overdose and overdose deaths attributed to OxyContin declined after its abuse-deterrent formulation was introduced, with decreases ranging between 34% and 65%.60,72-74 During the same period, the rates of overdose deaths attributed to other prescription or illicit opioids increased or remained stable, suggesting that consumers may have switched to abusing other products.60,72,75 For example, an analysis from the Wharton School and RAND Corporation estimated that each percentage point reduction of OxyContin misuse after reformulation was associated with an increase of 3.1 heroin deaths per 100,000 population. 75 Drug Diversion & Prescription Opioid Utilization Evidence on drug diversion is extremely limited. We identified three publications that reported on diversion; two of the three analyses were conducted by the same author using different periods of follow-up. All three publications relied on population-adjusted longitudinal surveillance data from the RADARS Drug Diversion Program.57,59,60 In the Drug Diversion Program, law enforcement officers and regulatory agencies submit quarterly data on the number of new arrests, street buys and sales involving prescription products. Drug diversion is a measure of law enforcement activity and is limited by available resources within reporting jurisdictions, local law enforcement priorities, the drugs targeted by investigators, and variations in reporting over time.22,76 Population-adjusted rates of diversion declined over five years following the reformulation of OxyContin, reaching an 89% decrease by June 2015 (from 1.95 per 1,000,000 in the year prior to reformulation to 0.21 per 1,000,000 at year 5 following reformulation); diversion of other opioids also decreased during this period, albeit at a significantly lower rate (from 13.4 to 9.8 per 1,000,000).57 OxyContin prescription sales also declined during this period.72 Details can be found in Table 12 in the main body of the report. ©Institute for Clinical and Economic Review, 2017 Page ES14 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Controversies and Uncertainties The use of surrogate outcomes (measures of drug liking, take drug again, etc.) in the abuse potential premarket studies of an ADF constitutes an important source of uncertainty concerning the effectiveness of ADFs. There is considerable uncertainty around whether these surrogate endpoints are predictive of real-world abuse and whether the studies that evaluated them reflect how opioids are consumed in the real world. These studies used small, selected populations of non- opioid-dependent recreational drug users who received single, controlled doses of each product under investigation, which may not reflect real-world opioid use or misuse. Data from real-world evidence poses a different kind of challenge. We found no prospective studies conducted in inception cohorts that measured real-world incidence of abuse among ADF and non-ADF users. Instead, the current evidence of real-world impact is limited to time series, which are subject to potential confounding factors and other biases. For example, these analyses do not consider other interventions that may have taken place during the study period, such as expansion of prescription drug monitoring plans, implementation of Risk Evaluation and Mitigation Strategies (REMS), and provider education, among many others. In addition, time series may be subject to autocorrelation (i.e., statistical relation between pre- and post-values), which may lead to underestimation of standard errors and overestimation of intervention effects; or conversely, they may be subject to over-dispersion, defined as greater-than-expected variability in observed data based on the assumed distribution.77 Moreover, the time series we reviewed used different timeframes of analysis and different databases, often only with a short duration of follow-up. While the trends are relatively consistent, the estimates of magnitude vary and the results of the different studies cannot directly be compared. For prospective inception cohort studies, evidence on the use of clinical risk abuse stratification tools would be important to support clinical decision-making on whether ADFs should be used for any patient who gets an opioid or only those patients at a certain threshold of abuse risk. Results of a recent systematic review on this question came to the conclusion that the evidence on clinical tools for identifying patients that are at higher risk for abuse was insufficient or absent.30 Evidence on the progression from medical use to non-medical use as well as on the “natural history” of abuse and addiction is also needed. It is believed that chewing an ER opioid is an important step towards dependence and addiction, followed by intranasal and intravenous routes of abuse,27 which explains the use of certain physical or chemical barriers in the development of ADFs. However, none of the studies in the assessment included addiction as an outcome. Furthermore, the overall net benefit of introduction of ADFs into the system cannot be fully determined from the available evidence in these studies. Although limited evidence from most of the time series studies suggest a decrease in OxyContin-specific abuse and overdose following ©Institute for Clinical and Economic Review, 2017 Page ES15 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents reformulation, many of the studies also found a shift towards abuse of other prescription opioids and heroin, the extent of which may not be fully captured. There may be a tipping point at which more widespread access to ADFs would show system-wide benefits; however, current evidence from one survey suggest that only about 3% of a small cohort of long-term abusers of OxyContin stopped abusing drugs as a result of reformulation, while many others continued to abuse OxyContin or switched to other forms of opioids, including heroin.61 Uncertainty also remains on the association between the introduction of ADFs and increases in the rates of heroin use or deaths. Evidence from time series studies suggest a rise in the use of heroin following OxyContin reformulation.63,67 As discussed above, one study by RAND and Wharton that explored the relationship between state variation in OxyContin misuse and heroin death found that each percentage point reduction of OxyContin misuse after reformulation was associated with an increase of 3.1 heroin deaths per 100,000.75 However, other studies have shown that rates of heroin use and overdoses began increasing prior to the introduction of ADFs.78,79 Finally, we currently do not have any real-world evidence for the other ADFs, as their entry into the US market is very recent. While postmarket studies are mandatory with FDA approval, the first postmarket studies for ADFs other than OxyContin are not scheduled for completion until 2018 and 2019, for Hysingla® ER and Embeda®, respectively.80 Summary Using the ICER Evidence Matrix, we assigned evidence ratings for each of the ADFs of focus compared to non-ADF prescription opioids. ADFs and their non-ADF counterparts are bioequivalent, producing the same analgesic benefits, and have the same profile of adverse effects when used as prescribed.24 For patients being considered for an opioid for therapeutic purposes, we judge the comparative clinical effectiveness of OxyContin to be "C+" for the risk of abuse, primarily based on the surrogate outcomes of "likability" used in premarket studies, and the evidence on the changes in the rates of abuse reported in post-market studies. Even though we have reasonably high certainty that OxyContin does not provide inferior net health benefit compared to non-ADF comparators, without stronger real-world evidence that OxyContin reduces the risk of abuse and addiction among patients, our judgment is that the evidence can only demonstrate a "comparable or better" net health benefit (C+). For all other ADFs, excluding OxyContin, we judge the evidence to be “promising but inconclusive” (P/I) for use in individual patients being considered for an opioid. Similar to OxyContin, all other ADFs demonstrate potential comparability or better results than their non-ADF counterparts based on the surrogate outcomes of “likability” in premarket studies. Furthermore, they are considered bioequivalent in producing the same analgesic benefits, and have the same adverse effects when ©Institute for Clinical and Economic Review, 2017 Page ES16 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents used as prescribed. However, while many of these formulations may present advances in technology relative to OxyContin and include alternative physical or chemical barriers, agonists and antagonists, or aversive agents, there is no real-world evidence published on any of these other ADFs to demonstrate improved health outcomes or reductions in the risk of abuse. Considering the high dependence on “likability” studies, and the lack of real world evidence, our judgment is that we cannot determine the magnitude of abuse reduction at this time, leading to our P/I rating. We believe there can be even less certainty in a judgment on the comparative clinical effectiveness of ADFs versus non-ADF opioids if the question relates to the net health impact of introducing or substituting ADFs for non-ADFs to the broad population of individuals who use opioids for therapeutic and non-therapeutic purposes. The evidence on the impact of OxyContin reformulation shows a decrease in OxyContin-specific abuse, but also a shift in some cases toward other routes of administration, other prescription opioids, and heroin. Given the limited evidence base on this mix of positive and negative outcomes, we do not feel there is adequate evidence to discount the possibility that the balance would be net harmful overall across the entire population, especially early in the introduction of ADFs. We therefore judge there to be insufficient evidence (“I”) with which to judge the net health benefit, at the population level, of the introduction or substitution of ADFs for non-ADF opioids. Table ES5. ICER Rating on the Comparative Net Health Benefit of ADF versus Non-ADF Prescription Opioids Intervention Comparator ICER Rating Individual patient prescribed an opioid for therapeutic purposes OxyContin Non-ADF Extended Release Opioid C+ All other ADFs: Non-ADF Opioid P/I Embeda® Targiniq® ER Hysingla® ER MorphaBond® Xtampza® ER Troxyca® ER Arymo® ER Vantrela® ER RoxyBond® IR Overall population, including potential non-therapeutic users ADF Non-ADF Opioid I ©Institute for Clinical and Economic Review, 2017 Page ES17 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Other Benefits or Disadvantages In this section of our review, we seek to provide information on other benefits or disadvantages offered by the intervention to the individual patient, the delivery system, public health or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Patients report feeling stigmatized when prescribed opioids, given their widespread and publicized potential for abuse. Some patients report that having an ADF prescription would diminish this stigma, meaning they have a prescription that purportedly cannot be abused. For physicians, ADFs could, as part of a multi-pronged strategy, allow physicians to feel comfortable treating severe pain adequately without feeling forced to limit prescriptions as they might be otherwise. Discussions about the necessary controls on opioid prescribing need to also take into account the need for chronic pain patients to have reliable access to pain medication as part of a comprehensive pain management program.81 Due to the higher costs of ADF therapy, there may be new prior authorization requirements that require clinicians’ time and have an impact on productivity and patient care. In public comments received from hospice workers, they noted that increased costs and prior authorization requirements could impact productivity at small provider practices and hospice programs, as well as their ability to adequately care for patients in need of pain management. The hospice workers also noted that out-of-pocket costs due to higher costs of the therapies could inhibit access to opioids for patients in need. Legislation and policy mandating or encouraging use of ADFs often includes other components targeted at reducing opioid abuse and misuse. However, no evidence seems to have been generated to date on the effects of these multi-component strategies, or on the importance of ADF policy relative to other components. Safety issues have been raised with abuse-deterrent technologies after tampering for intravenous use for Opana®ER (oxymorphone)15,82 and for the ADF RoxyBond®.26 The reformulation of Opana ER in 2012 with a high-molecular-weight polyethylene oxide physical and chemical barrier led to a shift from intranasal to intravenous abuse.83 An HIV and Hepatitis C virus outbreak in Indiana was caused by using the tampered product with shared needles, and the outbreak was controlled by implementing a needle exchange program.84 In Tennessee, a cluster of thrombotic microangiopathy is thought to be related to intravenous exposure of substances produced by the tampering of the polyethylene oxide barrier used as abuse-deterrent technology in Opana ER.25,82 These risks could also arise with the intravenous abuse of other ADFs that also use a polyethylene ©Institute for Clinical and Economic Review, 2017 Page ES18 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents oxide barrier, such as the ADFs Arymo, Hysingla, and OxyContin. These risks could also arise with the intravenous abuse of other ADFs that use similar technologies. Finally, ADFs are currently available only for the extended-release opioid formulations that comprise around 10% of all prescription opioid use. Broader understanding of the benefits of ADF formulations are urgently needed, with the first immediate-release ADF approved by the FDA as of April 2017, but not yet available on the market. Cost-Benefit and Potential Budget Impact of Abuse-Deterrent Opioid Formulations The aim of our analysis was to estimate and compare the costs and benefits of using extended- release (ER) ADF opioids or non-ADF opioids for chronic pain. We developed a model to explore two key research questions: 1) what are the potential net costs and outcomes of using ADFs compared to non-ADF opioids, and 2) what levels of effectiveness in abuse reduction and in price difference would be needed for ADF opioids to achieve cost neutrality or net savings relative to non-ADF opioids? The benefits were defined in terms of the reduction in abuse-related outcomes, such as the number of incident cases of abuse and the number of opioid overdose-related deaths. Our primary analyses focused on hypothetical cohorts of chronic pain patients receiving ADF and non-ADF opioids respectively. We also conducted a state-specific policy analysis that analyzed the health and economic burden associated with opioid use in the state of Massachusetts if all non-ADF ER opioid prescription users in the state were to be converted to ADF ER opioid prescriptions. Due to the varied nature of the underlying cause for chronic pain, as well as the lack of published data on quality of life in opioid users, we employed a cost-benefit approach rather than a cost-utility (i.e., “cost per QALY”) structure. Additionally, primary analyses did not include the health outcomes or costs of externalities such as diversion or switching to heroin or other non-ADF opioids that may occur in reaction to the abuse-deterrent properties of ADF opioids, due to a lack of data attributing these patterns to ADF opioid use. The effects of diversion were explored in a scenario analysis, however. This de novo model, built from a health system perspective, consisted of cohorts of 100,000 non- cancer chronic pain patients new to ER prescription opioids. The population was 45% female, and mean age was 37 years.85 Separate cohorts were assumed for patients newly starting ADF and non- ADF opioids respectively. Figure ES1 represents the therapeutic use- and abuse-related pathways. All patients enter the model as therapeutic users, defined as those chronic pain patients who used prescription opioids for only pain-alleviating purposes and not for abuse. As a therapeutic user, a patient could discontinue opioid use due to end of treatment or die from non-abuse related causes. Annual probabilities of discontinuing therapeutic use were obtained from a claims analysis that followed patients with prescribed opioids over nearly five years, and reported on the proportion of ©Institute for Clinical and Economic Review, 2017 Page ES19 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents patients without an opioid prescription refill over a 6-month period.86 Patients entering the model in the first year, as well as those who continued as therapeutic users in subsequent years, had an annual probability of opioid abuse. This rate of diagnosed opioid abuse, obtained from a claims study that identified the rate of abuse from all non-ADF opioids as well as post-reformulation OxyContin (the first FDA-approved ADF opioid), was estimated for ADF and non-ADF opioids at 2.82% and 3.65% respectively.87 A proportion of those who abused had an assumed annual probability of ceasing to abuse opioids, at 10% after which they dropped out of the model. Other patients who abuse had an annual probability of death from opioid overdose, at 5.9 per 100,000 abuse patients, or other causes.79,88 The remainder of those who abuse continue to a subsequent year of abuse. The model employs annual cycles over a five-year time-horizon, taking a health care system perspective. We chose a five-year time horizon because we assumed that few patients would be prescribed opioids continuously for longer than five years. Figure ES1. Model Schematic Representing One Cycle for the Prescription Opioid Hypothetical Cohort Continue therapeutic use Abuse Therapeutic use Discontinue therapeutic use Death (all cause) Prescription opioids Continue abuse Cease to abuse Death from Abuse overdose Overdose Continue abuse Death (all cause) Patients in the ADF and non-ADF opioid cohorts follow the same pathway ©Institute for Clinical and Economic Review, 2017 Page ES20 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents The ADF and non-ADF opioids included in this model are listed in Appendix G, Table G1. We calculated a weighted average daily opioid drug cost for both ADF and non-ADF opioids, using the market share of drugs within the ADF and non-ADF classes and a 90mg Morphine Equivalent Dose (MED) daily dosage for each drug.8,89 We combined this market-share data with opioid costs as reported in the Federal Supply Schedule (FSS) to arrive at a weighted-average drug price of $11.60 per day for ADF opioids and $5.82 for non-ADF opioids.90 Health care costs were assigned to the ADF and non-ADF cohorts, with patients abusing opioids having higher health care resource utilization and costs than therapeutic users. These costs included costs of emergency room visits, inpatient and outpatient visits, and associated professional fees. Costs were obtained from a claims study conducted by the Commonwealth of Massachusetts Health Policy Commission specifically for this report.91 The mean annual cost was estimated to be $19,285 for therapeutic users and $31,005 for those who abuse opioids. All costs were expressed in 2016 dollars, and adjusted as necessary based on the medical care component of the U.S. Consumer Price Index.92 The model was informed by several key assumptions, including: • The rates of abuse with ADF or non-ADF opioids were kept constant throughout the time horizon of the model, owing to a lack of published data on variability in the rates of abuse over time. • We assumed an annual rate of cessation of opioid abuse of 10%, with patients who stop abuse incurring 50% of drug and non-drug costs in the year of cessation of abuse, prior to dropping out of the model. • We assumed the same health care resource utilization costs for ADF and non-ADF opioid therapeutic users, and for ADF and non-ADF opioid patients who abused these opioids, in the absence of data suggesting an impact of an ADF opioid on other health care costs. • Our model uses inputs from commercially-insured populations, as complete data for Medicare or Medicaid populations were not available. • We assumed the same rate of discontinuation of therapeutic opioid use in both the ADF and non-ADF opioid cohorts due to a lack of data on the individual cohorts. • We did not include effects of diversion or switching to other opioids or to heroin in our base- case analysis, due to lack of consistent data. The full list of assumptions and corresponding rationale for each is available in Section 6 of the full report. We conducted one-way sensitivity analyses, varying model parameters on the incidence of abuse, the efficacy of ADF opioids, the cessation of abuse, and drug costs. Given the limited data on the effectiveness of newer ADF opioids, we conducted threshold analyses, varying the rate of abuse to ©Institute for Clinical and Economic Review, 2017 Page ES21 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents determine reductions in the annual rate of abuse that would attain cost neutrality for ADFs relative to non-ADF opioid use. We undertook a similar cost-neutrality analysis by varying the costs of ADF opioids relative to non-ADF opioids. Finally, while opioid diversion and switching play a critical role in ascertaining the health and economic impact of the opioid abuse epidemic, we did not include these effects in our base-case analysis due to a lack of robust evidence. However, we conducted a scenario analysis to test for cost-neutrality between the ADF and non-ADF opioid cohorts by introducing different assumed rates of diversion into the model, based on data published by the Substance Abuse and Mental Health Services Administration (SAMHSA) finding that indicated that there are approximately 1.25 cases of diverted opioid abuse for every case of prescription opioid abuse.93 Using this as a reference point for the non-ADF opioid cohort, we estimated the reduction in relative risk of diversion in the ADF opioid cohort that would achieve cost-neutrality between the two cohorts. Finally, we also included a modified societal perspective as a scenario analysis, including the costs of criminal justice and incarceration, as well as costs of productivity loss due to opioid abuse.92,94 In the state-specific policy model, we analyzed the health and economic burden associated with opioid use in Massachusetts under a policy in which all non-ADF ER opioid prescription users in the state were switched to ADF ER opioid prescriptions. Key changes to this model compared to the cost-benefit model were: • Replacing the hypothetical cohort population in the cost-benefit model with a population based on the prevalent estimates of prescription ER opioid users in Massachusetts, derived from the state’s prescription drug monitoring program.95 • Deriving opioid drug costs based on a pharmacy claims analysis done by the Commonwealth of Massachusetts Health Policy Commission for this report, rather than the approach used in the cost-benefit model above.91 • Using the state-specific opioid overdose death rate.96 A more detailed explanation of model changes and key assumptions used for the state-specific policy model are available in Section 6 of the full report. Base Case Results Over a five-year time-horizon, our base case analysis indicated that there were approximately 2,300 fewer new cases of abuse in the ADF opioid cohort and approximately 6,600 fewer abuse-years incurred compared to the non-ADF opioid cohort, with a small reduction in opioid overdose-related deaths of less than one. ©Institute for Clinical and Economic Review, 2017 Page ES22 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table ES6. Abuse-Related Outcomes for ADF and Non-ADF Opioid Cohorts of 100,000 Chronic Pain Patients with ER Opioid Prescriptions Outcome (at 5 years) ADF cohort Non-ADF cohort Increment (ADF cohort – Non-ADF cohort) New case of abuse 8,229 10,532 -2,303 Person-years of abuse 23,322 29,943 -6,621 Overdose deaths 1.38 1.77 -0.39 Even with the cost-offsets within the health system from having fewer patients abusing opioids, use of ADF opioids resulted in an additional $533 million net spending over five years from the health care system perspective (Table ES7). The lower abuse-related costs of ADF opioids compared to non-ADF opioids were outweighed by the higher prescription costs of ADF opioids. Table ES7. Total Estimated Health-Care Costs of Patients Prescribed ADF and Non-ADF Opioids Over Five Years ADF opioids Non-ADF opioids Difference (ADF – non-ADF) Therapeutic use* $7,845,606,246 $7,692,466,543 $153,139,703 Abuse* $939,121,323 $1,205,748,255 -$266,626,932 Prescription opioid costs (entire $1,303,908,313 $657,301,870 $646,606,443 cohort) Total $10,088,635,882 $9,555,516,668 $533,119,214 *Excludes prescription opioid costs. Includes health care resource utilization and non-opioid prescription costs Using ADF opioids resulted in additional costs of $231,500 for preventing one new case of abuse and approximately $80,500 for preventing one abuse-year. Given the small benefit observed in overdose deaths, the cost to prevent an overdose death was estimated to be approximately $1.4 billion (Table ES8). Table ES8. Cost Per Incremental Outcome of ADF Opioid versus Non-ADF Opioid Incremental outcome Cost To prevent one new abuse case $231,514 To prevent one new abuse year $80,517 To prevent one overdose death $1,362,339,569 ©Institute for Clinical and Economic Review, 2017 Page ES23 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Sensitivity Analysis Results Results from our cost-neutrality threshold analyses indicated that increasing the effectiveness of ADF opioids to the point where they fully eliminate abuse still resulted in additional costs of approximately $113 million over five years. Cost-neutrality was achieved when the ADF opioid- weighted market share price was discounted by 41%, from $11.60 to $6.86 per day, keeping the base case incidence of abuse in each cohort constant. One-way sensitivity analyses indicated that the ADF opioid costs had the most significant influence on the model results among the parameters tested (see Section 6 for further details). Figure ES2. Incremental Health System Cost of ADFs at Increasing Levels of Effectiveness (Decreasing Incidence of Abuse) $600 $533 Incremental burden to health system $500 $430 $400 (In Millions) $326 $300 $220 $200 $113 $100 $0 2.818%* 2.11% 1.41% 0.70% 0.00% Incidence of abuse with ADF opioids *Represents base case Scenario analyses Diversion We included diversion as a scenario analysis. We tested the level of reduction in relative risk of diversion of ADF opioids that would be needed to attain cost-neutrality relative to non-ADF opioid use. We conducted this analysis at three different estimates of diversion of opioids: 1.25, one, and 0.75 cases of diverted abuse for every one case of prescription non-ADF opioid abuse.93 The cases of diverted abuse were added to the cases of prescription abuse in each cohort. Assuming 1.25 ©Institute for Clinical and Economic Review, 2017 Page ES24 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents cases of diverted opioid abuse for every case of prescription opioid abuse, the risk of diversion of ADF opioids would need to be 35% lower compared to that with non-ADF opioids to achieve cost- neutrality between the ADF and non-ADF opioid cohorts. Similarly, assuming 1.0 and 0.75 diversion cases per abuse case would require reductions of 44% and 59% in the risk of diversion of ADF opioids, respectively, to achieve cost-neutrality. More details on this scenario analysis are available in the section 6 of the report. Modified societal perspective The societal costs of each case of abuse were estimated to be approximately $3,400 annually for criminal justice and incarceration, and approximately $16,600 annually for lost productivity.92,94,97 Birnbaum et al. derived health care and societal costs using data from a claims analysis that included privately insured patients, in which the per-patient opioid abuse-related health care and productivity loss costs as well as associated caregiver costs were estimated. Criminal justice and incarceration costs were calculated using data from the Criminal Justice Expenditures and Employment Extract Program. Including these societal costs in our model, the difference in total net spending between the ADF and non-ADF cohorts over five years was reduced, but still represented an increase of $393 million in the ADF cohort. A breakdown of total costs within each cohort, including societal costs, is available in Table 23 in Section 6. State-specific Policy Analysis for Massachusetts We conducted a state-specific model analysis of Massachusetts in which we used the actual number of prevalent cases of prescription opioid use in the state, and calculated health outcomes and costs of converting all non-ADF prescription opioids to ADF prescriptions over one year. This analysis does not take into account the health and economic impact associated with diversion, switching to other opioids or heroin, or societal costs. Our analysis used mean daily costs of $15.90 for ADF opioids and $3.44 for non-ADF opioids, based on a claims study conducted by the Commonwealth of Massachusetts Health Policy Commission.91 Using 2015 data claims data, we estimated approximately 173,000 prevalent non-cancer chronic pain patients using prescription ER opioids in Massachusetts, of which approximately 60,000 were prescribed ADF opioids and approximately 113,000 prescribed non-ADF opioids. All 173,000 prevalent users were non-cancer chronic pain patients. Converting all non-ADF to ADF prescriptions was estimated to result in approximately 850 fewer cases of abuse in one year, at an estimated cost of approximately $599,000 to prevent one case of abuse (Table ES9). While abuse-related costs would decline (from approximately $225 million to $204 million), prescription opioid costs would more than double, leading to an increase in costs statewide of $475 million annually. ©Institute for Clinical and Economic Review, 2017 Page ES25 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table ES9. Outcomes When Converting All Non-Cancer Chronic Pain Patients with Prescription ER Non-ADF Opioids to ADF Opioids in Massachusetts in One Year Mixed ADF/non-ADF All ADF opioid use Difference opioid use Abuse cases 5,229 4,387 -842 Prescription opioid costs $489,925,522 $1,002,689,521 $512,763,999 Abuse-related costs* $224,828,862 $203,548,318 -$21,280,544 Total healthcare costs $5,331,764,758 $5,806,899,717 $475,134,959 Cost to prevent one new -- -- $599,131 case of abuse using ADF opioids *Combination of prescription (opioid and non-opioid) and resource utilization costs Summary and Comment We analyzed the cost-benefit of ADF opioids compared to non-ADF opioids in a hypothetical cohort model of non-cancer chronic pain patients, as well as a state-specific policy model. In the hypothetical cohort cost-benefit model, ADF opioids prevented 2,300 new cases of abuse per 100,000 patients treated over five years, but cost the health system an additional $533 million over that time span. We estimated that using ADF opioids costs the health care system an additional $231,500 to prevent one new case of abuse and approximately $80,500 in additional health system costs to prevent one year of abuse. Health care cost neutrality could not be achieved even when the effectiveness of ADF opioids in preventing abuse was increased to 100%, with ADF opioids still incurring an additional cost of $113 million over five years. However, cost neutrality could be achieved if ADF opioids were discounted by 41% from the current market-basket price. We also conducted this analysis using a modified societal perspective which included estimates of the productivity loss and criminal justice and incarceration costs. In this analysis, ADF-opioids were estimated to cost approximately $393 million more than non-ADF opioids over five years. Our state policy model, focused on Massachusetts, estimated that converting all existing non-ADF opioid prescriptions to ADF prescriptions over one year would prevent approximately 850 new cases of abuse at a cost of $599,000 for every new case of abuse prevented, and increase statewide costs by approximately $475 million. There are several key limitations of our analyses. Our model assumed a static rate of opioid abuse that does not change over time. We found no published evidence on rates of abuse over time and so our model may under- or over-estimate the actual burden of abuse over five years. In addition, costs and health care resources utilized by therapeutic users and those who abuse opioids do not ©Institute for Clinical and Economic Review, 2017 Page ES26 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents change over time in our model. We found one study that reported variations in health care cost for patients with opioid abuse in the six months prior to and 18 months after abuse diagnosis, but did not find similar estimates for costs over a longer time-frame.98 Varying these costs over time would impact the overall cost to the health care system, depending on the direction and magnitude of this cost variation over time. Although there are ADF opioids with more advanced technologies and perhaps greater potential in reducing abuse now on the market, we used effectiveness data from an OxyContin study owing to lack of abuse-related effectiveness data for other ADF opioids. In addition, our primary model does not include diversion to a population outside the existing cohort. To fully capture the cost to the health care system and to society of such diversion, we would also need to include the costs of switching to other opioids or heroin among individuals who do not abuse ADF opioids. The balance of these two effects of ADF opioids cannot be determined from current data, and modeling just the potential benefits of ADF opioids in reducing diversion without also including estimates of the potential harms from increased abuse of heroin or other opioids would not provide policymakers with a balanced view of the likely effects of increased ADF opioid use. In summary, our economic modeling analyses indicate that ADF opioids have the potential to substantially reduce the incidence of abuse in opioid-prescribed chronic pain patients relative to non-ADF opioids, but at significantly higher costs to the health care system. Even when important societal costs are included, ADF opioids were still estimated to increase overall costs. The advent of new ADF opioids with potentially superior abuse-deterrent properties, as well as the lack of robust evidence on opioid diversion and switching to other opioids or heroin, call for further research that will generate real-world evidence to understand the true health and economic impact of ADF opioids on the opioid abuse epidemic. Public Deliberation and Evidence Votes At the July 20, 2017 meeting, the New England CEPAC Panel voted on key questions concerning whether ADFs reduce the risk of abuse in comparison to non-ADF opioids. The first three questions focused on the risk of abuse for the individual patient who is prescribed an opioid. The fourth question sought to elucidate the net health benefit for the broader population of patients who obtain opioids through both legitimate prescriptions for therapeutic use and abuse/diversion. In the final two questions, the CEPAC was asked to consider three policy applications for managing the introduction of ADF opioids. More details on the voting results are provided in Section 7 of the full report. ©Institute for Clinical and Economic Review, 2017 Page ES27 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 1. For a patient being considered for a prescription of an immediate release opioid, is the evidence adequate to demonstrate a reduced risk of abuse for patients using RoxyBond versus non-ADF immediate release opioids? Yes: 2 No: 10 2. For a patient being considered for a prescription of an extended release opioid, is the evidence adequate to demonstrate a reduced risk of abuse for patients using OxyContin versus non-ADF extended release opioids? Yes: 9 No: 3 3. For a patient being considered for a prescription of an extended release opioid, is the evidence adequate to demonstrate a reduced risk of abuse for patients using all other ADFs, excluding OxyContin, versus non-ADF extended release opioids? Yes: 1 No: 11 4. For a patient being considered for a prescription of an extended release opioid, is the evidence adequate to demonstrate a net health benefit on a population level with the use of OxyContin versus non-ADF extended release opioids? Yes: 2 No: 10 5. Clinicians and policymakers are making efforts to reduce the numbers of patients started on opioids, limit the time course and refills for opioid prescriptions, and enhance monitoring for potential diversion and misuse of opioids. In addition, ADF-substitution policies are being considered to shift opioid prescriptions toward abuse-deterrent formulations. Considering the broad potential impact of substitution policies on patients, diversion, and illicit opioid use, which of the following policies do you believe would produce the most overall health benefit? a. Develop a way for physicians to work with academics and payers and policymakers to determine a way to target ADFs to high risk individuals and families. b. Allow physicians to determine whether to shift current patients to ADF opioids; require all new opioid prescriptions to be written for an ADF opioid. ©Institute for Clinical and Economic Review, 2017 Page ES28 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents c. Require all current non-ADF prescriptions to be substituted with ADF and all new prescriptions to be written for an ADF opioid. A: 10 B: 2 C: 0 6. Clinicians and policymakers are making efforts to reduce the numbers of patients started on opioids, limit the time course and refills for opioid prescriptions, and enhance monitoring for potential diversion and misuse of opioids. In addition, ADF-substitution policies are being considered to shift opioid prescriptions toward abuse-deterrent formulations. Considering the broad potential impact of substitution policies on patients, diversion, and illicit opioid use, which of the following policies do you believe would be the best for policymakers to pursue? a. Develop a way for physicians to work with academics and payers and policymakers to determine a way to target ADFs to high risk individuals and families. b. Allow physicians to determine whether to shift current patients to ADF opioids; require all new opioid prescriptions to be written for an ADF opioid. c. Require all current non-ADF prescriptions to be substituted with ADF and all new prescriptions to be written for an ADF opioid. A: 12 B: 0 C: 0 Key Policy Implications and Recommendations Following its deliberation on the evidence, the New England CEPAC Panel engaged in a moderated discussion about the use of ADFs for preventing abuse, diversion, and overdose death with a Policy Roundtable that included one patient representative, two clinical experts, two payer representatives, one state policy representative and one representative from a coalition of innovators and manufacturers. Many of the Roundtable themes revolved around the challenges of balancing the introduction of ADFs, which are currently available only for ER opioids, with the resulting shift in abuse towards other prescription opioids and illicit opioids, such as heroin and fentanyl, and in defining the appropriate role of ADFs as part of a multifaceted strategy for combating the opioid abuse crisis. The Policy Roundtable discussion with the New England CEPAC Panel reflected multiple perspectives and opinions, and therefore, none of the recommendations below should be taken as ©Institute for Clinical and Economic Review, 2017 Page ES29 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents a consensus view held by all participants. Below are the top-line policy implications; for more information, please see Section 7.4 in the main report. 1. ADFs should have a growing role in clinical practice since we believe they are safer for the individual patient. Still, policymakers should be fully aware that there is some evidence that demonstrates their impact on shifts in abuse towards other illicit drugs following introduction of ADFs. Mandatory ADF substitution laws may cause more harm to the overall population by shifting abuse towards other, potentially more lethal opioids and heroin. Policymakers and clinical leaders should therefore consider measures that would phase in ADFs while ensuring adequate support for other arms of a multi-pronged approach to the opioid crisis. 2. In addition to uncertainty regarding the overall health effects of rapid substitution of ADF opioids for non-ADF opioids, at current price differentials between ADFs and non-ADFs any rapid requirement for substitution with ADFs would prove unaffordable. Policymakers should therefore avoid approaches to encouraging the use of ADFs that would be so costly that resources would be drained from other efforts needed to address the opioid crisis and drive up costs for patients and the health system at an unsustainable rate. 3. As part of their responsibility to address the national opioid crisis, manufacturers and payers must recognize a shared commitment to making ADFs affordable to patients and to the health system. Manufacturers of ADFs should moderate the exercise of their monopoly pricing power; and payers should accept that paying a premium for ADFs is reasonable and that barriers such as increased out of pocket payments should not be placed in the way of helping appropriate patients receive reasonably-priced ADFs. 4. The federal government should act with urgency to convene clinical experts, clinical pharmacists, patients, and payers to develop consistent methods to stratify the risk for abuse and diversion of opioids. As universal substitution of ADFs for all patients may not be advisable or feasible, these methods should be disseminated and used to help determine when patients and their environments present a high enough risk to warrant prescription with a tamper-resistant ADF. 5. The term “abuse-deterrent formulation” is confusing for prescribers, patients, and the public, and lends to misunderstanding about the risks for addiction and misuse of ADFs. It should be abandoned as quickly as possible. The FDA should reconsider whether it can use “tamper-resistant formulation” (TRF) instead, because many clinical experts believe this is the most accurate and useful term. If this is not possible, the FDA should explore other ©Institute for Clinical and Economic Review, 2017 Page ES30 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents labeling options; however, clinicians, payers, and policymakers face no barriers to using different terms and therefore should adopt “tamper resistant” as the preferred term for describing these formulations. 6. Opioids represent the greatest public health crisis this country has faced in many years. Public health policymakers at the federal level should educate the public about the risks of all opioids – ADFs and non-ADFs – through a major public health campaign, perhaps modeled on the techniques employed nationally to reduce smoking. 7. Medical school curricula and physician licensing exams should require physicians to demonstrate a robust understanding of the role of ADFs in clinical practice, specifically addressing misconceptions about the addictive nature of ADFs. Specific questions on ADFs should be added to the already required training on opioids. 8. Prescribing physicians should help patients understand that ADFs are not less addictive than non-ADFs. Physician groups, individual physicians, and clinical pharmacists should develop or share federally-developed materials on the proper storage and use of all opioids. 9. Robust clinical studies are needed to demonstrate the natural history of opioid abuse and the impact of ADFs on abuse among prescribed patients as well as the broader effects on diversion and drug switching. 10. Given that over 90% of opioid prescriptions are for immediate-release (IR) formulations, and that currently, no IR ADFs are on the market, further investment and development by manufacturers for IR ADFs is critical. 11. Exploring and removing barriers to access to non-pharmacologic treatments for pain patients will have the dual effect of ensuring access to treatment for patients while also addressing the public health concerns related to prescribing opioids. ©Institute for Clinical and Economic Review, 2017 Page ES31 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 1. Background 1.1 Introduction Background Opioids are substances that act on specific receptors in the brain and produce a variety of effects such as pain relief, euphoria, respiratory depression, constipation and others.99 They are either directly extracted from opium, obtained from the pods of poppy varieties, or produced semi-synthetically and synthetically. Opioids are used to treat acute and chronic pain that arises from a variety of causes, ranging from trauma to advanced illness. Every year, 100 million people in the U.S. suffer from pain, with 9-12% of these individuals experiencing pain that is considered chronic (i.e., lasting longer than three months).1 Opioid therapy is an essential component of chronic pain management for many patients, but the addictive and euphoric properties of these drugs make them vulnerable to misuse, abuse, addiction, and possible death by overdose. Since 1999, the number of deaths from prescription opioids in the U.S. has increased nearly fourfold, rising in parallel with the volume of dispensed prescriptions;2 since 2009, use of prescription opioids has killed more persons annually than car accidents.100 The health care utilization consequences are also significant; for every one death from prescription opioids, it is estimated that there are 10 treatment admissions for abuse, 32 emergency room visits for misuse or abuse, 130 people who become dependent on opioids, and 825 people who report non-medical use of these drugs.3 A variety of measures have been implemented to attempt to mitigate opioid abuse, one of which is the introduction of abuse-deterrent formulations (ADFs) of these drugs. An increasing number of ADF forms of prescription opioids, approved by the FDA based on guidance published in 2015,5 have reached the market during the last few years and nine ADF products are in the late-stage pipeline (Stage III or FDA submission).14,4 The following table provides an overview of different approaches for obtaining abuse-deterrence: ©Institute for Clinical and Economic Review, 2017 Page 1 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 1. Overview of Abuse-Deterrent Approaches Abuse-deterring approach Properties Examples Physical and chemical barriers Resists cutting, grinding, High-molecular-weight pulverizing; dissolving produces a polyethylene oxide in OxyContin, viscous substance that cannot be Arymo®, Hysingla® drawn into a syringe Agonist/antagonist combination Opioid with a corresponding Naltrexone in Embeda® antagonist; antagonist released only through tampering. Aversive agent Opioid is combined with an Sodium lauryl sulfate used in aversive agent released during Oxaydo® for intranasal abuse- tampering deterrence. Oxaydo® is not an FDA approved ADF. Prodrug Opioid is released after the parent Activation of PF614 in the drug is ingested and metabolized gastrointestinal tract by pancreatic (usually requires stomach enzyme); trypsin leading to the production of opioid is not activated through free oxycodone.101 alternative route of administration (e.g., snorting) As specified by the FDA, the abuse-deterrent technology does not change the addictive properties of the opioid itself; while ADFs may deter abuse, they are not abuse proof.5 This report focuses on the effectiveness, safety, and economic impact of ADFs relative to non-ADF opioid treatment, and considers the evidence and potential cost-benefit of different strategies to replace non-ADF formulations with ADFs in specific populations. Scope of the Assessment The scope for this assessment is described on the following pages using the PICOTS (Population, Intervention, Comparators, Outcomes, Timing, and Settings) framework. We conducted a systematic literature review using best practices for search strategy development and article retrieval. Evidence was culled from randomized controlled trials as well as high-quality systematic reviews; observational studies were considered given the difficulty of conducting randomized controlled trials for non-medical use of opioids. Our evidence review included input from experts, patients and patient advocacy organizations, data from regulatory documents, information submitted by manufacturers, and other grey literature when the evidence met ICER standards (for more information, see https://icer- review.org/methodology/icers-methods/icer-value-assessment-framework/grey-literature-policy/). ©Institute for Clinical and Economic Review, 2017 Page 2 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Analytic Framework The analytic framework for this assessment is depicted in Figure 1. Figure 1. Analytic Framework Patient and Non-ADF population prescription Intermediate outcomes: outcomes in • Physical Dependence clinical trials • Addiction • Stigma of opioid use • Overdose deaths ADF • Misuse and abuse prescription Diversion Health system outcomes: • Diversion Rates • Health system costs • Patterns of illegal drug use Societal outcomes: • Criminal activity • Productivity • Education outcomes Populations The population of focus for the systematic review of the clinical impact of ADFs included all persons using opioids for therapeutic (i.e., both as prescribed and misused) and non-therapeutic purposes (i.e., abuse, addiction). For modelling purposes, the population has been defined more narrowly, as described in section 6 of the present report. Interventions The interventions of interest were abuse-deterrent opioid formulations with an FDA label (ADFs). Opioids with abuse-deterrent properties but without an FDA label recognizing these properties were ©Institute for Clinical and Economic Review, 2017 Page 3 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents not included in the assessment. Currently, there are ten opioid products (nine extended-release [ER] and one immediate-release [IR]) that have U.S. FDA-approved abuse-deterrent labeling, and all were included in this review.102 However, only five products are available in the U.S. marketplace as of June 28, 2017. RoxyBond®, the only ADF for an immediate release (IR) oxycodone formulation, although not originally included in the scope of this review, was most recently approved by the FDA on April 20, 20174 and has therefore been included in the evidence review. Oxycodone: • OxyContin (oxycodone extended release; available on the market) • Xtampza® (oxycodone extended release; available on the market) • Troxyca® ER (oxycodone + naltrexone extended release; approved, but not available on the market as of June 27, 2017) • Targiniq® ER (oxycodone + naloxone extended release; approved, but not available on the market as of June 27, 2017) • RoxyBond® IR (oxycodone immediate release; approved, but not available on the market as of June 27, 2017 Hydrocodone: • Hysingla® ER (hydrocodone extended release; available on the market) • Vantrela® ER (hydrocodone extended release; approved, but not available on the market as of June 27, 2017) Morphine: • Embeda® (morphine + naltrexone extended release; available on the market) • Morphabond® (morphine extended release; approved, but not available on the market as of June 27, 2017) • Arymo® ER (morphine extended release; available on the market) Comparators The comparators of primary interest included non-abuse-deterrent formulations of specific opioids as appropriate. Outcomes Patient & Population Level: The impact of ADFs on individual patients was assessed by evaluation of the following outcome measures, including addiction rates and other clinical outcomes, many of which are surrogate outcomes currently being used by the FDA in granting marketing approval. Importantly, ©Institute for Clinical and Economic Review, 2017 Page 4 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents outcomes related to pain alleviation and tolerability were not included, as ADFs are considered bioequivalent to their relevant non-ADF counterparts.24 • Patient/Population Level Outcomes o Abuse Potential Endpoints ▪ VAS measures (0-100) of drug liking, take drug again, and overall drug liking ▪ Tampering o Real World Evidence of Abuse and Misuse ▪ Overdose and fatality ▪ Abuse/misuse • Physical evidence of misuse/abuse • Self-reported misuse/abuse • Route of administration for misuse/abuse ▪ Addiction • Health System Level Outcomes o Health system costs o Drug loss and diversion rates o Patterns of illegal drug use o Doctor shopping o Prescription utilization • Societal Level: Where evidence is available, we also sought to capture the societal impact of ADFs, including outcomes related to the criminal justice system, worker productivity, and education. The analysis of outcomes was based on a systematic literature review of peer-reviewed publications and on evidence from the grey literature meeting ICER standards (for more information, see https://icer-review.org/methodology/icers-methods/icer-value-assessment-framework/grey-literature- policy/). Timing Evidence on intervention effectiveness and harms were derived from studies of any duration. Settings All relevant settings were considered, including inpatient, clinic, office, and home settings. ©Institute for Clinical and Economic Review, 2017 Page 5 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 2. The Topic in Context 2.1 Overview Opioids are risky medications. While they are the strongest analgesics available, they also are responsible for abuse, addiction, and death. This is not surprising, as both therapeutic and harmful effects rely on the same receptor in the central nervous system and are therefore present in all opioids. Together with Canada, the U.S. has the highest per capita consumption of prescription opioids worldwide.103-105 Around 10% of patients receiving opioids for the first time use them for more than three months. Of those patients, about 25% begin non-medical use and 10% become addicted.106 In this report, the term ADF is only used for drugs with abuse-deterrent properties as recognized by the FDA. In order to be defined as ‘abuse-deterrent’ in the drug label, there must be sufficient evidence of abuse deterrence according to FDA standards, which is based on premarket studies and mandatory real world studies after drug approval.5 In April 2015, the FDA issued non-binding recommendations encouraging manufacturers to produce abuse-deterrent formulations (ADFs) of opioids. Between 2016- 2017, the FDA approved five new ADFs; today, nine extended-release (ER) opioids and one immediate release (IR) opioid have FDA-approved labeling describing a variety of abuse-deterrent properties (Table 2). ADFs are relatively new, branded therapies for treating pain, and are generally more expensive than their non-ADF equivalents (both branded and generic versions). Complicating this picture is the absence of any true generic opioids in an ADF form (see Section 2.3 for further details). In 2016, the VA spent approximately $100 million overall on opioids. If all opioids were to be replaced with ADFs the costs would be increased 10-fold on the average, and this “would result in approximately $1 billion yearly for these products and could represent as much as 20 percent of the VA pharmacy budget”.18 Policymakers will therefore be challenged on how to structure conversion to ADFs in a responsible and economically feasible manner.19 While ADFs may be more expensive opioid therapies, they also may achieve cost savings by reducing abuse and abuse-related events in both patients who are prescribed opioids and in individuals who do not obtain opioids through a prescription (e.g., “diversion”). In Section 4, we evaluate the evidence on the impact of ADFs on abuse, and in Section 6, we evaluate the cost-benefit of ADFs that includes the added costs of the prescription and potential cost-savings savings in abuse-related care, in both patients who obtain opioids through a prescription, and patients who obtain opioids through diversion. It is also important to consider the impact of ADFs on overall trends in opioid abuse, since persons already abusing specific opioids may shift to other opioids or routes of administration if a specific opioid is replaced with an abuse-deterrent formulation. In recent years, the dramatic and parallel increase in ©Institute for Clinical and Economic Review, 2017 Page 6 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents the use of prescription opioids and deaths from overdose of these substances has leveled off, but a continuing increase in opioid deaths is now driven by heroin and illegally-produced synthetic opioids such as fentanyl.2 In Section 4, we summarize the available literature on any correlations between the introduction of ADFs with alternative opioid abuse patterns or methods of administration. Finally, despite the overall increase in opioid prescriptions, many patients with chronic pain receive inadequate analgesia.107 Some patients report increased difficulty maintaining access to treatment with prescription opioids, as described below in the section detailing insights gained from discussions with patients and patient groups. Patient groups focused on addiction cautioned that individuals who are unable to abuse a particular opioid may substitute an easier-to-abuse option. ©Institute for Clinical and Economic Review, 2017 Page 7 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 2. Opioid Products with FDA-Approved Abuse-Deterrent Labeling Brand Name Type of Year of Reported Abuse-Deterrence Mechanism Commercially Opioid Approval Available± OxyContin® Oxycodone 2010 When dissolved, forms a viscous gel that is difficult to inject through a hypodermic Yes (reformulated) needle. Embeda® Morphine 2014 Capsules of ER morphine pellets that contain a sequestered core of naltrexone; if the Yes pellets are swallowed, the morphine is gradually released and absorbed, while the naltrexone core passes through the gut intact. If the pellets are crushed, chewed, or dissolved, the naltrexone is released, blocking morphine-induced euphoria. Targiniq® ER Oxycodone 2014 Combination pill containing extended-release (ER) oxycodone and naloxone; if the No formulation is crushed and administered intravenously or intranasally, high naloxone concentrations block opiate-induced euphoria and can induce withdrawal symptoms. Hysingla® ER Hydrocodone 2015 When dissolved, forms a viscous gel that is difficult to inject through a hypodermic Yes needle MorphaBond® Morphine 2015 Formulated with inactive ingredients that make the tablet harder to adulterate while No maintaining ER characteristics if the tablet is subjected to physical manipulation or chemical extraction. Xtampza® ER Oxycodone 2016 Capsules containing microspheres formulated with oxycodone base and inactive Yes ingredients that make the formulation harder to manipulate. Troxyca® ER Oxycodone 2016 Contains pellets that consist of oxycodone that surround sequestered naltrexone. When No taken orally, the naltrexone is intended to remain sequestered and patients receive ER oxycodone. When the pellets are crushed, the naltrexone is released and counteracts the effects of oxycodone. Arymo® ER Morphine 2017 A polymer matrix tablet technology with controlled-release properties as well as physical Yes and chemical barriers that resist manipulation. The technology results in a viscous hydrogel on contact with liquid, making the product very difficult to draw into a syringe. Vantrela® ER Hydrocodone 2017 Incorporates abuse-deterrent technology designed to resist drug extraction through the No most common routes: oral, intranasal, and intravenous. **RoxyBond® Oxycodone 2017 Includes inactive ingredients that make the tablets harder to misuse by physical No manipulation, chemical extraction, or both; in vitro data suggest physicochemical properties that are expected to make abuse through injection difficult. 20 *Modified from Becker, 2017. ** Only ADF approved as immediate-release. ±As of June 28, 2017. ©Institute for Clinical and Economic Review, 2017 Page 8 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 2.2 Use and Abuse of Prescription Opioids Use of Prescription Opioids Opioids affect the mu receptor in the spinal cord and in the brain to reduce pain.108 The mu receptor in the brain is also central to the feelings of reward or pleasure, leading to abuse.109 The analgesic effects are mediated mainly through the mu receptors’ release of substance P in the spinal cord,108,110 the central neurotransmitter for pain, whereas the rewarding effects involves the dopaminergic system which is implicated in all addictive behavior, including that of alcohol and nicotine.111 Most opioids are short-acting and require dosing approximately every four hours, with the exception of methadone, which requires dosing only two to three times per day.112 Methadone is intrinsically long acting (LA), while the other opioids require special formulations that enable extended release (ER). Methadone and the ER opioids are often lumped together under the term ER/LA opioids. In December 1995, OxyContin™ was the first ER opioid approved by the FDA.4 ER formulations represent about 10% of all opioids prescribed.12,13 ADFs of extended-release (ER) opioids aim to prevent what is known as “dose-dumping,” the rapid release through tampering of drug contained in an ER dosage form, which is at much higher concentrations than that found in IR opioids. For example, taking a 12 mg Palladone capsule, an ER hydromorphone no longer on the market, together with 8 ounces of 40% ethanol increased the average blood hydromorphone concentration by 6 times compared with co-ingestion with water.9 Increasing the blood opioid concentration can also be achieved by altering the route of administration. The oral bioavailability of hydromorphone is about 10%, which means that an intravenous injection of the same substance increases the bioavailability tenfold in addition to the more rapid onset and rise through the dose- dumping route.10,11 During the beginning of the second half of the 20th century, opioids were infrequently used in the treatment of chronic pain.113 In 1992, the Agency for Health Care Research and Quality (AHRQ) issued a guideline for acute pain management stating that “patients have a right to treatment that includes prevention of or adequate relief from pain and that fears of postsurgical addiction to opioids are generally groundless.”114 Pain management was promoted and recognized as a human right115,116 and pain included as a “fifth vital sign” in the pain management standards of the Joint Commission on Accreditation of Healthcare Organizations, to be monitored with the same vigilance as blood pressure, pulse, temperature, and respiratory rate.117,118 In 2012, patient experience in regard to pain management became one component of the newly created Hospital Value-Based Purchasing (HVBP) program, which ties a portion of hospital payment to performance on quality and cost, possibly encouraging physicians to increase the prescription of opioids.119 These new ©Institute for Clinical and Economic Review, 2017 Page 9 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents professional standards, combined with aggressive marketing120 led to a fourfold increase in the volume of dispensed prescription opioids between 2000 and 2010.121 Defining Terms of Abuse The concepts and terminology of drug-related problems are constantly evolving and sometimes contradictory between the different medical specialties. For example, in psychiatry, the Diagnostic and Statistical Manual (DSM-III) chose the term dependence to refer to uncontrolled drug-seeking behavior.122 In other branches of medicine, the term dependence refers to physical dependence.123 In the fifth revision of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), the terms abuse and dependence have been replaced with the term substance use disorders.122 For our report, we will use the terms abuse, dependence, and addiction with following meanings: • Abuse refers to the “intentional, non-therapeutic use of a drug product or substance, even once, to achieve a desirable psychological or physiological effect.”5 • Dependence refers to physical dependence: “a state of adaptation that is manifested by a drug class–specific withdrawal syndrome that can be produced by abrupt cessation, rapid dose reduction, decreasing blood level of the drug, and/or administration of an antagonist.”123 • Addiction refers to a “primary, chronic, neurobiological disease, with genetic, psychosocial, and environmental factors influencing its development and manifestations. It is characterized by behaviors that include one or more of the following: impaired control over drug use, compulsive use, continued use despite harm, and craving.”123 Abuse and the Opioid Epidemic In today’s pharmacotherapy, the use of opioids is clinically limited to the treatment of acute or chronic pain as well as the treatment of opioid addiction with methadone. Pain is the most common complaint leading a patient to a physician, and opioids are the most common medications prescribed in the U.S.124 Diseases of the musculoskeletal system and connective tissue are the predominant indication for both IR and ER opioids, accounting for approximately half of all prescriptions. Cancer accounts for 5 to 10% of prescriptions,12,125 including treatment in palliative care but also chronic pain in long-term cancer survivors.126 The terms opioid overprescribing and overconsumption are often used to describe the current level of opioid use in the US.127 However, it is extremely difficult to define an appropriate level of overall population-based therapeutic opioid use. Using data from the International Narcotics Control Board, an Adequacy of Opioid Analgesic Consumption measure has been proposed to compare the level of opioid use worldwide for pain treatment. This measure is based on country mortality data ©Institute for Clinical and Economic Review, 2017 Page 10 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents for cancer, HIV, and injuries, and is normalized with the average opioid consumption for the top 20 countries with the highest Human Development index.103 A level of 100% or greater of the Adequacy of Opioid Analgesic Consumption measure is defined as an adequate level of consumption. In 2010, the US had a figure of 230%.104 Worldwide, opioid consumption is highest in Canada and the U.S. (Figure 2). In the U.S., the volume of dispensed prescription opioids has stabilized recently due to different policy initiatives.128 Figure 2. Comparison of Total Opioid Consumption105 Our clinical knowledge of abuse and addiction is increasingly informed by our understanding of neurobiological mechanisms.129 Addiction is, however, not simply “a disease of exposure;” a person must also be vulnerable (e.g., genetically), and the exposure must occur at a vulnerable time, such as under conditions of stress or due to age.130 ©Institute for Clinical and Economic Review, 2017 Page 11 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Understanding the characteristics and pathways of individuals at higher risk of abuse is quite challenging. For example, we do have some information on the routes of abuse for patients entering drug rehabilitation programs, but are understanding is limited concerning the pathways leading from misuse to abuse, including specific information on recreational abuse, and finally to addiction. Furthermore, these pathways probably differ among different age groups. In susceptible individuals, the exposure to opioids for relieving pain can lead to a spiral of abuse, addiction, and death. Overall, around 10% of patients receiving opioids for the first time will use them for more than three months106. The probability of opioid use after one year increases from 6% among patients with at least one day of opioid therapy to 13.5% for persons whose first episode of use was for ≥8 days, and to approximately 30% when the first episode of use was for ≥31 days131. According to a systematic review, between 8% and 12% of patients receiving opioids for longer than three months become addicted132. Therefore, if 10% of patients used opioids for >3 months, and 8- 12% of those patients become addicted, then the population-level risk for any new opioid use is about 1%, a significant risk given the size of the population receiving opioids. These data provide some insight into the risks for patients arising from opioid prescriptions, but they do not include the risks from diversion, described later in this section. Abuse risk for patients is an area of active research, but there are currently no validated tools for predicting increased risk for abuse for specific patients or patient groups31,36. The pathways of progressing from medical use to non-medical use, to abuse and addiction, has not been very well studied. It is generally believed that chewing an ER opioid is an important step towards addiction, followed by intranasal and intravenous routes of abuse.27 However, even among patients entering drug rehabilitation programs, oral abuse of the IR formulation or the manipulated ER formulation remains the major route, with the exception of morphine, which intravenous abuse is the most common route (Table 3). Table 3. Estimated Prevalence of Routes of Abuse29* Prescription Opioid Oral Snort Inject Analgesic Hydrocodone 88% 25% <10% Oxycodone 76% 45% 22% Morphine 40% 29% 66% Methadone 71% 10% <10% *Abusers often use more than one route Opioid abusers often manipulate ER opioids and opioids with a low bioavailability through the oral route for a “‘dose-dumping’ effect (i.e., an increased maximum concentration of the opioid in the brain in the shortest possible time); this is associated with the occurrence of a rapid high and other reinforcing effects, which drive further abuse potential.”133 ©Institute for Clinical and Economic Review, 2017 Page 12 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Prescription opioids with abuse-deterrent properties are meant to prevent progression of patients to abuse and addiction. They are not influencing the demand for the drug and they are not a treatment for addiction.28 This means that persons already abusing specific opioids are likely to shift to other opioids or routes of administration if a specific opioid is replaced with an abuse- deterrent formulation. For example, this was the case for Opana ER, where the replacement in 2012 of the original formulation with an abuse-deterrent formulation for the intranasal route resulted in a shift of abuse instead toward the intravenous route.15 Since 2011, the continuing rise in opioid deaths are no longer attributable solely to prescription opioids but also to illicit opioids, mainly heroin and illegally manufactured fentanyl (Figure 3). Death rates and opioid overdoses are concentrated in states with large rural populations, such as Kentucky, West Virginia, Alaska, and Oklahoma.134 Hot-spots in prescription opioid overdoses show a spread in time from rural to suburban areas.135 While the death rates have increased overall, the greatest increases have been observed in New England states, with the most significant increase in New Hampshire.136 Figure 3. Overdose Deaths Involving Opioids 2000-20152 As ADFs enter the market, it is also critical to understand trends in abuse, since persons already abusing specific opioids may shift to other opioids or routes of administration if a specific opioid is replaced with an abuse-deterrent formulation. In Section 4, we summarize the available literature ©Institute for Clinical and Economic Review, 2017 Page 13 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents that correlate the introduction of ADFs with alternative opioid abuse patterns or methods of administration. Diversion Many individuals who abuse opioids do not receive a prescription from a prescriber. This is known as diversion, or the transfer of a prescription drug from a lawful to an unlawful channel for distribution or use.34 Diversion can occur at any of the different points in the drug delivery process: via the original manufacturing site, the wholesale distributor, the physician's office, the retail pharmacy, or the patient.112 The annual National Survey on Drug Use and Health of the Substance Abuse and Mental Health Services Administration, an annual self-report survey of the civilian, non- institutionalized population in the United States, provides the only population-based data on the sources of prescription opioids for non-medical use.137 About 50% of people who misused prescription opioids got them from a friend or relative for free, while 22% got them from a doctor and only 4% bought them from a drug dealer.32,33 The figures vary with the intensity of the abuse: people abusing prescription opioids up to 30 days in a year are receiving the drug for free in 62% of cases. This goes down to 26% for those abusing more than 200 days a year, with drugs increasingly being bought from friends or from drug dealers.138 The volume of prescription opioids diverted annually for non-medical use is extremely difficult to estimate. According to one estimation using several public and private databases, about 4% of all prescription opioid doses dispensed in 2002-2003 were used non-medically.139 Street prices of specific opioids can be a good indicator of drug availability, demand, and abuse potential.35 It is important to understand how the introduction of ADFs impact diversion and the availability of illicit opioids in order to capture their true impact on overall abuse, including for abusers who obtain opioids through diversion and not through a prescription. We summarize the available evidence on diversion in the Comparative Clinical Effectiveness section on page 41. 2.3 The FDA Designation for Abuse-Deterrent Formulations of Opioids The first two abuse-deterrent formulations of an opioid were introduced in the U.S. in 1960 and 1978. Lomotil® and Motofen® contained opioids for the treatment of diarrhea with atropine being added as an aversive agent to prevent abuse. Talwin NX® followed in 1982, using naloxone as an opioid antagonist.140 Following the general process of drug approval, the FDA did not list abuse- deterrent properties, or “tamper-resistant” properties as they were known at that time, in the label without epidemiological evidence on the real-world effectiveness of abuse-deterrent formulations.141 As a result, some opioids with abuse-deterrent technologies are available on the market without official recognition in the FDA drug label.142 Among the different approaches to ©Institute for Clinical and Economic Review, 2017 Page 14 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents diminish the abuse potential of opioids, as described in the background section, ADFs currently on the market use only physical and chemical barriers and agonist/antagonist combinations. In 2013, the FDA published the draft guidance Abuse-Deterrent Opioids — Evaluation and Labeling that was approved in 2015.5 The guidance contains four types of data requirements: the first three categories are premarket studies, mandatory for FDA approval, while category 4 is mandatory to be conducted after said approval (Table 4). It should be noted that OxyContin was approved with an ADF label in 2010,4 prior to the mandatory requirement of category 4 studies. Table 4. FDA Data Requirements for the Approval of an ADF Label5,143 Category Type of studies Description143 1 Laboratory Manipulation and Studies designed to evaluate physiochemical properties, Extraction Studies characterize a product’s abuse-deterrent properties and the degree of effort required to defeat those properties 2 Pharmacokinetic Studies Studies designed to compare pharmacokinetic profiles of an intact and manipulated ADF product to a comparator drug through one or more routes of administration 3 Clinical Abuse Potential Studies conducted in drug-experienced, recreational user Studies populations designed to assess the impact of potentially abuse- deterrent properties 4 Postmarket Studies Studies designed to determine whether an ADF product results in meaningful reductions in abuse, misuse and related adverse clinical outcomes Results of category 1 to 3 premarket studies are surrogate outcomes for abuse liability, meaning that they can be considered reasonably likely to predict clinical benefit.144 Premarket studies do not involve pain patients, but healthy, non-dependent recreational drug users between the ages of 18 and 55 years. The scientific foundation and interpretations of these studies are constantly evolving;5 while clinical abuse potential studies have been validated for analytic performance, they have not been validated regarding their relationship to being able to predict clinical benefit (i.e., their ability to predict real world abuse).21 The methodology of the category 3 clinical abuse potential studies is described in section 4.3 of this report. Results of category 1 studies have not shown either to reliably predict outcomes of Category 2 and 3 studies.5 Postmarket studies (i.e., following regulatory approval) are also required by the FDA, and are designed to measure the real-world impact of ADFs on patterns of abuse and misuse. As noted in a recent FDA paper prepared for an FDA public meeting on July 10-11, 2017,23 studies of prescription drug abuse differ from traditional pharmacoepidemiologic investigations: ©Institute for Clinical and Economic Review, 2017 Page 15 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents • Product-specific exposure can be problematic to determine because it often occurs outside the health care system; outcomes commonly occur in individuals not prescribed opioid products (i.e., “diversion”). • Many factors that can affect both the ability to assess and the overall levels and trends in prescription drug abuse are not captured in clinical information (e.g., state and federal law enforcement and policy changes, regional trends). • No national-level data resource can provide estimates of prescription drug abuse, at all levels of severity, and link those data to relevant clinical and other information needed to form a comprehensive assessment of the problem. • Available data resources generally capture one aspect of interest (abuse, clinical, or mortality data) without the ability to link to other relevant datasets. • Outcomes that come to medical attention cannot generally be linked to a specific product or products.”22 The FDA public meeting on July 10-11, 2017 aims to improve the use of “existing data sources and methods to evaluate the impact of these products in the real world, as well as what new data sources and study designs could be developed or enhanced to ensure these efforts result in the best possible answers to inform regulatory decision-making.”23 Currently, nine ER opioids and one IR opioid have received labeling describing abuse-deterrent properties (see Table 2). The only generic ADFs on the market are authorized generics of OxyContin:15,16 the pills are identical to the original OxyContin, and the price is higher than that of extended release generics approved through the traditional process.17 This absence of true generic competition contributes to the higher prices of currently available ADFs. ADFs and their non-ADF counterparts have the same profile of adverse effects when used as prescribed.24 However, when abused, the ADFs may present particular safety issues, such as precipitated severe withdrawal symptoms when an ADF with an agonist/antagonist combination is chewed or crushed. The reformulation of Opana ER in 2012 with a high-molecular-weight polyethylene oxide physical and chemical barrier led to a shift from intranasal to intravenous abuse.83 “The high street cost of the product coupled with the method of preparation contributed to IV users sharing the drug solution and the equipment used to prepare and inject it,” leading to an outbreak of HIV and HCV infections in Indiana.83,145 The outbreak was controlled by implementing a needle exchange program.84 In Tennessee, a cluster of thrombotic microangiopathy is thought to be related to intravenous exposure of substances produced by the tampering of the polyethylene oxide barrier used as abuse-deterrent technology in Opana ER.146,147 Polyethylene oxide is also present in nine other ER opioids listed for an oral route of administration, including the ADFs Arymo®, Hysingla®, ©Institute for Clinical and Economic Review, 2017 Page 16 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents and OxyContin.15 Up to now, there does not seem to been any reports on similar safety concerns after tampering for intravenous abuse of these other opioids, but this type of impact is very hard to detect. These safety issues with the abuse-deterrent technology in Opana ER convinced the panel members of the FDA Drug Safety and Risk Management Advisory Committee and the Anesthetic and Analgesic Drug Products Advisory Committee at their meeting on March 13-14, 2017, to conclude that the benefits of reformulated Opana ER no longer continue to outweigh its risks.82 On June 8, 2017, the FDA requested that Endo Pharmaceuticals remove Opana ER from the market.148 In their decision, the FDA did not seem to have considered the argument made by the manufacturer concerning the benefit of the unique metabolic profile of oxymorphone.149 Indeed, oxymorphone is the only opioid with no known pharmacokinetic drug-drug interactions, an important safety consideration in older and medically complicated patients, who may be taking multiple medications.150 Safety issues with excipients after tampering for intravenous abuse have also been raised concerning the IR ADF RoxyBond® at the FDA Drug Safety and Risk Management Advisory Committee and the Anesthetic and Analgesic Drug Products Advisory Committee at their meeting on April 5, 2017.26 2.4 Policy Interventions: Clinical Guidelines and State Policies The context for understanding the potential benefits of ADFs is complex, as these technologies are often part of a multipronged strategy to combat the public health epidemic of prescription opioid deaths. This strategy often includes educating clinicians to reduce initiation of opioid use, shortening the duration of prescriptions, monitoring of prescriptions, and in some states, mandatory substitution of opioid prescriptions with ADFs. Further details on some of these initiatives are described below. In 2016, the Centers for Disease Control and Prevention (CDC) released the CDC Guideline for Prescribing Opioids for Chronic Pain for patients 18 and older in primary care settings. This new guideline constitutes the most recent professional reference for treatment decisions for chronic pain (outside of active cancer treatment, palliative care, and end-of-life care). The primary recommendation prioritizes nonpharmacologic and non-opioid therapy for chronic pain.36 The CDC guidelines recommend a universal approach of urine testing to be performed at least annually for all patients receiving an opioid for chronic pain;36 and judged that the evidence on clinical tools for identifying patients at higher risk for abuse was insufficient or absent.30 None of the 12 recommendations of the CDC guidelines meets a high standard of evidence, but they are judged to ©Institute for Clinical and Economic Review, 2017 Page 17 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents reduce harm and likely improve chronic pain control in the United States.37 The guidelines do not currently mention ADFs for treating patients with pain. The use of opioids in clinical practice is influenced by legislation and regulation at different levels. The production, distribution, and prescribing of prescription opioids is regulated by the Controlled Substances Act (CSA) enacted in 1970.151 Prescribers need to be registered with the Drug Enforcement Agency.152 Numerous states are regulating the duration of opioid prescriptions,152 and all states, excluding Missouri, have instituted prescription drug monitoring programs (PDMPs).152,153 Many payers, including the Centers for Medicare and Medicaid Services (CMS), have instituted programs monitoring opioid prescriptions, identifying patients deemed at risk for misuse or abuse.152 State governments have also stepped up efforts to address the opioid epidemic, with executive led taskforces, physician education, and legislation to establish prescription monitoring programs, restrict the duration and/or quantity available in an opioid prescription, and allocate more funding for abuse treatment options. In August of 2014, Massachusetts became the first state to pass legislation to require pharmacies to automatically substitute ADFs for chemically equivalent non- ADF opioid prescriptions. The law requires insurance carriers to cover ADFs in the same way they cover non-ADF opioids, with no additional cost burden to patients. The Massachusetts law creates a deliberative drug formulary commission to determine substitutable abuse-deterrent formulations of opioids for their chemically-equivalent generic equivalents. In order to prescribe a non-ADF opioid, a physician must explicitly state ‘No Substitution’ and provide adequate rationale. While the Massachusetts law was originally set to go into effect in 2016, implementation has been delayed because state officials are still establishing regulatory guidance for insurers and pharmacy providers. In 2015, Maine also passed ADF legislation requiring all health insurance carriers to provide coverage for ADFs, making them the preferred drugs on any formulary or preferred drug list for both acute and chronic uses. The law prohibits step therapy with non-ADF opioids before use of ADF opioids. In order to pass the legislation, the legislature voted to override the Governor’s veto. In New Hampshire, Vermont, Connecticut, and Rhode Island, legislators have introduced similar legislation that is still undergoing debate. In June 2016, the New Hampshire legislature passed a law that established a commission to study the preventative abuse potential and cost impact of ADFs. With the 2016 election underway—and a new governor entering the executive office—the commission was delayed, and the legislative requirement became void in 2017. In 2016, bills were introduced in 20 other states relating to ADF coverage. In 2015 and 2016, Maryland, Florida, and West Virginia passed similar legislation requiring that ADFs are covered with parity to non-ADF equivalents and prohibiting step therapy with a non-ADF opioid. Four states, ©Institute for Clinical and Economic Review, 2017 Page 18 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents including Delaware, New Hampshire, Oklahoma, and Virginia, passed resolutions requiring further study of ADFs. At this point, however, data on the impact of these state policy and systems-level interventions are limited and inconsistent.38 Within this legislative context, the assessment of ADFs as an effective and efficient strategy for curbing the epidemic of death from opioid overdose is urgently needed. During the last decade, numerous policy initiatives have emerged for combatting the epidemic of death from opioid overdose. For example, in 2015, the U.S. Department of Health and Human Services announced three priority areas to combat opioid abuse: (1) opioid prescribing practices to reduce opioid use disorders and overdose, (2) expanded use and distribution of naloxone, and (3) expansion of medication-assisted treatment (MAT) to reduce opioid use disorders and overdose.154 ADFs are not part of any of the proposed actions in these priority areas. In 2016, the Obama administration requested $27.6 billion for the fiscal year 2016 to support efforts under the 2015 National Drug Control Strategy to reduce drug use and its effects.155 This strategy lead to a memorandum to combat the prescription drug abuse and heroin epidemic that directed federal departments and agencies to provide training to prescribers and to improve treatment for prescription drug abuse and heroin use.156 Again, ADFs are not part of any of the proposed actions. The FDA Opioids Action Plan seem to be the only policy initiative on the federal level that prioritizes ADFs.1 In July 2012, the FDA implemented Risk Evaluation and Mitigation Strategies (REMS) class-wide for ER and long-acting opioids that requires manufacturers of these agents to distribute educational information to clinicians and patients and involves clinicians in monitoring of patients and counseling them on safe use.152 2.5 Insights Gained from Discussions with Patients and Patient Groups As part of our review, we spoke with patient organizations focused on chronic pain and addiction. Patient organizations focused on chronic pain stressed the need for continued, affordable patient access to opioid therapy for daily function while also recognizing the need to curb opioid misuse and addiction. It was felt that the different policy initiatives for reducing the overall use of opioids contributed to increasing difficulties in obtaining prescriptions for long term opioid therapy. Patients with chronic pain were nervous that higher co-payments for ADFs compared to non-ADF ER opioids could act as a potential barrier to accessing needed opioid therapy. Some patients with chronic pain saw the ADF designation as potentially smoothing access to necessary medication, as it might reduce the typical level of stigma associated with controlled substances. The importance of assessing the total clinical, economic, and social value of ADFs was widely recognized by the different stakeholders as an essential step for their rational use. ©Institute for Clinical and Economic Review, 2017 Page 19 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Overall, organizations representing patients with chronic pain reported patients’ difficulties accessing specialized multidisciplinary pain care. Some patients believe that access to integrated pain management, including medications and complementary approaches such as acupuncture, physical therapy, and mind–body practices would contribute to diminishing the need for prescription opioids. Patient advocates who worked with patients struggling with addiction helped to illustrate how patients progress to opioid addiction, often beginning with the recreational oral abuse of opioids. One advocate who worked with teenagers described how her young patients abused pills orally and recreationally before getting addicted and entering her treatment program. She also described the stigma for young users in injecting opioids intravenously. These patient advocates saw potential in ADFs to prevent the progression of abuse from oral use to snorting and injecting opioids. However, they also cautioned that individuals who are unable to abuse a particular opioid may substitute an easier-to-abuse option. ©Institute for Clinical and Economic Review, 2017 Page 20 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 3. Summary of Coverage Policies To understand the insurance landscape for abuse-deterrent formulations of opioids, we reviewed publicly available 2017 coverage policies and formularies for the six New England state Medicaid programs, Centers for Medicare and Medicaid Services (CMS), and 12 majors “Silver-level” plans on individual marketplaces across New England. We identified coverage policies for four of the nine drugs in this review, including OxyContin, Xtampza, Hysingla ER, and Embeda. Arymo ER was not covered by any plans in the review. Vantrela ER, Troxyca ER, Morphabond, RoxyBond, and Targiniq ER are not yet commercially available in the U.S. OxyContin (oxycodone) is most likely to be covered, although more than half of plans still require prior authorization. Xtampza (oxycodone) is the least likely to be covered, and is covered by fewer than one-quarter of plans reviewed. Embeda (morphine) is covered by nearly two-thirds of plans reviewed, and it is least likely to require prior authorization. Despite their different active ingredients, many plans (60%) either cover Embeda (morphine) or Hysingla (hydrocodone), but not both (see Appendix B, Table B1) – and all plans have quantity limits for all opioid therapies in this review. Table 5. Percentage of New England Commercial Plans that Cover Abuse-deterrent Formulations of Opioids and Coverage Restrictions For those plans with coverage: Covered Prior Authorization Quantity Limits OxyContin 92% 58% 100% Xtampza 23% 100% 100% Hysingla 62% 67% 100% Embeda 69% 44% 100% Coverage policies for ADFs are distinct from coverage policies for generic ER opioids. In general, commercial carriers require prior authorization for ADF ER opioids, requiring patients to try non- abuse-deterrent, generic equivalents, or preferred brands first. For example, at Neighborhood Health Plan in Massachusetts, patients can access morphine ER tablets without prior authorization but with quantity limits, while an abuse-deterrent opioid like OxyContin requires prior authorization and step edits. Still, plans vary substantially in their policies for ADF ER opioids, with some plans requiring very simple step edits through preferred therapies and others requiring very detailed risk assessment ©Institute for Clinical and Economic Review, 2017 Page 21 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents and monitoring for abuse. Harvard Pilgrim and Connecticare require step therapy with an immediate release opioid or preferred extended release opioid before authorizing coverage for Hysingla, Embeda, or Xtampza. Anthem Maine has the most extensive prior authorization documentation, closely following recent CDC guidelines. They require prescribers to demonstrate proof of querying the state prescription monitoring program database, as is required by state law, and collecting urine samples every six months for continued coverage. The other Anthem programs in New England do not have such policies. Despite utilization management of these opioid pain therapies, many plans have special allowances for patients with cancer pain. Examples of these policies are included in Table 6 and can be found in Appendix B. Medicaid In New England Medicaid programs, the majority of ADF opioid therapies are non-preferred and require prior authorization with quantity limits. Embeda, however, is a preferred therapy in half of New England state Medicaid programs. In these states, use of Embeda does not require prior authorization. Many states throughout New England adhere to strict guidelines in their prior authorization documentation. New Hampshire, for instance, requires that prescribers query the Prescription Drug Monitoring database, have a written pain agreement, demonstrate a history of addiction (alcoholism and substance use disorder), and see a pain specialist before authorizing use of any long acting opioid. Maine Medicaid requires that patients have a chronic pain management plan and revisit their prescriber in order to reauthorize their prescription. Massachusetts is perhaps the least burdensome, requiring prescribers only to demonstrate proof of intolerance and need of therapy. ©Institute for Clinical and Economic Review, 2017 Page 22 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 6. Examples of Prior Authorization Policies Connecticare Anthem Maine New Hampshire Medicaid Example 1: Embeda Example 2: Embeda Example 3: All ER Opioids Initiating Coverage Preferred Agents Exalgo, fentanyl patch Fentanyl patch (generic), fentanyl patch (generic for (Duragesic), morphine sulfate ER levorphanol, methadone, Duragesic®) Kadian® morphine tabs (MS Contin), Nucynta ER, methadose, morphine sulfate ER, sulfate SA (generic MS Contin®) oxymorphone ER (Opana ER—MD OxyContin (brand), tramadol ER oramorph SA (generic for MS must write for original (generic), oxymorphone ER, Contin®) formulation on prescription) hydromorphone ER. Step therapy Yes, must fail two preferred Individual has been maintained on agents a short-acting opioid analgesic, including opioid analgesia as Failure on two other narcotics inpatient for post-surgical pain; OR for pain treatment for which the Individual transitioning from one requested long acting narcotic is long-acting opioid analgesic to indicated another long-acting opioid analgesic Cancer and/or Not listed Requests for increased quantity can Hospice patients and end of life Palliative Care be approved for the diagnosis of patients are exempt from prior Exemption cancer related pain. authorization. Risk Assessment or Not listed Yes, including a pain treatment plan Confirmation that patient has a Agreement with treatment goals written pain agreement Pain Specialist Not listed No Patient has been referred to a pain management clinic or other clinical specialist Querying Not listed No New Hampshire Prescription Prescription Drug Monitoring Program Monitoring (PDMP) has been reviewed Program (PMP) within the last 60 days Authorization Time 60 pills (1-2 months depending on 3 months 3 months dosage) Continuing Coverage Authorization Time 6 months 6 months 6 months Cancer Exemption Yes Yes: Authorized for 1 year for Yes ongoing treatment; Lifetime for palliative treatment Risk Assessment or Not listed Yes Yes Agreement Pain Specialist Not listed No Yes Querying PMP Not listed Yes Yes Urine Drug Screen Not listed Yes Not listed ©Institute for Clinical and Economic Review, 2017 Page 23 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 4. Comparative Clinical Effectiveness 4.1 Overview To inform our analysis of the comparative effectiveness of abuse-deterrent formulations of opioids, we abstracted evidence from available clinical and observational studies, whether in published, unpublished, or abstract form. The drugs of interest are included in Table 2 above. We sought evidence on the effects of ADFs on outcomes relevant to patients, the health system, and society, as listed below. • Patient/Population Level Outcomes o Abuse Potential Endpoints ▪ VAS measures (0-100) of drug liking, take drug again, and overall drug liking ▪ Tampering o Real World Evidence of Abuse and Misuse ▪ Overdose and fatality ▪ Abuse/misuse • Physical evidence of misuse/abuse • Self-reported misuse/abuse • Route of administration for misuse/abuse ▪ Addiction • Health System Level Outcomes o Health system costs o Drug loss and diversion rates o Patterns of illegal drug use o Doctor shopping o Prescription utilization • Societal Level: Outcomes related to the criminal justice system, worker productivity, and education. ©Institute for Clinical and Economic Review, 2017 Page 24 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 4.2 Methods Study Inclusion Criteria We included evidence from randomized controlled trials (RCTs) and observational studies (e.g., surveys, database and registry studies). We did not include studies that focused exclusively on the analgesic properties of ADFs without reporting on any abuse-related endpoints. We also excluded studies that “simulated” an ADF (e.g., combining intravenous oxycodone with naltrexone to simulate the ADF form of this combination being abused by intravenous route) rather than administering the actual agent of focus for the review. In recognition of the evolving evidence base for ADFs, we supplemented our review of published studies with data from conference proceedings, regulatory documents, information submitted by manufacturers, and other grey literature that met ICER standards for review (for more information, see https://icer-review.org/methodology/icers-methods/icer-value-assessment- framework/greyliterature-policy/). We excluded abstracts that also reported data available in peer- reviewed publications. Data Sources and Searches Procedures for the systematic literature review assessing the evidence on abuse-deterrent opioids followed established methods in systematic review research.157 We conducted the review in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines except for items 15 and 22 on the checklist of 27 items158. Further detail of is available in Appendix A, Table A1. We searched MEDLINE, PsycINFO, Cochrane Database of Systematic Reviews, Cochrane Central Register of Controlled Trials, and EMBASE for relevant studies. We limited each search to English- language studies of human subjects and excluded articles indexed as guidelines, letters, editorials, narrative reviews, case reports, or news items. To supplement the above searches and ensure optimal and complete literature retrieval, we performed a manual check of the references of recent peer-reviewed publications and public reports. Further details on the search algorithms, methods for study selection, quality assessment, and data extraction and synthesis are available in Appendix A. ©Institute for Clinical and Economic Review, 2017 Page 25 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Assessment of Level of Certainty in Evidence We used the ICER Evidence Rating Matrix (see Figure 4) to evaluate the evidence for a variety of outcomes. The evidence rating reflects a joint judgment of two critical components: a) The magnitude of the difference between a therapeutic agent and its comparator in “net health benefit” – the balance between clinical benefits and risks and/or adverse effects AND b) The level of certainty in the best point estimate of net health benefit.159 Figure 4. ICER Evidence Rating Matrix ©Institute for Clinical and Economic Review, 2017 Page 26 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 4.3 Results Study Selection Our literature search identified 1,424 potentially relevant references. A total of 41 references met our inclusion criteria, of which 15 were premarket studies that evaluated abuse potential endpoints (13 publications; two conference abstracts/posters), and 26 were postmarket studies that primarily evaluated real world impact on levels of abuse and misuse (19 publications; seven conference abstracts/posters). All of the premarket studies were RCTs, while the postmarket studies were entirely observational. Premarket studies that met our inclusion criteria were identified for all ADF interventions of interest, with the exception of Targiniq ER and RoxyBond IR, for which we identified relevant data in its FDA prescribing information. Postmarket studies were only found for OxyContin. The primary reasons for study exclusion included the use of a simulated ADF or use of opioids with abuse-deterrent properties that are not labeled by the FDA as an ADF, study outcomes that focused exclusively on pain (we assumed bioequivalence of ADF and non-ADF formulation), and non-comparative study designs. Quality of Individual Studies We rated only the studies that were published in peer-reviewed journals. All 13 published premarket studies were rated to be to be of fair quality using criteria from the U.S. Preventive Services Task Force (USPSTF).160 The studies considered some, but not all important outcomes and used acceptable measurement instruments which were generally applied equally. Of 19 published postmarket studies, we rated 15 as fair quality and four as poor quality using the National Institutes of Health (NIH) Quality Assessment Tool for pre-post studies with no control group as guidance.161 We did not assign a quality rating to references that were obtained from the grey literature (e.g. conference proceedings). Overall, 83% of our study set received funding from pharmaceutical companies, while another 10% was supported by the RADARS system, an independent nonprofit postmarketing surveillance system that is supported by subscription fees from pharmaceutical manufacturers. ©Institute for Clinical and Economic Review, 2017 Page 27 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Patient/Population Level Outcomes: Studies that Evaluated Abuse Potential Premarket studies evaluated the oral and intranasal abuse potential of each of the ADFs by asking recreational drug users to rate how much they liked the drug as well as their likelihood to take the drug again. All studies found significantly less likeability for ADFs versus non-ADF opioids, although the magnitude of difference between ADFs and comparators varied. Similar trends were observed for responses to questions regarding the likelihood of taking the drug again. Of note, there is no established threshold for what constitutes a clinically-important difference in any “abuse potential” endpoint, so the clinical significance of the findings remains unclear even if statistical differences were noted. Overview We identified 16 studies that evaluated the abuse potential of ER ADFs, of which 15 were premarket studies covering all interventions of interest except Targiniq ER. We did not identify any publication or conference presentation on the premarket findings of Targiniq ER that met our inclusion criteria. However, for completeness, we included premarket findings from two evaluations of Targiniq ER presented in the FDA prescribing information as part of our results. Additionally, although not originally in our scope of review, unpublished findings from a premarket study of RoxyBond IR available in FDA prescribing information was also included. All premarket studies were randomized, double-blind, active- and placebo-controlled crossover trials. The trials were broadly divided into two categories: those that assessed oral abuse potential (see Table 7) and those that assessed intranasal abuse potential (see Table 8). Study participants were healthy, non-dependent recreational drug users between the ages of 18 and 55 years. However, one study of the intranasal abuse potential of Targiniq ER, which was identified in the FDA prescribing information, was conducted among dependent opioid users and employed a similar study design. Trial populations were predominantly male (67-90%) and Caucasian (65-90%). Participants who had a positive urine drug screen or were physically dependent on opioids, alcohol, or other drugs were excluded from all but one study measuring the intranasal abuse potential of Targiniq ER. In addition, all trials had a screening phase which consisted of a naloxone challenge test (to determine physical dependence) and a drug discrimination test (to evaluate whether the study subject could distinguish the non-ADF comparator from a placebo). Participants were excluded from the study if they failed any part of the screening phase. There was no universal comparator, but ADFs were generally compared with non-ADFs in the same class. For example, oxycodone ADFs were compared with IR oxycodone; ©Institute for Clinical and Economic Review, 2017 Page 28 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents hydrocodone ADFs were compared with IR hydrocodone; and morphine ADFs were compared with ER morphine. Key measures of abuse potential included maximum levels of “drug liking” (“at this moment, my liking for this drug is…”), which was a primary endpoint in the studies of focus, as well as secondary endpoints of “overall drug liking” (typically measured at 12 and 24 hours post-dose), and “take drug again” (“I would take this drug again” measured at 12 and 24 hours post-dose). Drug liking endpoints were measured using a bipolar 0 to 100mm Visual Analog Scale (VAS), in which 0 represents “strong disliking”, 50 represents a neutral response, and 100 represents “strong liking”. Response to whether the subject would take the study drug again was measured on a unipolar scale of 0 to 100, where 0 represents “definitely would not take drug again” and 100 represents “definitely would take drug again”. Of note, there is no established threshold for what constitutes a clinically-important difference in any of these endpoints, so the clinical significance of the findings described below remain unclear even if statistical differences were noted. Results: Premarket RCTs VAS scores of drug liking, take drug again, and overall drug liking for each of the ADFs under consideration are shown in Tables 7 (oral abuse potential studies) and 8 (intranasal abuse potential studies) below. Relative to non-ADF comparators, both crushed and intact forms of each extended- release ADF produced statistically-significantly lower scores for drug liking. Although scores were lower, the magnitude of difference varied considerably across agents. Drug liking in oral abuse potential studies ranged from a 7-point difference between crushed Arymo ER and crushed morphine sulfate ER to a 25-point difference between Hysingla ER and hydrocodone IR solution.39,40 Similarly, the incremental difference in drug liking varied across intranasal abuse potential studies, ranging from seven points (crushed Vantrela ER vs. hydrocodone powder) to 36 points (crushed Targiniq ER vs. oxycodone IR powder).41,42 Crushed versions of each ADF generally produced higher drug liking scores than intact oral versions, but both remained lower than the non-ADF comparators. Although the magnitude of difference was typically minimal, there were a few instances of notable differences (e.g., the drug liking scores for intact vs. crushed Troxyca ER were 59.3 and 74.5, respectively).44 Similar trends were observed for overall drug liking, although statistical significance was not reached in a study of the oral abuse potential of crushed Arymo ER versus crushed morphine ER and a study of crushed Troxyca ER versus crushed oxycodone IR.40,44 As with drug liking measures, all studies except one showed less likelihood to take an ADF again versus a non-ADF comparator. The only study that did not follow this pattern was a trial on the oral abuse potential of crushed Troxyca ER, for which scores of take drug again did not statistically differ from crushed oxycodone IR (see Table 7).44 ©Institute for Clinical and Economic Review, 2017 Page 29 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Similar to extended-release ADFs, crushed form of RoxyBond IR produced statistically-significantly lower scores of drug liking (magnitude of difference: 12; p<0.0001) and take drug again (magnitude of difference: 20; p<0.0001) relative to the non-ADF comparator in one intranasal abuse study.56,162 No oral abuse study was identified for RoxyBond IR. Results: Observational Study A prospective cohort study from Peacock and colleagues may offer additional context to the findings reported in the premarket studies of abuse potential.163 A total of 522 Australian individuals who regularly tampered with opioids were interviewed; investigators sought to evaluate the level of tampering of reformulated OxyContin, as well as perceived attractiveness of original versus reformulated OxyContin. Compared to original OxyContin, fewer people rated reformulated OxyContin as easy to cut-up (21% vs. 79%; p<0.05) and dissolve (14% vs. 74%; p≤0.01).163 Additionally, whereas only 5% of participants reported that tampered original OxyContin was unpleasant to use, 50% perceived reformulated OxyContin in this way (p<0.01).163 ©Institute for Clinical and Economic Review, 2017 Page 30 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 7. Premarket Studies Evaluating the Oral Abuse Potential of ADFs ADF Dose Intact & crushed ADFs & active VAS score, Emax (n) comparators¥ Drug Take drug Overall drug liking again liking Extended-release (ER) OxyContin -- No oral abuse potential study 43 Xtampza ER 40mg Xtampza ER- intact 68.8* 70.2* 69.4* (n=38) Xtampza ER- crushed 73.4* 73.7* 74.2* IR oxycodone- crushed 81.8 75.4 76.2 Troxyca ER44 60mg Troxyca ER- intact 59.3* 48.7* 53.3* (n=41) Troxyca ER- crushed 74.5* 72.5 74.3 IR oxycodone- crushed 89.8 81.5 81.8 ǂ45 Targiniq ER Targiniq ER-intact 54.7 38.5 NR (n=29) -- Targiniq ER-chewed 54.6 32.6 NR Oxycodone IR solution 77.9 61.4 NR Hysingla ER39 60mg Hysingla ER- intact 63.3† 32.6† 54.9† (n=35) Hysingla ER- crushed 69† 43† 56.8† Hydrocodone IR solution 94 86.7 84.1 46 † † Vantrela ER 45mg Vantrela ER- intact 53.9 46.4 49.2† (n=41) Vantrela ER- crushed 66.9† 58.7† 59† Hydrocodone IR 85.2 75.2 75 Embeda47 120mg Embeda- crushed 65.2† 57.7† 58.6† (n=33) Morphine sulfate ER- crushed 80.8 70.7 69.8 48 † Embeda 120mg Embeda- intact 67.6 NR NR (n=32) Embeda- crushed 68.1 † NR NR Morphine solution 89.5 NR NR Morphabond ER -- No oral abuse potential study 40 Arymo ER 60mg Arymo ER- intact 62† 56† 57† (n=38) Arymo ER- crushed 67* 61.5* 63.5 Morphine sulfate ER- crushed 74 68 67.5 Immediate-release (IR) RoxyBond IR -- No oral abuse potential study ¥: Placebo arms not included in table, non-ADF comparator arms indicated by bold font; *p≤0.05 vs. active comparator; †p≤0.001 vs. active comparator; ǂ study conducted in opioid-dependent population ©Institute for Clinical and Economic Review, 2017 Page 31 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 8. Premarket Studies Evaluating the Intranasal Abuse Potential of ADFs ADF Dose Crushed ADFs & active VAS score, Emax (n) comparators¥ Drug liking Take drug Overall again drug liking Extended-release (ER) OxyContin49 30mg OxyContin- crushed NR 64* 69.7* (n=30) Original OxyContin- crushed NR 89.6 87.4 Oxycodone IR powder NR 86.6 84.8 Xtampza ER50 40mg Xtampza ER- crushed NR† 47.8† 48.2† (n=39) Oxycodone IR- crushed NR 71.3 71.8 51 † * Troxyca ER 30mg Troxyca ER- crushed 60.5 58.9 60.2* (n=28) Oxycodone IR- crushed 92.8 88.4 85.4 Targiniq ER‡42 40mg Targiniq ER-Crushed 59.1 42.6 NR (n=23) Oxycodone IR powder 94.8 93.6 NR Hysingla ER52 60mg Hysingla ER- crushed 66.8† 34.6† NR (n=25) Hydrocodone powder 90.4 83.9 83.4 Vantrela ER41 45mg Vantrela ER- crushed 72.8* NR 68.5* (n=45) Hydrocodone powder 80.2 NR 77.1 Zohydro 83.2 NR 79.8 53 † † Embeda 30mg Embeda- crushed 69.6 60.6 60.8† (n=33) Morphine sulfate ER- crushed 87.6 84.9 83.8 54 * *‡ Morphabond ER 60mg Morphabond ER- crushed 71.1 66.4 NR† (n=25) Morphine sulfate ER- crushed 84.8 76.4 NR 55 † † Arymo ER 60mg Arymo ER- crushed 52.5 50 50.5† (n=46) Morphine sulfate ER- crushed 77.5 73 71 Immediate-release (IR) RoxyBond IR‡56 30mg RoxyBond IR - crushed 71.1† 62.2 NR (n=29) Oxycodone IR - crushed 82.9 † 82.1 NR ¥: Placebo arms not included in table, non-ADF comparator arms indicated by bold font;‡: Data from FDA label *p≤0.05 vs. active comparator; †p≤0.001 vs. active comparator ©Institute for Clinical and Economic Review, 2017 Page 32 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Patient/Population-Level Outcomes: Studies that Evaluated Real-World Evidence of Abuse and Misuse We identified 22 postmarket studies that evaluated real-world evidence on the impact of ADFs on abuse and misuse; all were non-randomized studies focusing primarily on OxyContin and comparators. Comparators were either prescription opioids (e.g. IR oxycodone, ER morphine) or illicit drugs (e.g. heroin). Some studies also compared OxyContin to other prescription opioids as a group, rather than examine individual opioids. This usually included hydrocodone, hydromorphone, morphine, oxymorphone, tramadol, tapentadol, and IR oxycodone. The majority of the studies were time series analyses that compared the time before and after the introduction of reformulated OxyContin. Data for these analyses were obtained from a variety of sources (see Table 9 for a complete list). Major outcomes examined in these studies include overdose and fatalities, abuse rates, and routes of administration for abuse/misuse. None of the studies included addiction as an outcome. ©Institute for Clinical and Economic Review, 2017 Page 33 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 9. Data Sources Used to Assess the Impact of Reformulation on OxyContin and Comparators Poison control calls or visits 1. National Poison Data System (NPDS): A poisoning surveillance system that captures 99.8% of poison exposures reported to all poison centers in the USA. 2. The Researched, Abuse, Diversion, and Addiction (RADARS) Poison Center Program: Collects intentional abuse data from participating poison control centers participating across the country. Such data consists of calls to poison centers reporting adverse drug-using experiences and usually requesting assistance. Individuals entering substance use disorder programs 3. The National Addictions Vigilance Intervention and Prevention Program (NAVIPPRO): This is a risk management surveillance program for prescription opioids which uses the Addiction Severity Index-Multimedia Version (ASI-MV) to collect data. ASI-MV is a continuous, real-time, national data stream that assesses pharmaceutical abuse by patients entering substance use disorder treatment by collecting product-specific, geographically-detailed information. 4. RADARS Opioid Treatment Program (OTP) and the Survey of Key Informants’ Patients Program (SKIP): Interviews new patients entering substance-abuse treatment about medications that they have abused. Population-based surveys 5. National Survey on Drug Use and Health (NSDUH): A survey of the civilian, non-institutionalized population aged 12 years and older; provides national estimates on the use of illicit drugs, alcohol and tobacco and nonmedical use of certain prescription drugs. It is conducted annually by the Substance Abuse and Mental Health Services Administration; uses state-based sampling design. Electronic health data and medical claims databases 6. IMS LRx database: Covers approximately 65% of all retail prescriptions filled in the U.S. and uses de- identified data with a unique ID that enables multiple prescriptions dispensed to a patient to be linked over time (over 150 million unique patients). 7. Truven MarketScan commercial database: Provides de-identified pharmacy and medical claims data for commercially insured patients throughout the US. Data on abuse is based on ICD-9-CM diagnosis code. Other data sources 8. RADARS Drug Diversion Program: Collects diversion information from municipal police departments (47%), multi-jurisdictional drug task forces (26%), county sheriff’s departments (17%), regulatory agencies such as medical and pharmacy boards (5%), and other (5%) of events related to law enforcement activities or actions related to drugs of abuse. 9. RADARS StreetRx Program: Uses a crowdsourcing website that gathers street price data for drugs using a publicly-accessible website. ©Institute for Clinical and Economic Review, 2017 Page 34 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Abuse Evidence on the impact of reformulated OxyContin on opioid abuse is mixed. The majority of time series studies found that after the abuse-deterrent formulation of OxyContin was introduced, there was a decline in the rate of OxyContin abuse, which ranged from 12% to 75% at different post-reformulation time points in different populations. However, the rate of abuse of other prescription opioids (ER oxymorphone, ER morphine, IR oxycodone) and heroin abuse may have increased during the same period. Furthermore, findings from direct interviews with recreational users showed that reformulated OxyContin may have limited impact on changing overall abuse patterns. We identified 16 studies that presented evidence on the impact of reformulated OxyContin on abuse in different populations. Most of the studies focused on the changes in the rates of abuse of OxyContin (and comparators), presented as the prevalent proportion of the study population that report or identified as abusing OxyContin and other comparator opioids during the specified time period. Examples of populations covered by these studies include patients entering substance use disorder programs (e.g., NAVIPPRO and RADARS SKIP studies), total U.S. population covered by a set of poison control centers (e.g., the RADARS poison center based studies), or commercially insured patients on OxyContin (e.g., the claims-based studies). The results of these studies are summarized below, and grouped into database/surveillance studies or self-reported outcome studies. Database/surveillance studies Changes in rates of abuse are reported in Table 10. Using the number of cases received at poison control centers (RADARS poison control data), three studies reported reductions in the population adjusted rates of OxyContin abuse at various post-reformulation time periods.57,59,60 For example, at five years post-reformulation, the population adjusted rate of OxyContin abuse was estimated to have declined by 75% (0.056 per 100,00 to 0.014 per 100,000). Concurrently, there was also a 33% decline in the estimated rates of abuse from other prescription opioids during the same period (0.387 per 100,000 to 0.260 per 100,000).57 Similarly, another study based on the National Poison Data System (NPDS) surveillance system found a significant reduction in the average number of calls received at poison control centers from OxyContin intentional abuse two years post-reformulation (pre-post change: -36%; p<0.0001).67 However, in contrast to the RADARS poison center based studies, there was a simultaneous 20% increase in the abuse of other single entity oxycodone (IR and generic ER oxycodone) (p<0.0001). Furthermore, the study found a 42% increase in heroin abuse (p<0.0001).67 Based on the RADARS SKIP survey and NAVIPPRO surveillance system, six additional studies observed a 22% to 48% decline in the prevalence of OxyContin abuse among individuals entering ©Institute for Clinical and Economic Review, 2017 Page 35 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents substance use disorder programs at various post-reformulation periods (see Table 10).58,60-64 In contrast, these studies observed a significant increase in the prevalence of abuse of other prescription opioids and heroin (see Table 10). For example, based on the RADARS SKIP, which focuses on patients with a primary diagnosis of opioid dependence, one study observed a 38% and 100% increase in the abuse of ER oxymorphone and heroin, respectively, at four years post- reformulation.62 Furthermore, a NAVIPPRO-based study observed a significant increase of 8% in the abuse of all prescription opioids (including OxyContin) among all patients assessed for substance at 1 year post-reformulation (pre-post relative risk = 1.08, p<0.0001). Two studies used the NSDUH database, which is designed to estimate the prevalence of non- medical use of drugs in the United States among individuals ages 12 years and older (see Table 10).65,66 In the first study, Jones et al. reported the prevalence of past-year OxyContin abuse from 2006 to 2013. The prevalence increased progressively from 0.5% in 2006 to 0.7% in 2010; following reformulation in 2010, the prevalence declined to 0.6% in 2011 and was at 0.5% in 2013. The authors, however, noted that the prevalence in 2013 was only significantly different from that of the reformulation year (2010), but not significantly different from that of the other years prior to reformulation (2006-2009).65 Similarly, a second study used population-adjusted rates to show that compared with 2009, the rate of past year initiation of OxyContin abuse decreased by 19%, 38%, 28%, and 51% in 2011, 2012, 2013, and 2014, respectively; statistical significance was not reported.66 We identified three additional studies that used the Truven MarketScan pharmacy and medical claims database to assess the changes in rates of diagnosed opioid abuse.68,87,164 In one study by Rossiter et al., which was conducted over a 6 months pre- and 6 months post-reformulation period, the rate of diagnosed abuse among patients primarily on reformulated OxyContin compared with patients that were primarily on any extended release opioid in the period before reformulation declined by 23% and 18% among commercially-insured patients and Medicaid patients, respectively (p<0.05). In contrast, there was a non-significant increase in the rate of diagnosed abuse among Medicare patients on post-reformulated OxyContin.87 Similarly, relative to the years prior to reformulation, Kadakia et al. found a decline in the rate of diagnosed abuse among commercially insured patients on OxyContin and a simultaneous increase in the rate of diagnosed abuse from ER morphine, ER oxymorphone and IR oxycodone at 3 years post-reformulation (see Table 10).68 A third study by Michna et al. observed that 28% of commercially insured patients originally on OxyContin (N=15,162) switched to other forms of non-ADF opioids six months post-reformulation; and also noted a significantly higher rate of diagnosed abuse among patients who switched to non- ADF ER opioids (6.7%) or IR opioids (11.3%) than those on reformulated OxyContin (3.5%) during a 15 months study period (p<0.001).164 ©Institute for Clinical and Economic Review, 2017 Page 36 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Interview/self-reported outcome studies To give context to the RADARS SKIP data, one study reported additional information from the RAPID program, which is a subset of the SKIP participants willing to give up their anonymity and participate in a follow-up interview. In interviews with 153 RAPID program participants with a history of long-term abuse of OxyContin, 33% of participants indicated that the reformulation had no effect on them and they continued to abuse OxyContin, another 33% indicated that they replaced OxyContin with other drugs as a result of the ADF, and only 3% indicated that the ADF influenced their decision to stop abusing drugs (see Figure 5).61 Out of those that changed to other drugs (N=51), 70% indicated they switched to heroin; 29% to other prescription opioids while 1 participant (2%) changed to cocaine.61 Additionally, we identified three studies conducted in Kentucky (USA), Canada, and Australia among patients with a long history of opioid abuse. All three studies found a decline in self-reported OxyContin abuse post-reformulation.69-71 However, evidence on changes in abuse patterns (positive urine drug screen or self report) of other opioids was mixed. For example, in the Australian cohort, there was no apparent increase in the self-reported levels of other pharmaceutical opioid use compared to the period prior to reformulation, although these data were limited to only three months post-reformulation, and original OxyContin was still in circulation.71 In contrast, the Kentucky study covered a one year post-reformulation period and found a significant increase in the past 30-day use of IR oxycodone following reformulation (96% vs. 74%; RR=1.3, 95%CI 1.19-1.42).69 Figure 5. Follow Up Interview with RAPID Participants (N=153), Subset of RADARS SKIP 61 Did ADF OxyContin influence the drugs that participants used for abuse? Yes, replaced OxyContin with No, continued to other drugs use OxyContin 33% 34% No, did not use Stopped abusing OxyContin enough drugs to change actions 3% 30% ©Institute for Clinical and Economic Review, 2017 Page 37 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 10. Changes in Abuse Pattern of OxyContin and Comparators Data source Timeframe compared Change in abuse pattern of OxyContin‡ % change of comparators Before After reformulation Outcome (population) % Heroin Prescription opioids (excludes reformulation change OxyContin) RADARS Poison center59 4Q08 - 3Q10 4Q10 - 1Q12 Mean quarterly rate (Cases at poison control centers) -38* NM All other opioids: NS RADARS Poison center57 3Q09 - 2Q10 1Q11 - 2Q15 Mean quarterly rate (Cases at poison control centers) -75* NM All other opioids: -33* RADARS Poison center60 3Q09 - 2Q10 1Q11- 4Q13 Mean quarterly rate (Cases at poison control centers) -55* NM All other opioids: -7* RADARS SKIP61,62 1Q09 - 2Q10 1Q11 – 2Q14 Past month prevalence (Patients with primary diagnosis of -42* +100 ER oxymorphone: +38* opioid abuse) RADARS SKIP60 3Q09 - 2Q10 1Q11- 4Q13 Past month prevalence (Patients with primary diagnosis of -30* NM All other opioids: +16* opioid abuse) RADARS SKIP63 4Q09 – 3Q10 4Q10 – 1Q12 Past month prevalence (Patients with primary diagnosis of -37 +78¥ All other opioids: +5¥ opioid abuse) NAVIPPRO58 2Q09 – 3Q10 3Q10 – 2Q12 Past month prevalence (Patients entering substance use -41* NM ER oxymorphone: +246* disorder treatment) ER morphine: NS NAVIPPRO64 1Q08 – 3Q10 3Q10 – 4Q11 Past month prevalence (Patients entering substance use -22* -11* ER oxymorphone: +191* disorder treatment) ER morphine: NS NAVIPPRO60 3Q09 - 2Q10 1Q11- 4Q13 NC -48* NM All other opioids: -3* NSDUH65 1Q09 – 4Q09 1Q13 – 4Q13 Past year prevalence (US household survey-12 years and -28¥(NS) NM -- older) NSDUH66 1Q09 – 4Q09 1Q13 – 4Q13 Past year initiation rate (US household survey-12 years and -28¥† NM -- older) NPDS67 3Q09 – 2Q10 3Q10 – 3Q12 Quarterly rates (Calls to poison control centers) -36* +42* Other single entity oxycodone +20* Claims data68 3Q09 – 3Q10 4Q10 – 4Q13 Diagnosed rate (Patients on OxyContin and comparator -35* NM ER oxymorphone: +236* opioids) ER morphine: +44* IR oxycodone: +36* Kentucky cohort69 Pre-3Q10 4Q10 – 1Q11 Past month prevalence (recreational users) -55† NM IR oxycodone: +23 Canada cohort 70 1 year prior 3Q12-4Q12 Positive urine drug screen (recreational users) -12* NM ER morphine: NS Australia cohort71 1Q14-1Q14 2Q14 – 3Q14 Past month prevalence (recreational users) -57* NM Other opioids: NS *p<0.01; † value not reported; ¥estimated; NM-not measured; NC-not clear; NS-Not significant; ‡There were some differences in the operational definition of abuse across sources. ©Institute for Clinical and Economic Review, 2017 Page 38 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Routes of Administration of Abuse Limited evidence suggests there was a reduction in both oral and non-oral abuse of OxyContin following reformulation, however, among those that continued to abuse OxyContin, there was a significant shift from non-oral routes to the oral route of abuse. We identified three studies that described changes in the route of administration of opioids following reformulation.57,58,62 As described above, the three studies reported a decline in the rate of OxyContin abuse following reformulation. However, the non-oral route of abuse declined at a significantly greater rate compared with the oral route of abuse.57,58 For example, Severtson et al. reported a 71% decline in OxyContin abuse through the oral route compared with 87% decline in the non-oral route at five years following reformulation (p=0.006).57 Furthermore, among patients that abused OxyContin before and after it was reformulated, Butler et al. found a significant increase in the reported use of oral routes of abuse from pre- to post-reformulation period (55% vs. 76%; p<0.0001), while there was a concurrent decrease in the non-oral route of abuse (see Figure 5). Notably, other comparator opioids did not show a similar pattern. Specifically, ER morphine products showed no change in the route of administration profile across study periods, while ER oxymorphone showed a significant increase from pre- to post-reformulation in snorting (62% vs. 69%; p=0.0162) and injection (9% vs. 16%; p=0.0124), and a significant reduction in the oral route (38% vs. 30%; p=0.0056).58 Similarly, a second study by Cicero et al. based on the RAPID program (N=117) found a significant decrease in the non-oral abuse of OxyContin in the post-reformulation period among patients with a history of OxyContin abuse, while there was an increase in the oral route of administration (see Figure 5).62 Figure 6. Changes in the Abuse Routes of OxyContin Among Participants Who Have Taken Pre- and Post-Reformulated Forms NAVIPPRO RAPID 100% 76% 80% 79% 80% 55% 53% 54% 60% 43% 36% 28% 40% 25% 21% 18% 16% 8% 20% 6%4% 0% Oral Snort Inject Smoke Oral Snort Inject Smoke Before reformulation After reformulation Before reformulation After reformulation ©Institute for Clinical and Economic Review, 2017 Page 39 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Overdoses and Fatalities Limited evidence suggests that rates of overdose and overdose deaths attributed to OxyContin declined after its abuse-deterrent formulation was introduced. Overdose data on other opioids do not show consistent trends across studies, although heroin overdose deaths increased during the post-reformulation period. Evidence on ADF-related overdose and overdose deaths is extremely limited. Data on the commercially-insured population from Truven MarketScan suggest that rates of OxyContin overdose/poisoning diagnoses decreased 34% from 0.42 per 100 person-years of opioid use in the year before reformulation to 0.28 per 100 person-years of opioid use in the three years following reformulation (p=0.0189); overdoses of ER morphine, ER oxymorphone, IR oxycodone, and IR hydromorphone were not statistically different after reformulation.60 Another analysis that used Optum claims data from a large commercial insurer found that overdoses due to prescription opioids decreased by 20% (from 5.48 to 4.38 per 100,000 members per quarter) during the two years following OxyContin reformulation, while the heroin overdose rate increased by 23% (from 1.15 to 1.41 per 100,000 members).72 Similarly, OxyContin-related overdose deaths appeared to decline in surveillance datasets following its reformulation. Using manufacturer-reported adverse event data, two Purdue Pharma LP- sponsored studies reported on overdose deaths.60,73,74 Depending on the period of analysis, reports of OxyContin-related overdose deaths decreased 56-65% (See Table 11 for details).60,73 By the third year after reformulation, the rate of overdose death had declined 85-87% to reach an average of 3.3 overdose deaths per quarter (vs. 26.0 overdoses/quarter in the year prior to reformulation).60,73 Changes in fatality data for comparator opioids are insufficiently reported in the identified literature to enable comparisons, however an analysis from the Wharton School and RAND Corporation estimated that each percentage point reduction of OxyContin misuse after reformulation was associated with an increase of 3.1 heroin deaths per 100,000.75 Additionally, investigators found no evidence that reformulation affected overall overdose rates across illicit and prescription drugs, suggesting that consumers substituted OxyContin for other opioids.75 ©Institute for Clinical and Economic Review, 2017 Page 40 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 11. Change in Overdose Fatalities After Reformulation of OxyContin Study Period before Period after Opioid Change in overdose Change in rate of reformulation reformulation fatality reports overdose fatality between pre- and reports, 3rd year post- post-reformulation reformulation (95% CI) (95% CI) Coplan Q3-2009 to Q2- Q1-2011 to Q4- OxyContin -65% -85% 201660 2010 2013 -83% to -27% 95% CI NR Comparator No data No data Sessler Q3-2009 to Q2- Q3-2010 to Q2- OxyContin -56% -87% 201473 2010 2013 95% CI NR -93% to -78% ER morphine No data No data (MSContin®)* *Reports of fatalities to the manufacturer of ER morphine (MSContin®) were too few to provide a statistical comparator trend; NR=not reported Health System Level Outcomes We identified six references that reported on health system level outcomes, including doctor shopping, drug diversion, and prescription opioid utilization; all were non-randomized studies focusing on OxyContin. The majority of the studies were time series analyses that compared the time before and after the introduction of reformulated OxyContin. Data for these analyses were obtained from RADARS Drug Diversion Program and IMS prescription records (see Table 9). We did not identify any studies that discussed health systems costs. Doctor Shopping Two studies reported that doctor-shopping decreased 50% after the introduction of reformulated OxyContin, while it increased 66% for ER oxymorphone and 5% for single-entity IR oxycodone. One means by which individuals may access opioids for non-therapeutic purposes is through doctor-shopping. Doctor-shopping is “the practice of engaging multiple prescribers and/or pharmacies to obtain excess drugs that can be diverted for non-medical use.”165 There is no accepted threshold to define doctor-shopping, although the two studies we identified that reported on this outcome required individuals to have overlapping prescriptions from two or more unique prescribers and at least three unique pharmacies over a six-month interval.60,165 Using data from IMS prescription records, both studies reported that doctor-shopping for OxyContin decreased 50% after reformulation, while it increased 66% for ER oxymorphone.60,165 Among comparators, changes ranged from an average 25% decrease for IR hydromorphone to a 66% increase for ER oxymorphone; doctor-shopping with IR oxycodone increased 5%. ©Institute for Clinical and Economic Review, 2017 Page 41 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents In an analysis from Chilcoat and colleagues, investigators found that limiting their analysis to a more restrictive definition of doctor-shopping to one associated with specific characteristics associated with abuse and diversion (i.e., younger age, cash payment, and high dosage strength) resulted in a greater estimated decline in doctor-shopping with OxyContin (90%) over the specified period of analysis.165,166 Drug Diversion & Prescription Opioid Utilization After the introduction of reformulated OxyContin, rates of OxyContin diversion fell. Evidence on corresponding changes in diversion rates of other prescription opioids is inconsistent. Sales of OxyContin declined over the same period, while increasing for other long-acting opioids. We identified three publications that reported on drug diversion (any intentional act that results in transferring a prescription medication from lawful to unlawful distribution or possession)60 using population-adjusted longitudinal surveillance data from the RADARS Drug Diversion Program (see Table 12).57,59,60 In the Drug Diversion Program, law enforcement officers from municipal police departments, drug task forces, county sheriff’s departments, and regulatory agencies such as medical and pharmacy boards submit quarterly data on the number of new arrests, street buys and sales involving prescription products. Drug diversion is a measure of law enforcement activity and is limited by available resources within reporting jurisdictions, local law enforcement priorities, the drugs targeted by investigators, and variations in reporting over time.22,76 In one study, the average OxyContin diversion rate declined 53% (95% CI 41% to 63%; p<0.001) per population relative to the average rate in the period before the introduction of reformulated OxyContin; the change rate was significantly greater than that observed for other prescription opioids (-6%; p<0.001), which included immediate-release oxycodone, hydrocodone, hydromorphone, morphine, and oxymorphone.59 A follow-up study by the same investigators showed that population-adjusted rates of diversion continued to decline over five years after reformulation, reaching an 89% decrease (from 1.95 per 1,000,000 in the year prior to reformulation to 0.21 per 1,000,000 at year 5 following reformulation); diversion of other opioids also decreased during this period, albeit at a significantly lower rate (from 13.4 to 9.8 per 1,000,000). 57 Another study from Coplan and co- investigators (2016) also used data from RADARS Drug Diversion Program and reported relatively consistent results (66% decrease in diversion of OxyContin by the end of 2013), although their analysis did not show any change in diversion of comparator opioids.60 Changes in OxyContin prescription sales followed a similar pattern to that of diversion rates, with sales falling 24% in the year following reformulation; statistically significant changes in the overall opioid market for extended- and immediate-release products were not detected.167 Data from a cohort of 31 million commercially-insured individuals suggest that the dispensing rate of OxyContin fell 39% over two years (from an expected 29.1 mg to 17.8 mg of morphine-equivalent dose per ©Institute for Clinical and Economic Review, 2017 Page 42 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents member per quarter) while the estimated dispensing rate for long-acting non-oxycodone formulations was 11% higher than the predicted trend (absolute change, 3.26 mg of morphine- equivalent dose per member per quarter).72 Table 12. Population-Adjusted Change in Diversion After OxyContin Reformulation Study Period before Period after OxyContin Other Population- Population- Statistical reformulation reformulation Rate of Opioids adjusted adjusted significance Diversion* Rate of change in change in Diversion* diversion of diversion of OxyContin other (95% CI) opioids (95% CI) Severtson Q4-2008 to Q4-2010 to Pre: 3.47 Pre: 28.0 -53% -6% p<0.001 201359Ω Q3-2010 Q1-2012 Post: 1.63 Post: 26.3 -63% to -41% 95% CI NR p<0.001 p=0.602 Severtson Q3-2009 to Q1-2011 to Pre: 1.95 Pre: 13.4 -89% -27% “statistically 201657¥ Q2-2010 Q2-2015 Post: 0.21 Post: 9.8 -92% to -85% -36% to -16% different” p=NR p=NR p=NR Coplan Q3-2009 to Q1-2011 to NR NR -66% +6% p<0.001 201660† Q2-2010 Q4-2013 -74% to -55% -8% to +24% p<0.001 p=0.418 Ω “Other opioids” includes immediate-release oxycodone products, hydrocodone, fentanyl, hydromorphone, morphine, oxymorphone, methadone, buprenorphine, tramadol, and tapentadol; ¥ “other opioids” includes hydrocodone, hydromorphone, morphine, oxymorphone, tramadol, tapentadol, and immediate-release oxycodone; † “other opioids” consists of all non-OxyContin Schedule II opioid analgesic tablets and capsules with the active agents of hydrocodone, hydromorphone, morphine, oxymorphone, and immediate–release oxycodone products (methadone and transdermal patches were excluded); *per 1,000,000 population Societal Level Outcomes We did not identify any studies that assessed the societal impact of ADFs, including outcomes related to the criminal justice system, worker productivity, and education. Controversies and Uncertainties The use of surrogate outcomes (VAS measures of drug liking, take drug again, etc.) in the abuse potential premarket studies for FDA approval of an ADF constitutes an important source of uncertainty concerning the effectiveness of ADFs. Considering that there is no established threshold for what constitutes a clinically-important difference in the “VAS abuse potential” endpoints assessed in these studies, interpretation of observed results remains ambiguous. In addition, there is considerable uncertainty around whether these surrogate endpoints are predictive of real-world abuse and whether the studies that evaluated them reflect how opioids are ©Institute for Clinical and Economic Review, 2017 Page 43 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents consumed in the real world. These studies used small, selected populations of non-opioid- dependent recreational drug users who received single, controlled doses of each product under investigation, which may not reflect real-world opioid use or misuse. The uncertainties surrounding the use of premarket studies as an outcome to predict real-world abuse have been stressed by the FDA, as recently as the advisory committee meeting concerning an ADF label for an IR oxycodone (RoxyBond®) in April 2017: “None of the nine products approved with abuse-deterrent labeling have actually shown, to FDA's satisfaction, postmarketing data that demonstrate reduced abuse in the real world.”168 Data from real-world evidence poses a different kind of challenge. We found no prospective studies conducted in inception cohorts that measured real-world incidence of abuse among ADF and non-ADF users. Instead, the current evidence of real-world impact is limited to time series, which are subject to potential confounding factors and other biases. For example, these analyses do not consider other interventions that may have taken place during the study period, such as expansion of prescription drug monitoring plans, implementation of Risk Evaluation and Mitigation Strategies (REMS), and provider education, among many others. In addition, time series may be subject to autocorrelation (i.e., statistical relation between pre- and post-values), which may lead to underestimation of standard errors and overestimation of intervention effects; or conversely, they may be subject to over-dispersion, defined as greater-than-expected variability in observed data based on the assumed distribution.77 Moreover, the time series we reviewed used different timeframes of analysis and different databases, often only with a short duration of follow-up. While the trends are relatively consistent, the estimates of magnitude vary and the results of the different studies cannot directly be compared. For prospective inception cohort studies, evidence on the use of clinical risk abuse stratification tools would be important to support clinical decision-making on whether ADFs should be used for any patient who gets an opioid or only those patients at a certain threshold of abuse risk. Results of a recent systematic review on this question came to the conclusion that the evidence on clinical tools for identifying patients that are at higher risk for abuse was insufficient or absent.30 Evidence on the progression from medical use to non-medical use as well as on the “natural history” of abuse and addiction is also needed. It is believed that chewing an ER opioid is an important step towards dependence and addiction, followed by intranasal and intravenous routes of abuse,27 which explains the use of certain physical or chemical barriers in the development of ADFs. However, none of the studies in the assessment included addiction as an outcome. Furthermore, the overall net benefit of introduction of ADFs into the system cannot be fully determined from the available evidence in these studies. Although limited evidence from most of the time series studies suggest a decrease in OxyContin-specific abuse and overdose following ©Institute for Clinical and Economic Review, 2017 Page 44 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents reformulation, many of the studies also found a shift towards abuse of other prescription opioids and heroin, the extent of which may not be fully captured in these studies. There may be a tipping point at which more widespread access to ADFs would show system-wide benefits; however, current evidence from one survey suggest that only about 3% of a small cohort of long-term abusers of OxyContin stopped abusing drugs as a result of reformulation, while many others continued to abuse OxyContin or switched to other forms of opioids, including heroin.61 Uncertainty also remains on the association between the introduction of ADFs and increases in the rates of heroin use or deaths. Evidence from time series studies suggest a rise in the use of heroin following OxyContin reformulation.63,67 As discussed above, one study by RAND and Wharton that explored the relationship between state variation in OxyContin misuse and heroin death found that each percentage point reduction of OxyContin misuse after reformulation was associated with an increase of 3.1 heroin deaths per 100,000.75 However, other studies have shown that rates of heroin use and overdoses began increasing prior to the introduction of ADFs.78,79 Finally, we currently do not have any real-world evidence for the other ADFs, as their entry into the US market is very recent. While postmarket studies are mandatory with FDA approval, the first postmarket studies for ADFs other than OxyContin are not scheduled for completion until 2018 and 2019, for Hysingla® ER and Embeda®, respectively.80 Summary Using the ICER Evidence Matrix, we assigned evidence ratings for each of the ADFs of focus compared to non-ADF prescription opioids. ADFs and their non-ADF counterparts are bioequivalent, producing the same analgesic benefits, and have the same profile of adverse effects when used as prescribed.24 For patients being considered for an opioid for therapeutic purposes, we judge the comparative clinical effectiveness of OxyContin to be "C+" for the risk of abuse, primarily based on the surrogate outcomes of "likability" used in premarket studies, and the evidence on the changes in the rates of abuse reported in post-market studies. Even though we have reasonably high certainty that OxyContin does not provide inferior net health benefit compared to non-ADF comparators, without stronger real-world evidence that OxyContin reduces the risk of abuse and addiction among patients, our judgment is that the evidence can only demonstrate a "comparable or better" net health benefit (C+). For all other ADFs, excluding OxyContin, we judge the evidence to be “promising but inconclusive” (P/I) for use in individual patients being considered for an opioid. Similar to OxyContin, all other ADFs demonstrate potential comparability or better results than their non-ADF counterparts based on the surrogate outcomes of “likability” in premarket studies. Furthermore, they are considered bioequivalent in producing the same analgesic benefits, and have the same adverse effects when ©Institute for Clinical and Economic Review, 2017 Page 45 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents used as prescribed. However, while many of these formulations may present advances in technology relative to OxyContin and include alternative physical or chemical barriers, agonists and antagonists, or aversive agents, there is no real-world evidence published on any of these other ADFs to demonstrate improved health outcomes or reductions in the risk of abuse. Considering the high dependence on “likability” studies, and the lack of real world evidence, our judgment is that we cannot determine the magnitude of abuse reduction at this time, leading to our “P/I” rating. We believe there can be even less certainty in a judgment on the comparative clinical effectiveness of ADFs versus non-ADF opioids if the question relates to the net health impact of introducing or substituting ADFs for non-ADFs to the broad population of individuals who use opioids for therapeutic and non-therapeutic purposes. The evidence on the impact of OxyContin reformulation shows a decrease in OxyContin-specific abuse, but also a shift in some cases toward other routes of administration, other prescription opioids, and heroin. Given the limited evidence base on this mix of positive and negative outcomes, we do not feel there is adequate evidence to discount the possibility that the balance would be net harmful overall across the entire population, especially early in the introduction of ADFs. We therefore judge there to be insufficient evidence (“I”) with which to judge the net health benefit, at the population level, of the introduction or substitution of ADFs for non-ADF opioids. Table 13. ICER Rating on the Comparative Net Health Benefit of ADF versus Non-ADF Prescription Opioids Intervention Comparator ICER Rating Individual patient prescribed an opioid for therapeutic purposes OxyContin Non-ADF Extended Release Opioid C+ All other ADFs: Non-ADF Opioid P/I Embeda® Targiniq® ER Hysingla® ER MorphaBond® Xtampza® ER Troxyca® ER Arymo® ER Vantrela® ER RoxyBond® IR Overall population, including potential non-therapeutic users ADF Non-ADF Opioid I ©Institute for Clinical and Economic Review, 2017 Page 46 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 5. Other Benefits or Disadvantages In this section of our review, we seek to provide information on other benefits or disadvantages offered by the intervention to the individual patient, the delivery system, public health or the public that would not have been considered as part of the evidence on comparative clinical effectiveness. Patients report feeling stigmatized when prescribed opioids, given their widespread and publicized potential for abuse. Some patients report that having an ADF prescription would diminish this stigma, meaning they have a prescription that purportedly cannot be abused. For physicians, ADFs could, as part of a multi-pronged strategy, allow physicians to feel comfortable treating severe pain adequately without feeling forced to limit prescriptions as they might be otherwise. Discussions about the necessary controls on opioid prescribing need to also take into account the need for chronic pain patients to have reliable access to pain medication as part of a comprehensive pain management program.81 Due to the higher costs of ADF therapy, there may be new prior authorization requirements that require clinicians’ time and have an impact on productivity and patient care. In public comments received from hospice workers, they noted that increased costs and prior authorization requirements could impact productivity at small provider practices and hospice programs, as well as their ability to adequately care for patients in need of pain management. The hospice workers also noted that out-of-pocket costs due to higher costs of the therapies could inhibit access to opioids for patients in need. Legislation and policy mandating or encouraging use of ADFs often includes other components targeted at reducing opioid abuse and misuse. However, no evidence seems to have been generated to date on the effects of these multi-component strategies, or on the importance of ADF policy relative to other components. Safety issues have been raised with abuse-deterrent technologies after tampering for intravenous use for Opana®ER (oxymorphone)15,82 and for the ADF RoxyBond®.26 The reformulation of Opana ER in 2012 with a high-molecular-weight polyethylene oxide physical and chemical barrier led to a shift from intranasal to intravenous abuse.83 An HIV and Hepatitis C virus outbreak in Indiana was caused by using the tampered product with shared needles, and the outbreak was controlled by implementing a needle exchange program.84 In Tennessee, a cluster of thrombotic microangiopathy is thought to be related to intravenous exposure of substances produced by the tampering of the polyethylene oxide barrier used as abuse-deterrent technology in Opana ER.25,82 These risks could also arise with the intravenous abuse of other ADFs that also use a polyethylene oxide barrier, such ©Institute for Clinical and Economic Review, 2017 Page 47 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents as the ADFs Arymo, Hysingla, and OxyContin. These risks could also arise with the intravenous abuse of other ADFs that use similar technologies. Finally, ADFs are currently available only for the extended-release opioid formulations that comprise around 10% of all prescription opioid use. Broader understanding of the benefits of ADF formulations are urgently needed, with the first immediate-release ADF approved by the FDA as of April 2017, but not yet available on the market. ©Institute for Clinical and Economic Review, 2017 Page 48 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 6. Cost-Benefit and Potential Budget Impact of Abuse-Deterrent Opioid Formulations 6.1 Overview We conducted analyses of the potential economic impacts of abuse-deterrent formulations (ADF) of opioids. We developed a de novo cost-benefit model to evaluate the costs and benefits of ADF opioid use, comparing a hypothetical population of chronic pain patients who were newly prescribed either extended-release (ER) ADF opioids or ER non-ADF opioids over a five-year time- horizon from the health care system perspective. Due to the varied nature of the underlying conditions leading to chronic pain and the lack of published data on utility parameters in opioid users, this model used a cost-benefit rather than a cost-utility (cost per QALY) framework. The aim of this analysis was to estimate and compare the costs and benefits of using ADF opioids or non-ADF opioids for chronic pain (e.g., reduced numbers of deaths associated with opioid abuse). Our model objective was to attempt to answer two key research questions: 1) what are the potential net costs and outcomes of using ADFs compared to non-ADFs, and 2) what levels of effectiveness in abuse reduction and in price difference would be needed for ADF opioids to achieve cost neutrality or net savings relative to non-ADF opioids? The benefits were defined in terms of the reduction in abuse-related outcomes, such as the number of incident cases of abuse, the number of opioid overdose-related deaths, and subsequent health care resource use. Importantly, this analysis did not explicitly include the costs of externalities such as diversion or switching to heroin and other non-ADF opioids that may occur in reaction to the abuse-deterrent properties of ADFs, due to lack of data directly attributing these patterns to ADF use and the focus of the model on clinical and economic impacts among the chronic pain patients themselves. We tested this as a scenario analysis using various assumed estimates for the level of diversion and the relative risk (RR) of diversion of ADF opioids. This analysis also did not compare the benefits of ADFs to other strategies to address abuse of opioids, such as non-opioid pain management strategies, prescription monitoring, or addiction treatment programs. While our primary analyses focused on hypothetical cohorts of chronic pain patients receiving ADF and non-ADF opioids, we also conducted a state-specific policy analysis that analyzed the health and economic burden associated with opioid use in the state of Massachusetts if all non-ADF ER opioid prescription users in the state were to be converted to ADF ER opioid prescriptions. ©Institute for Clinical and Economic Review, 2017 Page 49 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 6.2 Cost-Benefit Model Methods Model Structure The cost-benefit cohort model was developed in Microsoft Excel (Microsoft, Redmond, WA) and (depicted in Figure 7), and consisted of nodes corresponding to outcomes of opioid use in 100,000 non-cancer chronic pain patients newly prescribed either ER ADF or non-ADF opioids. We did not include cancer patients in the model, as there may be different considerations when determining appropriate pain management for these patients (e.g., focus on immediate-release rather than ER opioids and possibly a higher dose of opioids compared to those of non-cancer patients). Separate cohorts were assumed for patients newly starting ADF and non-ADF opioids respectively. All patients enter the model as therapeutic users, defined as those chronic pain patients who used prescription opioids for only pain-alleviating purposes and not for abuse. As a therapeutic user, a patient could discontinue opioid use due to end of treatment or die from non-abuse related causes. Patients entering the model in the first year, as well as those who continued as therapeutic users in subsequent years, had an annual probability of opioid abuse. A proportion of those who abused had an assumed annual probability of ceasing to abuse opioids after which they drop out of the model. Other patients who abuse had an annual probability of death from opioid overdose or other causes. The remainder of those who abuse continue to a subsequent year of abuse. All clinical and cost inputs used in the model are described the sub-sections below. For this analysis, each cohort was assumed to receive long-term ER opioid prescriptions, defined as those for longer than 90 days.169,170 Health care costs were assigned to the ADF and non-ADF cohorts, with patients abusing opioids assumed to have higher health care resource utilization and costs than therapeutic users. ©Institute for Clinical and Economic Review, 2017 Page 50 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Figure 7. Model Schematic Representing One Cycle for the Prescription Opioid Hypothetical Cohort* Continue therapeutic use Abuse Therapeutic use Discontinue therapeutic use Death (all cause) Prescription opioids Continue abuse Cease to abuse Death from Abuse overdose Overdose Continue abuse Death (all cause) *Patients in the ADF and non-ADF opioid cohorts follow the same pathway The model employs annual cycles over a five-year time-horizon, taking a health care system perspective. We chose a five-year time horizon because we assumed that few patients would be prescribed opioids continuously for longer than five years. Costs and outcomes were calculated annually as well as cumulatively over the five-year period, and compared for the ADF and non-ADF opioid cohorts. Costs and outcomes were not discounted because of the relatively short time horizon. This base-case evaluation was conducted from a health care system perspective, and thus focused on direct health care costs only. ©Institute for Clinical and Economic Review, 2017 Page 51 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 14. Key Assumptions Assumption Rationale Source The rates of abuse with ADF or non-ADF Lack of published evidence on the variability opioids were kept constant throughout in the rates of abuse over time. the time horizon of the model. Baseline characteristics of the ADF Patient characteristics in this claims analysis Rice et al., 201485 opioid prescription cohort (reported were similar to those seen in a national below) were assumed to be the same survey of opioid use. as those in the non-ADF cohort. We assumed the same health care Lack of published evidence showing a Commonwealth of resource utilization costs for ADF and statistically significant impact of an ADF Massachusetts non-ADF opioid therapeutic users, and opioid on health care costs. Health Policy for ADF and non-ADF opioid patients Commission91 who abused opioids The base case analysis did not include Lack of robust published evidence on effects diversion from prescription use or of ADF use on drug-switching behavior abuse. among abusers obtaining through diversion. The model did not include outcomes ADFs are considered bioequivalent to their Schaeffer, 201224 related to pain alleviation and relevant non-ADF formulation. tolerability. Base case assumed abuse-deterrent OxyContin has majority of market share IMS data on file; effectiveness of ADF OxyContin in a among ADFs, and largest real-world evidence Rossiter et al., commercially-insured population when base available. 201487 calculating the difference in health and economic outcomes between ADF and non-ADF opioids. Daily dosage for both ADF and non-ADF Reflects dosage beyond which patient CDC report89 opioids is assumed to be 90mg monitoring is recommended. morphine equivalent dose (MED), split over three doses daily. The model does not account for Model aims to analyze potential benefit of switches to other prescription opioids ADF opioids as replacement for non-ADF or use of illicit opioids such as heroin. opioids in patients with new opioid prescriptions, focusing only on effects on abuse and not that of other opioid drugs. In addition, illicit opioid use and associated costs would fall outside a health care system perspective. Rate of discontinuation of regular use Lack of published evidence that rate of of opioids was assumed to be the same discontinuation of regular opioid use for ADF and non-ADF cohorts. differed. ©Institute for Clinical and Economic Review, 2017 Page 52 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Annual rate of cessation of opioid Lack of published evidence on this rate, or on abuse was assumed to be 10% in both utilization and costs in year of cessation of cohorts. In the year of cessation, the abuse. patient was assumed to incur 50% of abuse-related health-care resource use and costs prior to dropping out of model after cessation of abuse. Only inputs derived from a commercial- Lack of complete published evidence for Rossiter et al., insured population were used Medicare and Medicaid populations. 201487 Commonwealth of Massachusetts Health Policy Commission91 In the scenario analyses, diverted abuse Published evidence suggests that SAMSHA, 2016171 in the non-ADF cohort was assumed to prescription opioid abuse contributes to Severtson et al., be 1.25, 1 and 0.75 cases for every approximately an equal number of cases of 201659 prescription abuse. The relative risk of diverted abuse. We found only one study diversion of ADF opioids was varied as reporting relative risk of diverted abuse in well. ADF opioids, specifically OxyContin. Patients who abused ADF or non-ADF Lack of published evidence on varying opioids were awarded the same societal costs between patients who abuse societal costs ADF or non-ADF opioids. Target Population The population for the base-case hypothetical cohort in the cost-benefit model included adults aged 18 years and older with chronic non-cancer pain and new prescriptions for long-term ER opioid use. Baseline characteristics of the hypothetical cohort were assumed to be similar to those reported in an observational study using administrative claims data from 2006 to 2012 85 in which two groups of patients, one with evidence of regular opioid use and the other with evidence of abuse, were matched on age, gender, presence of other non-opioid substance-abuse diagnoses, and other comorbidities. Data on age and gender from this analysis determined background mortality for this model (Table 15). We modeled two distinct cohorts, each including 100,000 patients with: 1) new ADF ER opioid prescriptions, and 2) new non-ADF ER opioid prescriptions. As mentioned previously, we did not model the effects of diversion or switching to heroin or other opioids due to the lack of good quality data on these impacts. While we assume that these would occur more frequently with prevalent use and abuse of ER opioids than with the new users modeled here, the results of this model can provide insight into whether the net economic benefit of ADFs compared to non-ADFs might balance out the cost of switching to abuse of heroin or other opioids. ©Institute for Clinical and Economic Review, 2017 Page 53 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 15. Model Cohort Characteristics Opioid abuse Regular use Primary source Mean Age 36.5 years (14.6) 37 years (16.3) Rice et al., 201485 (SD) Male 56.4% 54.7% The comparison represents a matched sample Treatment Strategies We compared FDA-approved branded ADF opioids to branded and generic non-ADF opioids. Costs for a typical ADF and non-ADF opioid were calculated as a weighted average of their market share, based on the number of incident users of these opioids in Massachusetts.8 A list of opioids and their market share within the ADF and non-ADF groups is available in Appendix G, Table G1. Opioids with ADF properties but without an FDA-approved ADF label fell into the non-ADF opioid category in our analysis. While there are several ADF opioid formulations, we used efficacy data on OxyContin in the cost-benefit hypothetical cohort model because it is the only ADF for which data on effectiveness in deterring abuse were available. For each ER opioid, we assumed a strength of 90mg morphine equivalent dose (MED) as a daily dose, split into three doses of 30mg MED, except in the case of Nucynta®, for which the split was four doses a day to reach the 90mg MED threshold. Details on the drugs included are available in Appendix G, Table G2. Model Inputs Model inputs were estimated from several sources, including observational studies and published reports. The inputs that informed our model are described below, separated into clinical and cost inputs. Clinical inputs Rate of abuse The rate of abuse for ADF and non-ADF opioids was obtained from results reported by Rossiter et al. for a commercially insured population.87 We used data on the rate of abuse prior to OxyContin reformulation to simulate abuse in the non-ADF cohort, and data following reformulation to estimate abuse in the ADF cohort. Abuse was defined based on the ICD-9 diagnosis codes for opioid abuse, dependence and poisoning, as seen in Appendix G Table G3. All inputs can be found in Table 16. ©Institute for Clinical and Economic Review, 2017 Page 54 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Opioid discontinuation Opioid discontinuation in therapeutic users ranged from 17.2% in year one to 40.4% in year five after initiating ER opioid use, based on a claims analysis by Martin et al. using data from a national commercial health care network from January 2000 to December 2005.86 Patients with opioid prescriptions were followed overly nearly five years, with discontinuation defined as patients without an opioid prescription refill over six months. Discontinuation was assumed to be the same for therapeutic users in both the ADF and non-ADF cohorts (Table 16). The other reason for discontinuation of regular opioid use was all-cause mortality. Mortality The model accounts for mortality from opioid overdose (Table 16) as well as all-cause mortality (Appendix G, Table G4). The opioid overdose mortality was assumed to attributed to abuse-related overdose and not accidental overdose in therapeutic users. The risk of mortality from opioid overdose was assumed to be the same for patients with abuse in both the ADF and non-ADF cohorts. The background all-cause mortality matches the cohort’s age and sex characteristics and was obtained from the Social Security Administration’s actuarial life tables.88 The background all- cause mortality matches the cohort’s age and sex characteristics and was obtained from the Social Security Administration’s actuarial life tables.88 Table 16. Clinical Inputs Input Value Source Incidence of non-ADF ER opioid 3.647% Rossiter et al., 201487 abuse Incidence of ADF ER opioid abuse 2.818% Rossiter et al., 201487 (OxyContin) Annual percentage of Year 1 – 17.8% Martin et al., 201186 discontinuation of prescription Year 2 – 28.4% opioid use Year 3 -- 34.6% Year 4 – 38.2% Year 5 – 40.4% Death from opioid overdose 5.9/100,000 Compton et al., 201679 Costs All costs were calculated annually and included both drug and non-drug costs. All costs were inflated to 2016 dollars using the medical care component of the US Consumer Price Index.92 Drug costs We could not calculate net prices for all drugs using our standard source (SSR Health, LLC), as this source relies on publicly-disclosed net sales figures for branded drugs from publicly-traded ©Institute for Clinical and Economic Review, 2017 Page 55 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents companies and several of the opioids in this review were either generic or brands of privately- owned drug manufacturers. We therefore used data from the Federal Supply Schedule (FSS) to calculate discounted prices of all opioids.90 The FSS supports the acquisition of pharmaceutical drugs, medical equipment, and supplies and service contracts for the VA and other federal organizations. We weighted ADF and non-ADF prices by market share, based on IMS data on incident use of prescription ER opioids from February 2016 to January 2017 in Massachusetts. When there was more than one price for the same drug, as in the availability of multiple generics of the same non-ADF formulation, an average price per dose was calculated. Health care costs Health care costs were assigned to the ADF and non-ADF cohorts, with patients abusing opioids having higher health care resource utilization and costs than therapeutic users. These costs included costs of emergency room visits, inpatient and outpatient visits, and associated professional fees. Costs were obtained from a claims study conducted by the Commonwealth of Massachusetts Health Policy Commission specifically for this report.91 This claims analysis used the 2014 Massachusetts All-Payers Claim Database (APCD) that consists of commercial medical claims, pharmacy claims and personal spending for the three largest payers (Blue Cross Blue Shield, Harvard Pilgrim Health Care and Tufts Health Plan) in the state. The population included was patients in the data set who had an opioid prescription of 90 days or more. Of the 3,199 patients included in the study sample, 176 and 374 patients had diagnosis of abuse with ADF and non-ADF opioids, respectively, while 861 and 1,788 patients were therapeutic users of ADF and non-ADF opioids, respectively. We took a weighted average cost of health care resources used, weighted by patient sample size as mentioned above, to arrive at the mean annual resource utilization and non- opioid prescription costs for therapeutic users and patients who abuse opioids (Table 17). More details on the methodology for this analysis can be found in Appendix G. ©Institute for Clinical and Economic Review, 2017 Page 56 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table 17. Cost Inputs Input Value Source ADF Opioids – 90mg MED Cost per daily dose* $11.60 FSS, 201790 Annual cost $4,234 Calculation Non-ADF Opioids – 90mg MED Cost per daily dose* $5.82 FSS, 201790 Annual cost $2,124 Calculation Mean Annual Health Care Costs Mean annual costs Therapeutic use patients Abuse patients Health care resource utilization $19,285 $31,005 Commonwealth of Costs Massachusetts Health Non-opioid prescriptions $8,404 $7,140 Policy Commission91 *Market-share based weighted average cost of drugs within each category. Drugs are listed in Appendix table G1. Sensitivity Analyses Threshold Analysis Given the limited data on the effectiveness of newer ADF opioids, we conducted threshold analyses, varying the ADF effectiveness (by varying rate of abuse) to determine reductions in the annual rate of abuse that would attain cost neutrality for ADFs relative to non-ADF opioid use. We undertook a similar cost-neutrality analysis by varying the costs of ADF opioids relative to non-ADF opioids. One-Way Sensitivity Analysis One-way sensitivity analyses for key inputs used 95% confidence intervals or ranges based on plausible values from the published literature when available; where not available, input parameters were varied by +/- 25%. We conducted one-way sensitivity analyses, varying model parameters on the incidence of abuse, the efficacy of ADF opioids, the cessation of abuse, and drug costs. Scenario Analysis Diversion While opioid diversion and switching play a critical role in ascertaining the health and economic impact of the opioid abuse epidemic, we did not include these effects in our base-case analysis due to a lack of robust evidence. However, we conducted a scenario analysis to test for cost-neutrality between the ADF and non-ADF opioid cohorts by introducing different assumed rates of diversion into the model, based on data published by the Substance Abuse and Mental Health Services ©Institute for Clinical and Economic Review, 2017 Page 57 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Administration (SAMHSA) that indicated that there are approximately 1.25 cases of diverted opioid abuse for every case of prescription opioid abuse.93 Using this as a reference point for the non-ADF opioid cohort, we estimated the reduction in relative risk of diversion in the ADF opioid cohort that would achieve cost-neutrality between the two cohorts. Modified Societal Perspective The impact of opioid abuse expands beyond costs of the health care system. To account for this, we included a modified societal perspective in a scenario analysis, including the costs of criminal justice and incarceration, as well as costs of productivity loss due to opioid abuse as reported by Birnbaum at al.94 Birnbaum et al. derived health care and societal costs using data from a claims analysis that included privately insured patients, where they calculated the per patient opioid abuse-related health care and productivity loss costs cost as well as the associated caregiver costs. Criminal justice and incarceration costs were calculated using data from the Criminal Justice Expenditures and Employment Extract Program. Total societal costs were then calculated by attributing these costs per abuse patient to the number of abuse patients in the 2007 National Survey on Drug Use and Health (NSDUH). We assumed the same societal costs for patients who abused ADF or non-ADF opioids. Base Case Results Health outcomes of our base case analysis over a five-year time horizon are presented in Table 18. The results indicate that the ADF opioid cohort had approximately 2,300 fewer new cases of abuse and approximately 6,600 fewer abuse-years compared to the non-ADF opioid cohort, with a small reduction in opioid overdose-related deaths of fewer than one. Table 18. Abuse-Related Outcomes for ADF and Non-ADF Opioid Cohorts of 100,000 Chronic Pain Patients with ER Opioid Prescriptions Outcome (at 5 years) ADF cohort Non-ADF cohort Increment (ADF cohort – Non-ADF cohort) New case of abuse 8,229 10,532 -2,303 Person-years of abuse 23,322 29,943 -6,621 Overdose deaths 1.38 1.77 -0.39 Table 19 shows results for the total healthcare costs of the two cohorts, the total prescription opioid costs, and the incremental differences between the two cohorts. Even with the cost-offsets within the health care system from having fewer patients abusing opioids, use of ADF opioids resulted in an additional $533 million net spending over five years from ©Institute for Clinical and Economic Review, 2017 Page 58 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents the health care system perspective (Table 19). The lower abuse-related costs of ADF opioids compared to non-ADF opioids were outweighed by the higher prescription costs of ADF opioids. Table 19. Total Estimated Health-Care Costs of Patients Prescribed ADF and Non-ADF Opioids over Five Years ADF opioids Non-ADF opioids Difference (ADF – non-ADF) Therapeutic use* $7,845,606,246 $7,692,466,543 $153,139,703 Abuse* $939,121,323 $1,205,748,255 -$266,626,932 Prescription opioid costs (entire $1,303,908,313 $657,301,870 $646,606,443 cohort) Total $10,088,635,882 $9,555,516,668 $533,119,214 *Excludes prescription opioid costs. Includes health care resource utilization and non-opioid prescription costs Using ADF opioids resulted in additional costs of $231,500 for preventing one new case of abuse and approximately $80,500 for preventing one abuse-year. Given the small benefit observed in overdose deaths, the cost to prevent an overdose death was estimated to be approximately $1.36 billion (Table 20). Table 20. Cost Per Incremental Outcome of ADF Opioid Versus Non-ADF Opioid Incremental outcome Cost To prevent one new abuse case $231,514 To prevent one new abuse year $80,517 To prevent one overdose death $1,362,339,569 Sensitivity Analysis Results Threshold analysis We increased the effectiveness of ADF opioids in reducing abuse (i.e., decreased the rate of abuse in the ADF opioid cohort) to identify the point at which cost-neutrality with respect to total health system costs between the two cohorts would be achieved. Results from this analysis indicated that increasing the effectiveness of ADF opioids to the point where they fully eliminate abuse (where the rate of abuse is 0) still resulted in additional costs of approximately $113 million over five years. (Figure 8). ©Institute for Clinical and Economic Review, 2017 Page 59 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Figure 8. Incremental Health System Cost of ADFs at Increasing Levels of Effectiveness (Decreasing Incidence of Abuse) $600 $533 Incremental burden to health system $500 $430 $400 (In Millions) $326 $300 $220 $200 $113 $100 $0 2.818%* 2.11% 1.41% 0.70% 0.00% Incidence of abuse with ADF opioids *Represents base case In this scenario, when ADF opioids were assumed to have 100% effectiveness in preventing abuse, the cost per (a) new abuse case prevented was approximately $10,700, (b) abuse-year prevented was approximately $3,800, and (c) opioid overdose death prevented was approximately $63.7 million (Table 21). Table 21. Cost per Incremental Outcome of ADF Opioid Versus Non-ADF Opioid when ADF Effectiveness in Preventing Abuse Is Assumed to be 100% Incremental outcome Cost Preventing one new abuse case $10,712 Preventing one new abuse year $3,768 Preventing one overdose death $63,749,147 Cost-neutrality was achieved when the ADF opioid-weighted market share price was discounted by 41%, from $11.60 to $6.86 per day (at 90 mg MED per day), keeping the base case incidence of abuse in each cohort constant. This discount required to achieve cost-neutrality represents the discount from a market-share weighted average price of ADFs, and does not represent the discount required by any individual ADFs in the market. ©Institute for Clinical and Economic Review, 2017 Page 60 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents One-Way Sensitivity Analyses Detailed findings of the one-way sensitivity analyses can be found in Figure 9 and Table 22. Results were most sensitive to uncertainty to costs of ADF opioids followed by uncertainty in rate of cessation of abuse. Varying the parameters within plausible ranges did achieve cost-neutrality between the two cohorts. Figure 9. Tornado Diagram for Overall Health Care Cost-Difference Between ADF and Non-ADF Opioids Base case cost difference is $533,19,214. Table 22. Tornado Diagram Inputs and Results Parameters Low High Low Result High Result Input Input ADF opioid costs (+/- 25%) $8.70 $14.50 $207,142,136 $859,096,292 Rate of abuse cessation (0% to 0% 20% $479,886,976 $576,406,603 20%) Efficacy of ADF opioids (95% CI) 0.0251 0.0313 $488,072,826 $577,714,346 Rate of abuse (+/- 25%)* 0.0274 0.0456 $558,338,210 $508,699,399 *While the rate of abuse was varied, the percentage difference in this rate of abuse between ADF and non-ADF opioids was kept constant, at 22.7%. ©Institute for Clinical and Economic Review, 2017 Page 61 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Scenario Analyses Diversion We included diversion as a scenario analysis. We tested the level of reduction in relative risk of diversion of ADF opioids that would be needed to attain cost-neutrality relative to non-ADF opioid use. Data on rates of opioid diversion and abuse indicates that there are approximately 1.25 cases of diverted abuse for every case of prescription opioid abuse.171 We conducted this analysis at three different estimates of diversion of opioids: 1.25, 1.0, and 0.75 cases of diverted abuse for every one case of prescription non-ADF opioid abuse. The cases of diverted abuse were added to the cases of prescription abuse in each cohort. Assuming 1.25 cases of diverted opioid abuse for every case of prescription opioid abuse, the risk of diversion of ADF opioids would need to be 35% lower compared to that with non-ADF opioids to achieve cost- neutrality between the ADF and non-ADF opioid cohorts. Similarly, assuming 1.0 and 0.75 diversion cases per abuse case would require reductions of 44% and 59% in the risk of diversion of ADF opioids, respectively, to achieve cost-neutrality (Figure 10). The incremental total health care system costs associated with ADF opioids at different levels of diversion is available in Appendix G, Table G5. The results of this analysis must be viewed with two important considerations in mind. First, we have examined only the impact of reductions in the risk of diversion of the opioid used in the cohort initially. It is recognized that ADF formulations may deter diversion of that formulation but also increase abuse of other opioids and heroin. Second, the costs of diversion are not in fact incurred by the cohorts in our analysis (chronic pain patients); while some of these costs may represent real costs to the health system, this is entirely dependent on the extent to which diversion occurs with first-degree relatives of chronic pain patients or others covered by the same health system. ©Institute for Clinical and Economic Review, 2017 Page 62 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Figure 10. Incremental Costs of Diversion and Percentage Decrease in ADF Opioid Diversion Required to Achieve Cost-Neutrality $600 Incremental health system cost $500 $400 (In Millions) $300 $200 $100 35% 44% 59% $0 0% 10% 20% 30% 40% 50% 60% 70% Decrease in diversion with ADFs 1.25 cases of diverted abuse* 1 case of diverted abuse* 0.75 cases of diverted abuse* *For every case of prescription abuse with non-ADF opioids Modified Societal Perspective We included the costs of lost productivity and criminal justice and incarceration for those who abused ADF and non-ADF opioids. The societal costs of each case of abuse were estimated to be approximately $3,400 annually for criminal justice and incarceration, and approximately $16,600 annually for lost productivity.92,94,97 Including these societal costs, the difference in total net spending between the ADF and non-ADF cohorts over five years was reduced, but still represented an increase of $393 million in the ADF cohort. A breakdown of total costs within each cohort, including societal costs, is shown in Table 23. Table 23. Total Estimated Societal Costs of Patients Who Abuse Prescription ADF and Non-ADF Opioids over Five Years ADF opioids Non-ADF opioids Difference (ADF – non-ADF) Societal costs $492,445,032 $632,255,624 -$139,810,592 Total costs (Health system + $10,581,080,914 $10,187,772,292 $393,308,622 Societal costs)* *Includes therapeutic users’ costs ©Institute for Clinical and Economic Review, 2017 Page 63 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 6.3 Prior Published Evidence, Model Validation To the best of our knowledge, ours is the first model analyzing and reporting the health and economic outcomes of ADF and non-ADF opioids in both regular users and opioid abusers. This approach can provide a more realistic estimate of the overall burden of ADFs and non-ADFs to relevant policy stakeholders. Our model’s internal validity was assessed by stress-testing the model through variations in inputs across a wide range of estimates. In addition, we reviewed other published, ADF-related economic models to assess external validity. Rossiter et al. studied the incidence of abuse and costs associated with the introduction of reformulated OxyContin among ER opioid users.87 Their findings on the rate of abuse pre- and post- reformulation have been used in our model, and have been described in the methods section. For a commercially-insured population, the authors calculated an excess medical cost associated with a case of diagnosed abuse at $9,456 versus a control (that included patients without an opioid prescription), and an excess cost of $7,565 for an undiagnosed abuse case compared to a control. They applied rate of abuse (as seen in NSDUH data) and calculated costs to the US population in 2011, assuming a 1:5 ratio of diagnosed to undiagnosed abuse, and reported annual savings of $430 million associated with the use of ADF OxyContin rather than original oxycodone. In contrast, our study indicates a cost burden associated with all ADF opioids in the market. There are key methodological differences between the two models that lead to these different results. First, Rossiter et al. did not include opioid prescription drug costs incurred for patients, either those with regular use or those with abuse, so savings calculations were based on other healthcare services alone. Our model has indicated that this is a key driver of costs in the model, and outweighs the savings in other healthcare services to a substantial extent. Rossiter et al. justify excluding prescription drug costs by citing a lack of statistically significant difference in prescription costs between patients with abuse and regular use; however, they did not consider the cost difference between non-ADF opioids and OxyContin. Second, the Rossiter et al. study includes an abuse cohort alone and not a cohort of new users of opioids (regular use and incident abuse) as in the ICER model. Third, the Rossiter study attempted to account for diversion and switching to other opioids and heroin for abuse by assuming the reduction in incidence of abuse with OxyContin was only 75% of the reduction observed in their claims analysis. Fourth, the cost-savings reported by Rossiter et al. include commercial, Medicare and Medicaid populations, while the ICER model accounts for only the commercially-insured population. Finally, Rossiter et al. report findings at a single year while the ICER study projects results over a five-year period. White et al. developed a budget impact model that reported annual savings ranging from $0.6 billion to $1.6 billion from use of a theoretical ADF opioid in the US population, with the amount saved dependent on the uptake rate of ADF opioids.172 The theoretical ADF opioid was assumed to ©Institute for Clinical and Economic Review, 2017 Page 64 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents have therapeutic properties and the clinical efficacy of ADF ER oxycodone HCl. The number of cases of opioid abuse was derived from a claims database as well as the 2005 NSDUH survey. The authors reported total annual health care costs of $11 billion associated with prescription opioid abuse. While both the ICER model and the White et al. model report health care resources and costs avoided with ADF opioid use, the two models examine different populations. The ICER model projects abuse incidence in a new patient cohort, while White et al. model calculate abuse numbers by applying abuse prevalence derived from their database analysis to the US population. Perhaps most importantly, the White et al. model assumes the cost of ADF opioids to be the same as branded ER oxycodone while the ICER model uses a market-basket price for ADF and non-ADF opioids, with a substantial difference observed in daily costs. Winegarden, in an issue brief, calculated the net benefit of opioids using estimates from studies by Rossiter et al. and Kirson et al. 173,87,174 The cost savings per patient treated with ADF opioids was calculated by multiplying the reduction in rate of diagnosed abuse as reported by Rossiter et al. with the additional costs per abuse case from a health system perspective as reported by Rossiter et al. and from a societal (non-health care expenses) perspective as reported by Kirson et al. The two cost-savings results were then summed to arrive at a net cost-saving (benefit) per patient, at $4,645. The final net benefit per opioid patient on ADFs ranged from $1,834 to $4,033, depending on the additional costs of ADF opioids (least to most expensive opioids). From the resources Winegarden used to calculate the cost benefit of ADF opioids, it can be inferred that the author considered a US population cohort, unlike the ICER model, which employed a hypothetical cohort of new patients. Finally, we also reviewed a model developed by the Canadian Health Policy institute on the societal costs savings associated with the introduction of ADF opioids in Canada.175 We have not summarized the methods or results of this study since findings from this study cannot be translated or compared to those done from a US perspective although this study extrapolates clinical and cost estimates from US studies to a Canadian population. In summary, one of the major differences between the ICER model and other models are the populations that enter the models, with ours being a hypothetical cohort of new opioid users while other models used a US prevalent cohort. In addition, these models examine health care resource use and economic burden associated with opioid abuse cases but do not associate health care resource use and economic burden with regular use ADF and non-ADF opioids when calculating a potential benefit with ADF opioids. These differences lead to markedly different conclusions, with ADF opioids found to be cost-saving in these earlier models while leading to an additional cost burden in the ICER model. ©Institute for Clinical and Economic Review, 2017 Page 65 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 6.4 State-Specific Policy Analysis for Massachusetts We also developed a state-specific analysis as an extension of the cost-benefit cohort model, examining the health outcomes and economic impact of converting an existing prevalent population of non-cancer chronic pain patients on ER opioids from current ADF and non-ADF prescribing patterns to using entirely ADF opioids in Massachusetts over one year. We used data on the patterns of prescription opioid use in Massachusetts. Importantly, this model takes a prevalence approach, using data on both existing and new opioid users to inform state-specific findings. Methods The state-specific policy model uses the same general model structure as the cost-benefit cohort model (Figure 7). Methodological differences from the cohort model are described below. Model changes We replaced the hypothetical cohort population of 100,000 with the actual number of prevalent cases of prescription ER opioid use in each state. The model calculates health outcomes and costs over one year using 2014 claims data for a population of prescription ER opioid users who have been prescribed opioids for non-cancer pain. This model also employs a state-specific rate for deaths from opioid overdose (Appendix G, Table G6).96 Model assumptions 1. We included only non-cancer pain related ER opioid users by applying the ratio of state- specific cancer to non-cancer incident opioid use to the prevalent ER opioid use population.8,95,176 2. The proportion of prevalent opioid use that was ER was estimated by assuming the percentage of prescription opioid patients was equivalent to the percentage of ER prescription opioid fills, as reported in a 2012 IMS report.8 3. We assumed the market share for prevalent opioid use to be the same as that seen in the incident population, as we did not have data on this market share for the prevalent population. 4. Since we obtained opioid costs in Massachusetts directly from claims data, we did not have to calculate the average opioid costs using a 90mg MED per day rule in this case. The number of patients estimated to be on ADF and non-ADF opioids is shown in Appendix G, Table G7, along with data on drug market share and rate of death from overdose for Massachusetts ©Institute for Clinical and Economic Review, 2017 Page 66 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents (Appendix G, Tables G1 and G6). Mean daily cost for ADF and non-ADF opioids in Massachusetts were $15.90 and $3.44 respectively. The mean daily cost of opioids was obtained from a claims analysis undertaken by the Commonwealth of Massachusetts Health Policy Commission (HPC), specifically for this report.91 The sample of patients in HPC’s claims database was limited to 2014 claims data for those with commercial insurance through Blue Cross Blue Shield MA, Harvard Pilgrim, and Tufts Health Plan. The mean daily cost of opioids was obtained from a claims analysis undertaken by the Commonwealth of Massachusetts Health Policy Commission (HPC), specifically for this report.91 The sample of patients in HPC’s claims database was limited to 2014 claims data for those with commercial insurance through Blue Cross Blue Shield MA, Harvard Pilgrim Health Care, and Tufts Health Plan. Results We estimated a total of approximately 173,000 prevalent users of prescription ER opioids in Massachusetts, using 2015 data, of which approximately 60,000 were prescribed ADF opioids and approximately 113,000 prescribed non-ADF opioids. If all prescription opioid users in the state were prescribed ADF opioids, there would be approximately 850 fewer cases of abuse, at an estimated cost of approximately $599,000 to prevent one abuse case. in one year while prescription opioid costs would increase an additional $513 million. While abuse-related costs would decline (from $225 million to $204 million), prescription opioid costs would more than double, leading to an increase in costs statewide of $475 million annually. (Table 24). Table 24. Outcomes When Converting All Non-Cancer Chronic Pain Patients with Prescription ER Non-ADF Opioids to ADF Opioids in Massachusetts in One Year Mixed ADF/non-ADF All ADF opioid use Difference opioid use Abuse cases 5,229 4,387 -842 Prescription opioid costs $489,925,522 $1,002,689,521 $512,763,999 Abuse-related costs* $224,828,862 $203,548,318 -$21,280,544 Total healthcare costs $5,331,764,758 $5,806,899,717 $475,134,959 Cost to prevent one new $599,131 case of abuse using ADF opioids *Combination of prescription (opioid and non-opioid) and resource utilization costs ©Institute for Clinical and Economic Review, 2017 Page 67 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 6.5 Summary and Comment We analyzed the cost-benefit of ADF opioids compared to non-ADF opioids in a hypothetical cohort model of non-cancer chronic pain patients, as well as a state-specific policy model for Massachusetts. In the hypothetical cohort cost-benefit model, use of ADF opioids was estimated to prevent 2,300 new cases of abuse per 100,000 patients treated over five years, but to cost the health care system an additional $533 million over that time span. We estimated that using ADF opioids costs the health care system an additional $231,500 to prevent one new case of abuse and approximately $80,500 in additional health system costs to prevent one year of abuse. Health care cost neutrality could not be achieved even when the effectiveness of ADF opioids in preventing abuse was increased to 100%, with ADF opioids still incurring an additional cost of $113 million over five years. However, cost neutrality could be achieved if ADF opioids were discounted by 41% from the current market-basket price. We also conducted this analysis using a modified societal perspective which included estimates of productivity loss and criminal justice and incarceration costs. In this analysis, use of ADF opioids was estimated to cost approximately $393 million more than non-ADF opioids over five years. Our state policy model, focused on Massachusetts, estimated that converting all existing non-ADF opioid prescriptions to ADF prescriptions over one year would prevent approximately 850 new cases of abuse, at a cost of $599,000 for every new case of abuse prevented. The incremental overall health system costs of converting all non-ADF to ADF prescriptions over a year total to approximately $475 million. Limitations Our model has several limitations. 1) Our model assumed a static rate of opioid abuse that does not change over time. We found no published evidence on rates of abuse over time and so our model may under- or over-estimate the actual burden of abuse over five years. Owing to a lack of any published evidence on the directional change in rate of abuse over time, we did not modify this in a sensitivity analysis to test for its impact on the model outcomes. We have assumed this static rate of abuse to be the incident rate of abuse in our model. 2) We assumed death from overdose to occur only in the abuse population and not in the therapeutic use population, which therefore excludes any risk of accidental overdose. In addition, our model only accounts for overdose death as an event, and does not include the rate of overdose generally due to a paucity of available data. One can assume, however, that a significant proportion of utilization of emergency department and inpatient hospital services is attributed to opioid overdose. 3) costs and health care resources utilized by therapeutic users and those who abuse opioids do not change over time in our model. We found one study that reported variations in health care cost for patients with opioid abuse in ©Institute for Clinical and Economic Review, 2017 Page 68 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents the six months prior to and 18 months after abuse diagnosis, but did not find similar estimates for costs over a longer time-frame.98 Varying these costs over time would impact the over-all cost to the health care system, depending on the direction and magnitude of this cost variation over time.4) Our source for annual rates of ER opioid discontinuation was based on data for both IR and ER opioids. There were more IR opioid users compared to ER opioid users in that study. This, coupled with the longer duration of ER prescriptions, would indicate lower discontinuation rates for therapeutic use in our model. 5) Although there are ADF opioids with more advanced technologies and perhaps greater potential in reducing abuse are now on the market, we used effectiveness data from an OxyContin study, owing to lack of abuse-related effectiveness data for other ADF opioids. In addition, our primary model does not include diversion to a population outside the existing cohort. 6) Our analysis potentially underestimates the costs of resource utilization for patients who abuse, as it only includes cost data within the study period, and because not all abuse-related treatment is covered by health insurance and would be captured in claims data. Perhaps most importantly, our primary model analyses do not include diversion to a population outside the existing cohort. One might argue that such diversion represents a true cost to the health system, but so are the costs of switching to other opioids or heroin among individuals frustrated by ADF properties, which are also not included in this model due to a lack of robust data. We have conducted a scenario analysis examining different assumed levels and relative risks of diversion of ADF and non-ADF opioids, but these again focus only on the reduced costs associated with preventing diversion of the medication used to treat chronic pain in the cohort, and do not account for any increased use of legal or illicit opioids. Finally, in our state-specific analysis, we applied the efficacy of ADF and non-ADF opioids seen in a commercially insured population to the entire state-specific population using ER prescription opioids, owing to a lack of data split by commercial and non-commercial opioid prescription users. We expect the balance between the positive and negative effects of ADF opioids beyond the prescribed population will be a prominent element in the Policy Roundtable discussion to be held as part of the public meeting at which this report will be deliberated. Conclusion Our economic modeling analyses indicate that ADF opioids have the potential to substantially reduce the incidence of abuse in opioid-prescribed chronic pain patients relative to non-ADF opioids, but at significantly higher costs to the health care system. Even when important societal costs are included, ADF opioids were still estimated to increase overall costs. While our cost-benefit analysis reflects the current opioid landscape, this landscape is bound to change with the passage of new federal and state legislation, new ADF opioids entering the market, and the changing dynamics of opioid prescribing. The advent of new ADF opioids with potentially superior abuse-deterrent ©Institute for Clinical and Economic Review, 2017 Page 69 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents properties, as well as the lack of robust evidence on opioid diversion and switching to other opioids or heroin, call for further research that will generate real-world evidence to understand the true health and economic impact of ADF opioids on the opioid abuse epidemic. ©Institute for Clinical and Economic Review, 2017 Page 70 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 7. Summary of the Votes and Considerations for Policy 7.1 About the New England CEPAC Process During New England CEPAC public meetings, the New England CEPAC Panel deliberates and votes on key questions related to the systematic review of the clinical evidence. In this meeting, the CEPAC also voted on two policy questions to better evaluate the application of the clinical evidence and cost-benefit analysis for ADF opioids versus non-ADF opioids. The New England CEPAC members are not pre-selected based on the topic being addressed and are intentionally selected to represent a range of expertise and diverse perspectives. Acknowledging that any judgment of evidence is strengthened by real-life clinical and patient perspectives, subject matter experts are recruited for each meeting topic and provide input to New England CEPAC Panel members before the meeting to help clarify their understanding of the different interventions being analyzed in the evidence review. Subject matter clinical experts serve as a resource to the New England CEPAC Panel during their deliberation, and help to shape recommendations on ways the evidence can apply to policy and practice. At each meeting, after the New England CEPAC Panel votes, a Policy Roundtable discussion is held with the New England CEPAC Panel, clinical experts, and representatives from payers, manufacturers, patient groups, and in some cases, public policymakers. The goal of this discussion is to bring stakeholders together to apply the evidence to guide patient education, clinical practice, and coverage and public policies. Participants on Policy Roundtables are selected for their expertise on the specific meeting topic and are different for each meeting. At the July 20, 2017 meeting, the New England CEPAC Panel discussed issues regarding the application of the available evidence to help patients, clinicians, and payers address important questions related to the use of ADF opioids in reducing the risk of abuse in patients and the net health benefit across the broader population of people who may obtain opioids through both prescriptions and diversion. Following the evidence presentation and public comments (public comments from the meeting can be accessed here, starting at 1:25:00), the New England CEPAC Panel voted on key questions concerning whether ADFs reduce the risk of abuse in comparison to non-ADF opioids. The first three questions focused on the risk of abuse for the individual patient who is prescribed an opioid. The fourth question sought to elucidate the net health benefit for the broader population of patients who obtain opioids through both legitimate prescriptions for ©Institute for Clinical and Economic Review, 2017 Page 71 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents therapeutic use and abuse/diversion. In the final two questions, the CEPAC was asked to consider three policy applications for managing the introduction of ADF opioids. These questions are developed by the ICER research team for each assessment, with public comment, to ensure that the questions are framed to address the issues that are most important in applying the evidence to support clinical practice, medical policy decisions, and patient decision- making. The voting results are presented below, along with comments reflecting considerations mentioned by New England CEPAC Panel members during the voting process. 7.2 Clinical Effectiveness Voting Results 1. For a patient being considered for a prescription of an immediate release opioid, is the evidence adequate to demonstrate a reduced risk of abuse for patients using RoxyBond versus non-ADF immediate release opioids? Yes: 2 No: 10 Comments: In their deliberation, the New England CEPAC discussed the nature of pre-market studies, including how trial participants obtain the opioid for abuse, route of administration measured, and characteristics of the study participants. In the case of Roxybond, the only FDA- approved immediate-release ADF, study participants included recreational drug users who were given pre-crushed versions of both the ADF and the non-ADF opioid, in order to snort, and rated ‘likability’ and ‘take drug again’ on a 1-100 visual analog scale. Those members of the CEPAC who voted that the evidence was not adequate to demonstrate reduced risk of abuse acknowledged that while data from pre-market trials captured outcomes related to nasal abuse, there was not data on the effects on oral abuse, which is the most common form of abuse for immediate release oxycodone; in addition, there were no real-world studies examining the impact of this ADF on rates of abuse and/or diversion. Members of the panel who voted that the evidence was adequate suggested that despite the lack of real world evidence, the pre- market likability studies were sufficient to demonstrate an improvement over existing non-ADF opioids. ©Institute for Clinical and Economic Review, 2017 Page 72 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 2. For a patient being considered for a prescription of an extended release opioid, is the evidence adequate to demonstrate a reduced risk of abuse for patients using OxyContin versus non-ADF extended release opioids? Yes: 9 No: 3 Comments: The New England CEPAC recognized the more substantial body of evidence that evaluated real world abuse of OxyContin pre- and post- reformulation. Much of the discussion centered on the quality of “interrupted time-series” designs in available observational studies, including if the study authors were able to control for any confounding factors that may have influenced the results. While the panel was concerned with the lack of a randomized trial, the majority of the panel recognized that the observational studies gave consistent directional evidence to indicate a reduced risk of abuse for OxyContin in comparison to non-ADF extended release opioids. 3. For a patient being considered for a prescription of an extended release opioid, is the evidence adequate to demonstrate a reduced risk of abuse for patients using all other ADFs, excluding OxyContin, versus non-ADF extended release opioids? Yes: 1 No: 11 Comments: The panel deliberated on whether the real-world evidence from the OxyContin studies could apply to the other ADF extended release opioids. Clinical experts suggested that non- OxyContin extended release ADFs had at least as much laboratory abuse deterrence as OxyContin, but acknowledge that there is no way to compare data in the real world, since each ADF contains distinct abuse deterrent properties. In the end, the panel voted 11-1 that the evidence was not adequate to demonstrate a reduced risk of abuse for patients using all other ADF extended release opioids, excluding OxyContin, in comparison to non-ADF extended release opioids. 4. For a patient being considered for a prescription of an extended release opioid, is the evidence adequate to demonstrate a net health benefit on a population level with the use of OxyContin versus non-ADF extended release opioids? Yes: 2 No: 10 Comments: For this question, the panel had to consider whether the evidence demonstrated a net health benefit on a population level for OxyContin versus non-ADF ©Institute for Clinical and Economic Review, 2017 Page 73 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents extended release opioids, instead of on an individual patient level. In their deliberation, the panel discussed the evidence on diversion and potential switching to other opioids. One panel member highlighted the Cicero 2015 study that demonstrated that upon entrance of the abuse-deterrent OxyContin, those surveyed either kept on abusing OxyContin or switched to another drug for abuse. She recalled that only 3% of those surveyed stopped abuse altogether. Another panel member responded that those surveyed in the Cicero study were entering treatment programs because they were already addicted—he believed the evidence was sufficient to demonstrate that OxyContin prevented people from getting addicted to opioids in the first place. A majority of the panel voted that the evidence was not sufficient to demonstrate a net health benefit on a population level. 7.3 Policy Options Voting Results 5. Clinicians and policymakers are making efforts to reduce the numbers of patients started on opioids, limit the time course and refills for opioid prescriptions, and enhance monitoring for potential diversion and misuse of opioids. In addition, ADF-substitution policies are being considered to shift opioid prescriptions toward abuse-deterrent formulations. Considering the broad potential impact of substitution policies on patients, diversion, and illicit opioid use, which of the following policies do you believe would produce the most overall health benefit? a. Develop a way for physicians to work with academics and payers and policymakers to determine a way to target ADFs to high risk individuals and families. b. Allow physicians to determine whether to shift current patients to ADF opioids; require all new opioid prescriptions to be written for an ADF opioid. c. Require all current non-ADF prescriptions to be substituted with ADF and all new prescriptions to be written for an ADF opioid. A: 10 B: 2 C: 0 Comments: In this discussion, CEPAC members were asked to not consider cost as part of their vote. The question aimed to capture the net health benefit of policy interventions isolated from the budget impact of ADF opioids. This discussion came at the end of the Policy Roundtable with representatives from patient, payer, manufacturer, clinician, and government groups weighing in with their perspectives. The CEPAC discussed in depth the impact of introducing ADFs on increasing abuse of other opioids, both prescribed and illicit, including fentanyl and ©Institute for Clinical and Economic Review, 2017 Page 74 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents carfentanil, etc. CEPAC members discussed how to best balance the introduction of ADFs with shifts in rates of abuse, and some panel members expressed skepticism that any of these policy options would create meaningful impact on the population net health benefit, including overall opioid abuse. Several CEPAC members coalesced in their opinions that doctors need to collaborate with payers to determine what at-risk patient populations should be targeted with these therapies. However, one CEPAC member reflected that perhaps the decision should be made by public health administrators and not by physicians, so that society wrestles with escalating costs of ADFs as a public health issue, rather than placing the burden on individual physicians and payers. While asked to consider the impact of ADFs regardless of their budget impact, many CEPAC members found it difficult to dissociate spending on ADFs from the natural “opportunity costs” of such spending – namely, budget allotments for substance use disorder treatment options. Ten panel members voted for option A and two panel members voted for option B. 6. Clinicians and policymakers are making efforts to reduce the numbers of patients started on opioids, limit the time course and refills for opioid prescriptions, and enhance monitoring for potential diversion and misuse of opioids. In addition, ADF-substitution policies are being considered to shift opioid prescriptions toward abuse-deterrent formulations. Considering the broad potential impact of substitution policies on patients, diversion, and illicit opioid use, which of the following policies do you believe would be the best for policymakers to pursue? a. Develop a way for physicians to work with academics and payers and policymakers to determine a way to target ADFs to high risk individuals and families. b. Allow physicians to determine whether to shift current patients to ADF opioids; require all new opioid prescriptions to be written for an ADF opioid. c. Require all current non-ADF prescriptions to be substituted with ADF and all new prescriptions to be written for an ADF opioid. A: 12 B: 0 C: 0 Comments: This conversation was a continuation of the discussion above, but considered current costs and budget impact of ADFs. CEPAC members voted based on earlier discussions of the ICER cost-benefit model that projected an increase in spending of $533 million per 100,000 patients shifted to ADFs, as well as cost projections from the policy roundtable participants from the Veterans Administration and Prime Therapeutics. CEPAC ©Institute for Clinical and Economic Review, 2017 Page 75 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents members considered the health system tradeoffs of shifting all patients with an ER opioid to an ADF ER opioid, including the possibility of reducing other critical areas of addiction treatment. Panel members voted unanimously for option A. ©Institute for Clinical and Economic Review, 2017 Page 76 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 7.4 Roundtable Discussion and Key Policy Implications Following its deliberation on the evidence, the New England CEPAC Panel engaged in a moderated discussion about the use of ADFs for preventing abuse, diversion, and overdose death with a Policy Roundtable that included one patient representative, two clinical experts, two payer representatives, one state policy representative and one representative from a coalition of innovators and manufacturers. The Policy Roundtable discussion with the New England CEPAC Panel reflected multiple perspectives and opinions, and therefore, none of the recommendations below should be taken as a consensus view held by all participants. The Roundtable discussion was facilitated by Dr. Steven Pearson, MD, MSc, President of ICER, and the names of the Policy Roundtable participants are shown below. Table 25. Policy Roundtable Participants Policy Roundtable Marty Boldin, LICSW, MLADC, LCS Nathaniel Katz, MD, MS Governor’s Policy Advisor for Prevention, Treatment, & Recovery Chief Executive Officer Office of the Governor Christopher T. Sununu Analgesic Solutions Dan Cohen Jeanmarie Perrone, MD Chair Director of the Division of Medical Toxicology Abuse Deterrent Coalition Professor of Emergency Medicine University of Pennsylvania Patrick Gleason, PharmD, FAMCP, FCCP, BCPS Shaina Smith Senior Director, Health Outcomes Director of State Advocacy & Alliance Development Prime Therapeutics U.S. Pain Foundation Inc. C. Bernie Good MD, MPH Chair, Medical Advisory Panel for Pharmacy Benefits Management Veterans Administration Many of the Roundtable themes revolved around the challenges of balancing the introduction of ADFs, which are currently available only for ER opioids, with the resulting shift in abuse towards other prescription opioids and illicit opioids, such as heroin and fentanyl, and in defining the appropriate role of ADFs as part of a multifaceted strategy for combating the opioid abuse crisis. Policy recommendations are presented below. 1. ADFs should have a growing role in clinical practice since we believe they are safer for the individual patient. Still, policymakers should be fully aware that there is some evidence that demonstrates their impact on shifts in abuse towards other illicit drugs following introduction of ADFs. Mandatory ADF substitution laws may cause more harm to the overall population by shifting abuse towards other, potentially more lethal opioids and ©Institute for Clinical and Economic Review, 2017 Page 77 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents heroin. Policymakers and clinical leaders should therefore consider measures that would phase in ADFs while ensuring adequate support for other arms of a multi-pronged approach to the opioid crisis. ADFs are formulated to reduce the abuse potential for a recreational drug user who is abusing the drug through a specific route. ADFs may reduce abuse of specific drugs through either nasal or intravenous routes of abuse, but may also increase the risk of oral abuse, or switching abuse to more dangerous forms of illicit opioids such as heroin and fentanyl. Furthermore, evidence is lacking to demonstrate the overall public health impact of substituting ADFs for non ADFs. Rapid moves to shift all prescribed opioids into ADF forms may precipitate greater demand for access to treatments for substance-use disorder due to these substitution patterns. Members of the New England CEPAC did not vote to mandate substitution of ADFs because of this potential for shifts in abuse towards dangerous substitutes that can be more lethal, such as heroin or other illicit opioids; any potential shift in abuse could possibly overwhelm current treatment facilities that treat substance-use disorder, including Medication Assisted Therapy (MAT) programs. As highlighted in ICER’s 2014 report (https://icer- review.org/material/opioid-dependence-final-report/), while MAT programs are effective in treating opioid dependence, there are severe limitations on access due to regulatory, insurer, and social barriers. Manufacturers should demonstrate to lawmakers how ADFs might impact overall public health, including shifting patterns in abuse. Because of this broader population consideration, any introduction of ADFs should take a measured approach that is closely studied and monitored over a longer time frame. 2. In addition to uncertainty regarding the overall health effects of rapid substitution of ADF opioids for non-ADF opioids, at current price differentials between ADFs and non-ADFs any rapid requirement for substitution with ADFs would prove unaffordable. Policymakers should therefore avoid approaches to encouraging the use of ADFs that would be so costly that resources would be drained from other efforts needed to address the opioid crisis and drive up costs for patients and the health system at an unsustainable rate. When considering costs, the New England CEPAC unanimously voted against mandating a wholesale shift to ADF opioids. ADFs have an unsustainable budget impact to both public and private payers and will require coverage tradeoffs that may disadvantage patient care. ICER’s state policy model demonstrated that requiring all extended release opioids to be ©Institute for Clinical and Economic Review, 2017 Page 78 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents substituted for an ADF in Massachusetts would cost an additional $475 million annually. Since extended release opioids are only 10% of overall opioid prescriptions, the introduction of immediate release ADF opioids has the potential to have an even more unsustainable impact on health system budgets. One payer on the policy roundtable estimated that, for his 15 million members, converting all IR opioids to ADF IR opioids would add an additional $2.5 billion to their total pharmacy budget, or $13.57 per member per month (PMPM). One study estimated that new, curative, Hepatitis C drugs, if given to all candidate Medicare patients, would raise Part D premiums by ~$12PMPM.177 At current costs, mandating sweeping requirements that public and private payers cover ADFs is unsustainable. The New England CEPAC agreed that the best immediate step forward was for prescribers to collaborate with payers and patients to determine what patients and households are at highest risk for abuse and diversion, and target ADFs to those patients. If policymakers pursue a mandatory substitution strategy, a staged approach to implementation would allow physicians the opportunity to develop tools for identifying high risk patients to target with ADFs, and payers the opportunity to develop strategies to absorb the higher cost of moving to ADFs over generic opioids. 3. As part of their responsibility to address the national opioid crisis, manufacturers and payers must recognize a shared commitment to making ADFs affordable to patients and to the health system. Manufacturers of ADFs should moderate the exercise of their monopoly pricing power; and payers should accept that paying a premium for ADFs is reasonable and that barriers such as increased out of pocket payments should not be placed in the way of helping appropriate patients receive reasonably-priced ADFs. ICER’s analysis showed that ADF manufacturers can price their products at an 18% price premium over generic opioids while considering the reduction in costs related to the resulting decrease in abuse due to tamper-resistant formulations. This pricing strategy would communicate a commitment from manufacturers of trying to help solve a national public health crisis, while also allowing fewer restrictions on patients needing long-term opioids for chronic pain management. If manufacturers reduce their prices to align with value, payers could create mirrored coverage policies and cost-sharing structures for ADFs and non-ADFs, further improving patient access. Members on the roundtable agreed that if payers and manufacturers worked together to negotiate prices for ADFs that reflected this true health system value, payers should be willing to cover ADF opioids and eliminate any difference in coverage policies or cost- sharing between ADF opioids and non-ADF opioids. This would both improve access and ©Institute for Clinical and Economic Review, 2017 Page 79 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents affordability for patients, minimize disruptive prior authorization protocols for prescribers, and have the potential to reduce overall levels of abuse. 4. The federal government should act with urgency to convene clinical experts, clinical pharmacists, patients, and payers to develop consistent methods to stratify the risk for abuse and diversion of opioids. As universal substitution of ADFs for all patients may not be advisable or feasible, these methods should be disseminated and used to help determine when patients and their environments present a high enough risk to warrant prescription with a tamper-resistant ADF. Given that the costs of rapid switching to ADFs are not viewed as affordable, it is important to develop validated tools for risk stratification to help identify patients who are most at risk and to monitor patients in at-risk environments who should receive an ADF opioid. While there are several tools for prescribers to identify patients and families with the highest risk of abuse, no single tool is widely used or recommended. Payers are acting separately to flag high risk members based on their treatment history in claims analyses. Clinical pharmacists are rarely part of the care team when a patient is prescribed opioids for relieving pain. And patients often can speak best to the risk of abuse based on family medical history or the risk of diversion in their own household. Providers, payers, pharmacists and patients are acting in silos in making risk determinations, and there is no national guidance for how to best collaborate to identify patients and their families who are most at risk for abuse and diversion. The Federal government needs to make a rapid, concerted effort to convene prescribing clinicians, pharmacists, payers, and patients to develop universal recommendations and validated tools that can be used by physicians and clinical pharmacists to identify patients and households at greatest risk for abuse and diversion. Including patients in the development of this tool will ensure that pain patients are not stigmatized by the use of the risk tool. The Federal government needs to issue comprehensive guidance for all prescribing clinicians, pharmacists, and payers to stratify patient risk, including recommendations on which patient populations should be prescribed an ADF opioid versus a non-ADF opioid. For those who are most at risk, ADFs can act as a valuable tool to possibly reduce abuse and diversion in at-risk households. 5. The term “abuse-deterrent formulation” is confusing for prescribers, patients, and the public, and lends to misunderstanding about the risks for addiction and misuse of ADFs. It ©Institute for Clinical and Economic Review, 2017 Page 80 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents should be abandoned as quickly as possible. The FDA should reconsider whether it can use “tamper-resistant formulation” (TRF) instead, because many clinical experts believe this is the most accurate and useful term. If this is not possible, the FDA should explore other labeling options; however, clinicians, payers, and policymakers face no barriers to using different terms and therefore should adopt “tamper resistant” as the preferred term for describing these formulations. Using the term abuse deterrent formulation is often misunderstood to mean the formulation prevents abuse. All the FDA approved ADFs do not effectively deter abuse by swallowing, which is the most common forms of opioid abuse. Instead, ADFs make crushing to snort, or dissolving to inject more difficult and less rewarding. Furthermore, several participating clinicians expressed concern at the meeting that prescribing abuse-deterrent opioids may be perceived as a strategy for reducing the risk of opioid addiction. Results of a survey among primary care physicians corroborated this concern, demonstrating that nearly one-half of those surveyed erroneously reported that ADFs were less addictive than the corresponding non-ADF opioids.178 ADFs are not less addictive than non-ADF opioids—and it is critical that prescribers understand that while ADFs are tamper-resistant to prevent certain routes of abuse (like nasal or intravenous), they are still potentially addictive for any patients who are prescribed these formulations. ADF opioids can also still be addictive for members of the household who might divert opioids for recreational abuse. It is therefore important that federal policymakers include language about tamper-resistance when describing the abuse-deterrent properties of opioids. 6. Opioids represent the greatest public health crisis this country has faced in many years. Public health policymakers at the federal level should educate the public about the risks of all opioids – ADFs and non-ADFs – through a major public health campaign, perhaps modeled on the techniques employed nationally to reduce smoking. Public health campaigns to reduce smoking influenced the labeling on cigarette packaging, the creation of smoking cessation programs, and even entered the physician’s office, where patients and physicians discussed the risks of smoking. Cigarette smokers were often reminded of the risks of smoking in widescale advertising campaigns. Opioid use and abuse need to be treated similarly. Patients deserve education about the risks of opioid abuse and diversion through widespread advertising, changes to packaging, and clear conversations about risk in the physician’s office. To fund this campaign, the federal government could either allocate their own resources or consider an assessment on revenues from opioids from each opioid manufacturer into an educational trust fund. While ADF opioids may have ©Institute for Clinical and Economic Review, 2017 Page 81 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents the potential to reduce the risk of abuse through tampering and diversion, they are still addictive. It is important that the public understands the risks of all opioids, including ADFs. 7. Medical school curricula and physician licensing exams should require physicians to demonstrate a robust understanding of the role of ADFs in clinical practice, specifically addressing misconceptions about the addictive nature of ADFs. Specific questions on ADFs should be added to the already required training on opioids. Given the misunderstanding of many physicians regarding the addictive nature of ADF versus non-ADF opioids, it is critical that physicians are trained to understand what abuse- deterrent properties mean in terms of addiction risk and diversion. Furthermore, prescribers need also to be trained in evaluating patients and their households for risk of abuse and diversion. It is important that prescribers are required to understand these key concepts through mandatory training and licensure requirements. 8. Prescribing physicians should help patients understand that ADFs are not less addictive than non-ADFs. Physician groups, individual physicians, and clinical pharmacists should develop or share federally-developed materials on the proper storage and use of all opioids. Patients deserve information from their prescribing physicians about the limitations of tamper-resistant ADF opioids in reducing the risk of addiction in themselves or a family member, and the safety measures necessary to ensure that opioids are not diverted within their household. Prescribing clinicians and pharmacists can play a role in educating patients about the abuse-potential of tamper-resistant ADF opioids, how to safely store and monitor use of ADF opioids, and how to identify early signs of addiction both in themselves and in family members. 9. Robust clinical studies are needed to demonstrate the natural history of opioid abuse and the impact of ADFs on abuse among prescribed patients as well as the broader effects on diversion and drug switching. Of primary importance is a clear, scientific answer on how ADFs impact rates of abuse, evaluated through randomized clinical trials and/or rigorous, prospectively-defined observational studies. Secondarily, the body of knowledge will be improved if studies can answer important questions about drug switching and diversion, and their impact on public health. Clinical experts at the meeting maintained that these studies are practical and ©Institute for Clinical and Economic Review, 2017 Page 82 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents feasible. Manufacturers have a responsibility to the broader public health to make these trials a top research and development priority. The body of real world evidence to measure abuse of ADFs is limited to observational time- series studies of OxyContin. The studies do not follow specific patients or cohorts of patients to evaluate their risk for abuse or diversion, and none of the real-world studies are randomized. Furthermore, it is most important to understand the impact of ADFs on overall rates use of abuse, of both the prescribed product and other (often illicit) opioids. In order to best evaluate the impact of ADFs on the risk of a patient for abusing or diverting an opioid, manufacturers need to fund well-controlled studies that follow patients to determine how ADFs reduce the risk of abuse, abuse-related outcomes, and diversion within a cohort of patients. Importantly, the FDA specifically recommended such designs at a special advisory committee meeting following the approval of reformulated OxyContin.179 10. Given that over 90% of opioid prescriptions are for immediate-release (IR) formulations, and that currently, no IR ADFs are on the market, further investment and development by manufacturers for IR ADFs is critical. Since the first tamper-resistant formulation was approved in 2010, the vast majority of innovation has occurred in extended release tamper-resistant ADFs. This is because extended release opioids have the dangerous potential for dose-dumping, or extracting a long-acting dose of opioids for immediate abuse. Still, extended release opioids represent under 10% of all opioid prescriptions, and many people get addicted to opioids through immediate release forms. Currently, only one ADF IR opioid has been approved by the FDA, and it is not yet available on the market as of July 2017. Manufacturers need to prioritize investment in IR opioids, and be sure to capture their impact through high quality post- market evaluation. 11. Exploring and removing barriers to access to non-pharmacologic treatments for pain patients will have the dual effect of ensuring access to treatment for patients while also addressing the public health concerns related to prescribing opioids. Patients who are managing chronic pain need to have access to all treatments supported by evidence. Payers have a role to play in meeting the broader public health objective of less opioid addiction by loosening restrictions on access to evidence-based, non-pharmacologic treatments for pain such as physical therapy and cognitive-behavioral therapy. ©Institute for Clinical and Economic Review, 2017 Page 83 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents References 1. Califf RM, Woodcock J, Ostroff S. A Proactive Response to Prescription Opioid Abuse. N Engl J Med. 2016;374(15):1480-1485. 2. Centers for Disease Control and Prevention. Opioid Data Analysis. 2017; https://www.cdc.gov/drugoverdose/data/analysis.html. Accessed 2017-03-27. 3. Centers for Disease Control and Prevention. Policy Impact: Prescription Painkiller Overdoses. 2011; https://www.cdc.gov/drugoverdose/pdf/policyimpact-prescriptionpainkillerod- a.pdf#page=5, references at https://www.cdc.gov/drugoverdose/pdf/policy_impact_rx_painkiller_overdoses_references- a.pdf. Accessed 2017-03-21. 4. U.S. Food and Drug Administration (FDA). Timeline of Selected FDA Activities and Significant Events Addressing Opioid Misuse and Abuse. 2017; https://www.fda.gov/Drugs/DrugSafety/InformationbyDrugClass/ucm338566.htm. Accessed 2017-04-30. 5. U.S. Food and Drug Administration (FDA). Abuse-Deterrent Opioids — Evaluation and Labeling: Guidance for Industry. In:2015: http://www.fda.gov/downloads/drugs/guidancecomplianceregulatoryinformation/guidances/uc m334743.pdf. Accessed 2016-09-03. 6. McNaughton EC, Coplan PM, Black RA, Weber SE, Chilcoat HD, Butler SF. Monitoring of internet forums to evaluate reactions to the introduction of reformulated OxyContin to deter abuse. J Med Internet Res. 2014;16(5):e119. 7. U.S. Food & Drug Administration (FDA). FDA Approves New Formulation for Oxycontin. FDA News Release 2010; https://wayback.archive- it.org/7993/20161024033421/http:/www.fda.gov/NewsEvents/Newsroom/PressAnnouncement s/ucm207480.htm. Accessed 6/20, 2017. 8. IMS Health. Data on file. 2017. 9. Murray S, Wooltorton E. Alcohol-associated rapid release of a long-acting opioid. CMAJ : Canadian Medical Association journal = journal de l'Association medicale canadienne. 2005;173(7):756. 10. Gregory TB. Hydromorphone: evolving to meet the challenges of today's health care environment. Clin Ther. 2013;35(12):2007-2027. 11. Sloan P. Review of oral oxymorphone in the management of pain. Ther Clin Risk Manag. 2008;4(4):777-787. 12. Governale L. Outpatient Prescription Opioid Utilization in the U.S., Years 2000 – 2009. 2010. https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndLifeSupportDrugsAdvisoryCommittee/UCM220950.pdf. Accessed 2017-03-24. 13. Throckmorton D. FDA Perspective on Abuse-Deterrent Opioid Development. CBI Abuse Deterrent Formulations Summit, March 7-8, 2017. 2017. https://www.fda.gov/downloads/AboutFDA/CentersOffices/OfficeofMedicalProductsandTobacc o/CDER/UCM545923.pdf. Accessed 2017-03-15. 14. BioMedTracker. 2017; http://www.biomedtracker.com. Accessed 06/26, 2017. 15. U.S. Food and Drug Administration (FDA). Presentations for the March 13-14, 2017 Joint Meeting of the Drug Safety and Risk Management Advisory Committee and the Anesthetic and ©Institute for Clinical and Economic Review, 2017 Page 84 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Analgesic Drug Products Advisory Committee. In:2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM547235.pdf#page=20. 16. Drugs.com. Generic OxyContin Availability. 2017; https://www.drugs.com/availability/generic- oxycontin.html. 17. U.S. Food and Drug Administration (FDA). Abbreviated New Drug Application (ANDA): Generics. 2017; https://www.fda.gov/drugs/developmentapprovalprocess/howdrugsaredevelopedandapproved /approvalapplications/abbreviatednewdrugapplicationandagenerics/default.htm. Accessed 2017-04-13. 18. Manolis C, Bernie Good, B.C. and Shrank, W. Mandating Coverage Of Abuse-Deterrent Opioids Would Be A Costly Distraction From More Effective Solutions. HealthAffairsBlog. 2017. http://healthaffairs.org/blog/2017/05/26/mandating-coverage-of-abuse-deterrent-opioids- would-be-a-costly-distraction-from-more-effective-solutions/ Accessed 2016/06/27. 19. Manolis C, Bernie Good, and William Shrank. Mandating Coverage Of Abuse-Deterrent Opioids Would Be A Costly Distraction From More Effective Solutions. Health Affairs Blog. 2016. 20. Becker WC, Fiellin DA. Abuse-Deterrent Opioid Formulations - Putting the Potential Benefits into Perspective. N Engl J Med. 2017;376(22):2103-2105. 21. Fleming TR, Powers JH. Biomarkers and surrogate endpoints in clinical trials. Stat Med. 2012;31(25):2973-2984. 22. U.S. Food and Drug Administration (FDA). Data and Methods for Evaluating the Impact of Opioid Formulations with Properties Designed to Deter Abuse in the Postmarket Setting. In: Issues Paper. U.S. Department of Health & Human Services; 2017: https://www.fda.gov/downloads/Drugs/NewsEvents/UCM562743.pdf. Accessed 2017-06-14. 23. U.S. Food & Drug Administration (FDA). Statement from FDA Commissioner Scott Gottlieb, M.D. – FDA is taking new steps to help assess opioid drugs with abuse-deterrent properties. FDA Statement 2017; https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm562961.htm. Accessed 6/19, 2017. 24. Schaeffer T. Abuse-deterrent formulations, an evolving technology against the abuse and misuse of opioid analgesics. J Med Toxicol. 2012;8(4):400-407. 25. Lammle B. Opana ER-induced thrombotic microangiopathy. Blood. 2017;129(7):808-809. 26. Cipriano M. RoxyBond Gets US Advisory Panel OK For Abuse-Deterrent Claim Despite Excipient Concerns. 2017. https://pink.pharmamedtechbi.com/PS120383/RoxyBond-Gets-US-Advisory- Panel-OK-For-AbuseDeterrent-Claim-Despite-Excipient-Concerns. Accessed 2017-05-02. 27. Dart R. Public Health Need for Abuse-Deterrent ER Morphine. Egalet Presentations for the August 4, 2016 Joint Meeting of the Anesthetic and Analgesic Drug Products Advisory Committee and the Drug Safety and Risk Management Advisory Committee. 2010. 28. Butler SF. Reply to commentary. J Pain. 2013;14(4):361-362. 29. California Health Benefits Review Program. Analysis of California Assembly Bill AB 623 Abuse- Deterrent Opioid Analgesics. In:2015: http://chbrp.ucop.edu/index.php?action=read&bill_id=181&doc_type=3. Accessed 2016-09-03. 30. Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain - United States, 2016. MMWR Recomm Rep. 2016;65(1):1-49. https://www.cdc.gov/mmwr/volumes/65/rr/rr6501e1.htm#T1_down. Accessed 2017-01-22. ©Institute for Clinical and Economic Review, 2017 Page 85 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 31. Cicero TJ, Ellis MS, Kasper ZA. Psychoactive substance use prior to the development of iatrogenic opioid abuse: A descriptive analysis of treatment-seeking opioid abusers. Addict Behav. 2017;65:242-244. 32. Manchikanti L. National drug control policy and prescription drug abuse: facts and fallacies. Pain Physician. 2007;10(3):399-424. 33. Tetrault JM, Butner JL. Non-Medical Prescription Opioid Use and Prescription Opioid Use Disorder: A Review. The Yale journal of biology and medicine. 2015;88(3):227-233. 34. Smith SM, Dart RC, Katz NP, et al. Classification and definition of misuse, abuse, and related events in clinical trials: ACTTION systematic review and recommendations. Pain. 2013;154(11):2287-2296. 35. Dasgupta N, Freifeld C, Brownstein JS, et al. Crowdsourcing black market prices for prescription opioids. J Med Internet Res. 2013;15(8):e178. 36. Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain--United States, 2016. Jama. 2016;315(15):1624-1645. 37. Katz MH. Opioid Prescribing for Chronic Pain: Not for the Faint of Heart. JAMA Intern Med. 2016;176(5):599-601. 38. Haegerich TM, Paulozzi LJ, Manns BJ, Jones CM. What we know, and don't know, about the impact of state policy and systems-level interventions on prescription drug overdose. Drug and alcohol dependence. 2014;145:34-47. 39. Harris SC, Cipriano A, Kapil RP, et al. Oral Abuse Potential, Pharmacokinetics, and Safety of Once- Daily, Single-Entity, Extended-Release Hydrocodone (HYD) in Recreational Opioid Users. Pain Med. 2016. 40. Smith MD, Webster LR, Lawler J, Lindhardt K, Dayno JM. Human Abuse Potential of an Abuse- Deterrent (AD), Extended-Release (ER) Morphine Product Candidate (Morphine-ADER Injection- Molded Tablets) versus Extended-Release Morphine Administered Orally in Nondependent Recreational Opioid Users. Pain Med. 2016. 41. Bond M, Schoedel K, Rabinovich-Guilatt L, et al. Evaluation of the relative intranasal abuse potential of a hydrocodone extended-release tablet formulated with abuse-deterrence technology in nondependent, recreational opioid users. Journal of Pain. 2015;16(4):S82. 42. US Food and Drug Administration (FDA). Targiniq ER (oxycodone hydrochloride and naloxone hydrochloride extended-release tablets) Package Insert. https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/205777lbl.pdf. Accessed March, 2017. 43. Kopecky EA, Fleming AB, Levy-Cooperman N, O'Connor M, M. Sellers E. Oral Human Abuse Potential of Oxycodone DETERx® (Xtampza® ER). Journal of Clinical Pharmacology. 2016. 44. Setnik B, Bass A, Bramson C, et al. Abuse Potential Study of ALO-02 (Extended-Release Oxycodone Surrounding Sequestered Naltrexone) Compared with Immediate-Release Oxycodone Administered Orally to Nondependent Recreational Opioid Users. Pain Med. 2016. 45. U.S. Food and Drug Administration (FDA). Targiniq ER (oxycodone hydrochloride and naloxone hydrochloride extended-release tablets) Package Insert. https://www.accessdata.fda.gov/drugsatfda_docs/label/2014/205777lbl.pdf. Accessed March, 2017. 46. Bond M, Darwish M, Ma Y, Webster L. Evaluation of the abuse potential of an hydrocodone extended-release bitartrate tablet formulated with abuse-deterrence technology in nondependent, recreational opioid users. Drug and alcohol dependence. 2015;156(13). ©Institute for Clinical and Economic Review, 2017 Page 86 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 47. Setnik B, Sommerville K, Goli V, Han L, Webster L. Assessment of pharmacodynamic effects following oral administration of crushed morphine sulfate and naltrexone hydrochloride extended-release capsules compared with crushed morphine sulfate controlled-release tablets and placebo in nondependent recreational opioid users. Pain Medicine. 2013;14(8):1173-1186. 48. Stauffer J, Setnik B, Sokolowska M, Romach M, Johnson F, Sellers E. Subjective effects and safety of whole and tampered morphine sulfate and naltrexone hydrochloride (ALO-01) extended- release capsules versus morphine solution and placebo in experienced non-dependent opioid users: a randomized, double-blind, placebo-controlled, crossover study.[Erratum appears in Clin Drug Investig. 2011;31(8):598]. Clinical drug investigation. 2009;29(12):777-790. 49. Harris SC, Perrino PJ, Smith I, et al. Abuse potential, pharmacokinetics, pharmacodynamics, and safety of intranasally administered crushed oxycodone HCl abuse-deterrent controlled-release tablets in recreational opioid users. Journal of Clinical Pharmacology. 2014;54(4):468-477. 50. Webster LR, Kopecky EA, Smith MD, Fleming AB. A Randomized, Double-Blind, Double-Dummy Study to Evaluate the Intranasal Human Abuse Potential and Pharmacokinetics of a Novel Extended-Release Abuse-Deterrent Formulation of Oxycodone. Pain Med. 2016;17(6):1112- 1130. 51. Setnik B, Bramson C, Bass A, et al. Intranasal administration of crushed ALO-02 (extended- release oxycodone with sequestered naltrexone): A randomized, controlled abuse-potential study in nondependent recreational opioid users. Journal of clinical pharmacology. 2015;55(12):1351-1361. 52. Harris SC, Cipriano A, Colucci SV, et al. Intranasal abuse potential, pharmacokinetics, and safety of once-daily, single-entity, extended-release hydrocodone (HYD) in recreational opioid users. Pain Medicine. 2016;17(5):820-831. 53. Setnik B, Goli V, evy-Cooperman NL, Mills C, Shram M, Smith I. Assessing the subjective and physiological effects of intranasally administered crushed extended-release morphine formulations with and without a sequestered naltrexone core in recreational opioid users. Pain Research & Management. 2013;18(4):e55-e62. 54. Webster LR, Pantaleon C, Shah MS, et al. A Randomized, Double-Blind, Double-Dummy, Placebo- Controlled, Intranasal Drug Liking Study on a Novel Abuse-Deterrent Formulation of Morphine- Morphine ARER. Pain Med. 2016. 55. Webster LR, Smith MD, Lawler J, Lindhardt K, Dayno JM. Human Abuse Potential of an Abuse- Deterrent (AD), Extended-Release (ER) Morphine Product Candidate (Morphine-ADER Injection- Molded Tablets) vs Extended-Release Morphine Administered Intranasally in Nondependent Recreational Opioid Users. Pain Med. 2016. 56. U.S. Food and Drug Administration (FDA). RoxyBond IR (oxycodone hydrochloride) Package Insert. https://www.accessdata.fda.gov/drugsatfda_docs/label/2017/209777lbl.pdf. Accessed May 26, 2017. 57. Severtson SG, Ellis MS, Kurtz SP, et al. Sustained reduction of diversion and abuse after introduction of an abuse deterrent formulation of extended release oxycodone. Drug and alcohol dependence. 2016;168:219-229. 58. Butler SF, Cassidy TA, Chilcoat H, et al. Abuse rates and routes of administration of reformulated extended-release oxycodone: Initial findings from a sentinel surveillance sample of individuals assessed for substance abuse treatment. The Journal of Pain. 2013;14(4):351-358. ©Institute for Clinical and Economic Review, 2017 Page 87 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 59. Severtson SG, Bartelson BB, Davis JM, et al. Reduced abuse, therapeutic errors, and diversion following reformulation of extended-release oxycodone in 2010. The Journal of Pain. 2013;14(10):1122-1130. 60. Coplan PM, Chilcoat HD, Butler SF, et al. The effect of an abuse-deterrent opioid formulation (OxyContin) on opioid abuse-related outcomes in the postmarketing setting. Clinical Pharmacology and Therapeutics. 2016;100(3):275-286. 61. Cicero TJ, Ellis MS. Abuse-deterrent formulations and the prescription opioid abuse epidemic in the United States: Lessons learned from OxyContin. JAMA Psychiatry. 2015;72(5):424-429. 62. Cicero TJ, Ellis MS, Kasper ZA. A tale of 2 ADFs: Differences in the effectiveness of abuse- deterrent formulations of oxymorphone and oxycodone extended-release drugs. Pain. 2016;157(6):1232-1238. 63. Cicero TJ, Ellis MS, Surratt HL. Effect of abuse-deterrent formulation of OxyContin. N Engl J Med. 2012;367(2):187-189. 64. Cassidy TA, DasMahapatra P, Black RA, Wieman MS, Butler SF. Changes in prevalence of prescription opioid abuse after introduction of an abuse-deterrent opioid formulation. Pain Medicine. 2014;15(3):440-451. 65. Jones CM, Muhuri PK, Lurie PG. Trends in the Nonmedical Use of OxyContin, United States, 2006-2013. Clinical Journal of Pain. 2016. 66. De Veaugh-Geiss A, Coplan P, Chilcoat H, Sessler N, Singh R. Changes in nonmedical use of oxycontin® after reformulation with abuse deterrent properties. Postgraduate Medicine. 2016;128:22-23. 67. Coplan PM, Kale H, Sandstrom L, Landau C, Chilcoat HD. Changes in oxycodone and heroin exposures in the National Poison Data System after introduction of extended-release oxycodone with abuse-deterrent characteristics. Pharmacoepidemiology & Drug Safety. 2013;22(12):1274- 1282. 68. Kadakia A, Coplan P. Decrease in diagnosed abuse, addiction, and opioid poisoning among patients prescribed opioids after introduction of oxycontin with abuse-deterrent characteristics. Pharmacoepidemiology and Drug Safety. 2015;24:472-473. 69. Havens JR, Leukefeld CG, DeVeaugh-Geiss AM, Coplan P, Chilcoat HD. The impact of a reformulation of extended-release oxycodone designed to deter abuse in a sample of prescription opioid abusers. Drug and alcohol dependence. 2014;139:9-17. 70. Sankey C, Setnik B, Harsanyi Z, Michalko K, Yang Z, Geoffroy P. Opioid use following the introduction of an extended-release oxycodone formulation with tamper-resistant properties: Prospective historical chart review in methadone-maintained patients. Journal of Opioid Management. 2016;12(2):149-159. 71. Degenhardt L, Bruno R, Ali R, Lintzeris N, Farrell M, Larance B. The introduction of a potentially abuse deterrent oxycodone formulation: Early findings from the Australian National Opioid Medications Abuse Deterrence (NOMAD) study. Drug and alcohol dependence. 2015;151:56-67. 72. Larochelle MR, Zhang F, Ross-Degnan D, Wharam JF. Rates of opioid dispensing and overdose after introduction of abuse-deterrent extended-release oxycodone and withdrawal of propoxyphene. JAMA Internal Medicine. 2015;175(6):978-987. 73. Sessler NE, Downing JM, Kale H, Chilcoat HD, Baumgartner TF, Coplan PM. Reductions in reported deaths following the introduction of extended-release oxycodone (OxyContin) with an abuse-deterrent formulation. Pharmacoepidemiology & Drug Safety. 2014;23(12):1238-1246. ©Institute for Clinical and Economic Review, 2017 Page 88 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 74. Sessler NE, Downing JM, Kale HP, et al. Reductions in fatalities following introduction of a reformulated opioid with abuse-deterrent properties. Pharmacoepidemiology and Drug Safety. 2013;22:256. 75. Alpert A, Powell D, Pacula RL. Supply-Side Drug Policy in the Presence of Substitutes: Evidence from the Introduction of Abuse-Deterrent Opioids. The National Bureau of Economic Research;2017. 76. Sembower MA, Ertischek MD, Buchholtz C, Dasgupta N, Schnoll SH. Surveillance of diversion and nonmedical use of extended-release prescription amphetamine and oral methylphenidate in the United States. J Addict Dis. 2013;32(1):26-38. 77. Biglan A, Ary D, Wagenaar AC. The value of interrupted time-series experiments for community intervention research. Prev Sci. 2000;1(1):31-49. 78. Dart RC, Surratt HL, Cicero TJ, et al. Trends in opioid analgesic abuse and mortality in the United States. N Engl J Med. 2015;372(3):241-248. 79. Compton WM, Jones CM, Baldwin GT. Relationship between Nonmedical Prescription-Opioid Use and Heroin Use. N Engl J Med. 2016;374(2):154-163. 80. Abuse-Deterrent Opioid Formulations. JAMA. 2015;314(16):1744-1745. 81. Cheatle MD. Biopsychosocial Approach to Assessing and Managing Patients with Chronic Pain. The Medical clinics of North America. 2016;100(1):43-53. 82. U.S. Food and Drug Administration (FDA). Briefing Document-Postmarketing safety issues related to reformulated Opana ER®. In: Joint Meeting Of The Anesthetic And Analgesic Drug Products Advisory Committee And The Drug Safety And Risk Management Advisory Committee March 13 and 14, 2017 - Advisory Committee Briefing Materials: Available For Public Release.2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM545760.pdf. Accessed 2017-03-27. 83. U.S. Food and Drug Administration (FDA). Minutes for the March 13-14, 2017 Joint Meeting of the Drug Safety and Risk Management Advisory Committee (DSaRM) and the Anesthetic and Analgesic Drug Products Advisory Committee (AADPAC). In: U.S. Department of Health & Human Services; 2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM551226.pdf. Accessed 2017-04-14. 84. U.S. Food and Drug Administration (FDA). Transcript for the March 13, 2017 Joint Meeting of the Drug Safety and Risk Management Advisory Committee (DSaRM) and the Anesthetic and Analgesic Drug Products Advisory Committee (AADPAC). In: U.S. Department of Health & Human Services; 2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM553190.pdf#page=261. Accessed 2017-06-13. 85. Rice JB, Kirson NY, Shei A, et al. Estimating the costs of opioid abuse and dependence from an employer perspective: a retrospective analysis using administrative claims data. Appl Health Econ Health Policy. 2014;12(4):435-446. 86. Martin BC, Fan MY, Edlund MJ, Devries A, Braden JB, Sullivan MD. Long-term chronic opioid therapy discontinuation rates from the TROUP study. Journal of general internal medicine. 2011;26(12):1450-1457. ©Institute for Clinical and Economic Review, 2017 Page 89 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 87. Rossiter LF, Kirson NY, Shei A, et al. Medical cost savings associated with an extended-release opioid with abuse-deterrent technology in the US. J Med Econ. 2014;17(4):279-287. 88. Social Security Administration. Estimates from the 2016 Trustees Report. 2016. https://www.ssa.gov/OACT/STATS/table4c6.html. 89. Centers for Disease Control and Prevention. Calculating Total Daily Dose Of Opioids For Safer Dosage. https://www.cdc.gov/drugoverdose/pdf/calculating_total_daily_dose-a.pdf. Accessed 02/15/2017. 90. Office of Acquisition and Logistics (OAL). Federal Supply Schedule Contracting: Pharmaceutical Prices. In: Affairs USDoV, ed2017. 91. Commission CoMHP. MA Opioid Use - Data on file. In: Commonwealth of Massachusetts Health Policy Commission; 2017. 92. Bureau of Labor Statistics. Archived Consumer Price Index Detailed Report Information. United States Department of Labor; 2017. 93. Center for Behavioral Health Statistics and Quality. 2015 National Survey on Drug Use and Health: Detailed Tables. Substance Abuse and Mental Health Services Administration, Rockville, MD;2016. 94. Birnbaum HG, White AG, Schiller M, Waldman T, Cleveland JM, Roland CL. Societal costs of prescription opioid abuse, dependence, and misuse in the United States. Pain Med. 2011;12(4):657-667. 95. Health MDoP. MA Prescription Monitoring Program County‐Level Data Measures (Calendar Year 2015). 2016. 96. Prescription Opioid Overdose Deaths and Death Rate per 100,000 Population (Age-Adjusted). 2017. http://kff.org/other/state-indicator/prescription-opioid-overdose-deaths-and-death-rate- per-100000-population-age- adjusted/?dataView=1&currentTimeframe=0&selectedRows=%7B%22nested%22:%7B%22conn ecticut%22:%7B%7D,%22maine%22:%7B%7D,%22massachusetts%22:%7B%7D,%22new- hampshire%22:%7B%7D,%22rhode- island%22:%7B%7D,%22vermont%22:%7B%7D%7D%7D&sortModel=%7B%22colId%22:%22Loca tion%22,%22sort%22:%22asc%22%7D. Accessed 04/01/2017. 97. Development OFEC-oA. Hourly Earnings (MEI). 2017; Hourly wage index. Available at: https://stats.oecd.org/Index.aspx?DataSetCode=EAR_MEI#. Accessed 6/1, 2017. 98. Scarpati LM, Kirson NY, Jia ZB, Wen J, Howard J. Opioid Abuse: A Detailed Examination of Cost Drivers over a 24-Month Follow-up Period. Journal of Managed Care & Specialty Pharmacy.0(0):1-6. 99. Wikipedia. Opioid receptor. 2017; https://en.wikipedia.org/wiki/Opioid_receptor. Accessed March 20, 2017. 100. Paulozzi LJ. Prescription drug overdoses: a review. Journal of safety research. 2012;43(4):283- 289. 101. Kirkpatrick DL, Schmidt WK, Morales R, et al. In vitro and in vivo assessment of the abuse potential of PF614, a novel BIO-MD prodrug of oxycodone. J Opioid Manag. 2017;13(1):39-49. 102. U.S. Food and Drug Administration (FDA). FDA Facts: Abuse-Deterrent Opioid Medications. 2017-01-17; http://www.fda.gov/NewsEvents/Newsroom/FactSheets/ucm514939.htm. Accessed 2017-03-27. ©Institute for Clinical and Economic Review, 2017 Page 90 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 103. Seya MJ, Gelders SF, Achara OU, Milani B, Scholten WK. A first comparison between the consumption of and the need for opioid analgesics at country, regional, and global levels. J Pain Palliat Care Pharmacother. 2011;25(1):6-18. 104. Duthey B, Scholten W. Adequacy of opioid analgesic consumption at country, global, and regional levels in 2010, its relationship with development level, and changes compared with 2006. J Pain Symptom Manage. 2014;47(2):283-297. 105. Pain & Policy Studies Group. Custom Consumption Graphs for Opioid Medicines. 2013; https://ppsg-chart.medicine.wisc.edu/ Accessed 2017-04-13. 106. Nelson LS, Juurlink DN, Perrone J. Addressing the Opioid Epidemic. JAMA. 2015;314(14):1453- 1454. 107. Brennan MJ. Update on prescription extended-release opioids and appropriate patient selection. Journal of multidisciplinary healthcare. 2013;6:265-280. 108. Jensen TS. Opioids in the brain: supraspinal mechanisms in pain control. Acta Anaesthesiol Scand. 1997;41(1 Pt 2):123-132. 109. Fields HL, Margolis EB. Understanding opioid reward. Trends Neurosci. 2015;38(4):217-225. 110. Holden JE, Jeong Y, Forrest JM. The endogenous opioid system and clinical pain management. AACN Clin Issues. 2005;16(3):291-301. 111. Hou H, Wang C, Jia S, Hu S, Tian M. Brain dopaminergic system changes in drug addiction: a review of positron emission tomography findings. Neurosci Bull. 2014;30(5):765-776. 112. Inciardi JA, Surratt HL, Cicero TJ, Kurtz SP, Martin SS, Parrino MW. The "black box" of prescription drug diversion. J Addict Dis. 2009;28(4):332-347. 113. Alam A, Juurlink DN. The prescription opioid epidemic: an overview for anesthesiologists. Canadian journal of anaesthesia = Journal canadien d'anesthesie. 2016;63(1):61-68. 114. Agency for Health Care Quality and Research. Acute pain management: operative or medical procedures and trauma. 1992; https://archive.ahrq.gov/clinic/medtep/acute.htm#acutefind. Accessed 2017-03-24. 115. Fishman SM. Recognizing pain management as a human right: a first step. Anesth Analg. 2007;105(1):8-9. 116. Brennan F, Carr DB, Cousins M. Pain management: a fundamental human right. Anesth Analg. 2007;105(1):205-221. 117. Phillips DM. JCAHO pain management standards are unveiled. Joint Commission on Accreditation of Healthcare Organizations. JAMA. 2000;284(4):428-429. 118. Kotecha MK, Sites BD. Pain policy and abuse of prescription opioids in the USA: a cautionary tale for Europe. Anaesthesia. 2013;68(12):1210-1215. 119. Tefera L, Lehrman WG, Goldstein EG, Agrawal S. A Special Contribution from the Centers for Medicare and Medicaid Services: Valuing Patient Experience While Addressing the Prescription Opioid Epidemic. Annals of emergency medicine. 2017;69(2):181-183. 120. Schatman ME, Webster LR. The health insurance industry: perpetuating the opioid crisis through policies of cost-containment and profitability. J Pain Res. 2015;8:153-158. 121. Kolodny A, Courtwright DT, Hwang CS, et al. The prescription opioid and heroin crisis: a public health approach to an epidemic of addiction. Annu Rev Public Health. 2015;36:559-574. 122. O'Brien C. Addiction and dependence in DSM-V. Addiction (Abingdon, England). 2011;106(5):866-867. 123. American Academy of Pain Medicine, American Pain Society, American Society of Addiction Medicine. Definitions Related to the Use of Opioids for the Treatment of Pain: Consensus ©Institute for Clinical and Economic Review, 2017 Page 91 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Statement of the American Academy of Pain Medicine, the American Pain Society, and the American Society of Addiction Medicine. 2001; http://www.asam.org/docs/default- source/public-policy-statements/1opioid-definitions-consensus-2-011.pdf? Accessed 2017-02- 20, 2017. 124. Katz N. Opioids: after thousands of years, still getting to know you. Clin J Pain. 2007;23(4):303- 306. 125. U.S. Food and Drug Administration (FDA). Presentations for the April 5, 2017 Joint Meeting of the Drug Safety and Risk Management Advisory Committee and the Anesthetic and Analgesic Drug Products Advisory Committee. In:2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM551778.pdf. Accessed 2017-04-11. 126. Carmona-Bayonas A, Jimenez-Fonseca P, Castanon E, et al. Chronic opioid therapy in long-term cancer survivors. Clinical & translational oncology : official publication of the Federation of Spanish Oncology Societies and of the National Cancer Institute of Mexico. 2017;19(2):236-250. 127. Atkinson TJ, Schatman ME, Fudin J. The damage done by the war on opioids: the pendulum has swung too far. J Pain Res. 2014;7:265-268. 128. Pezalla EJ, Rosen D, Erensen JG, Haddox JD, Mayne TJ. Secular trends in opioid prescribing in the USA. J Pain Res. 2017;10:383-387. 129. Belin-Rauscent A, Fouyssac M, Bonci A, Belin D. How Preclinical Models Evolved to Resemble the Diagnostic Criteria of Drug Addiction. Biol Psychiatry. 2016;79(1):39-46. 130. Fudin J, Atkinson TJ. Opioid prescribing levels off, but is less really more? Pain Med. 2014;15(2):184-187. 131. Shah A, Hayes CJ, Martin BC. Characteristics of Initial Prescription Episodes and Likelihood of Long-Term Opioid Use - United States, 2006-2015. MMWR Morbidity and mortality weekly report. 2017;66(10):265-269. 132. Vowles KE, McEntee ML, Julnes PS, Frohe T, Ney JP, van der Goes DN. Rates of opioid misuse, abuse, and addiction in chronic pain: a systematic review and data synthesis. Pain. 2015;156(4):569-576. 133. Gasior M, Bond M, Malamut R. Routes of abuse of prescription opioid analgesics: a review and assessment of the potential impact of abuse-deterrent formulations. Postgrad Med. 2016;128(1):85-96. 134. Keyes KM, Cerda M, Brady JE, Havens JR, Galea S. Understanding the rural-urban differences in nonmedical prescription opioid use and abuse in the United States. American journal of public health. 2014;104(2):e52-59. 135. Cerda M, Gaidus A, Keyes KM, et al. Prescription opioid poisoning across urban and rural areas: identifying vulnerable groups and geographic areas. Addiction (Abingdon, England). 2017;112(1):103-112. 136. Rudd RA, Seth P, David F, Scholl L. Increases in Drug and Opioid-Involved Overdose Deaths - United States, 2010-2015. MMWR Morbidity and mortality weekly report. 2016;65(5051):1445- 1452. https://www.cdc.gov/mmwr/volumes/65/wr/mm655051e1.htm. Accessed Dec 30. 137. Substance Abuse and Mental Health Services Administration. National Survey on Drug Use and Health. 2017; https://nsduhweb.rti.org/respweb/homepage.cfm. Accessed 2017-04-13. 138. Jones CM, Paulozzi LJ, Mack KA. Sources of prescription opioid pain relievers by frequency of past-year nonmedical use United States, 2008-2011. JAMA Intern Med. 2014;174(5):802-803. ©Institute for Clinical and Economic Review, 2017 Page 92 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 139. Katz NP, Birnbaum HG, Castor A. Volume of prescription opioids used nonmedically in the United States. J Pain Palliat Care Pharmacother. 2010;24(2):141-144. 140. Mastropietro DJ, Omidian H. Abuse-deterrent formulations: part 1 - development of a formulation-based classification system. Expert Opin Drug Metab Toxicol. 2015;11(2):193-204. 141. U.S. Food and Drug Administration (FDA). Overview of the May 5, 2008 ALSDAC Meeting to Discuss NDA 21-272 for a New, Abuse-Resistant Formulation of Oxycontin. In: Washington D.C.2008: http://www.fda.gov/ohrms/dockets/ac/08/briefing/2008-4356b1-01-FDA.pdf. Accessed 2017-02-08. 142. Alexander L, Mannion RO, Weingarten B, Fanelli RJ, Stiles GL. Development and impact of prescription opioid abuse deterrent formulation technologies. Drug and alcohol dependence. 2014;138:1-6. 143. Nguyen V, Raffa RB, Taylor R, Pergolizzi JV, Jr. The role of abuse-deterrent formulations in countering opioid misuse and abuse. J Clin Pharm Ther. 2015;40(6):629-634. 144. U.S. Food & Drug Administration (FDA). CFR - Code of Federal Regulations Title 21, Part 314 -- Applications for FDA Approval to Market a New Drug. 2016; https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=314.510. 145. Conrad C, Bradley HM, Broz D, et al. Community Outbreak of HIV Infection Linked to Injection Drug Use of Oxymorphone--Indiana, 2015. MMWR Morbidity and mortality weekly report. 2015;64(16):443-444. 146. Centers for Disease Control and Prevention. Thrombotic thrombocytopenic purpura (TTP)-like illness associated with intravenous Opana ER abuse--Tennessee, 2012. MMWR Morbidity and mortality weekly report. 2013;62(1):1-4. 147. Ban BH, Verma A, Tudor M, Sethi J. Opana-induced thrombotic microangiopathy masquerading as thrombotic thrombocytopenic purpura. Oxf Med Case Reports. 2017;2017(6):omx026. 148. U.S. Food & Drug Administration (FDA). FDA requests removal of Opana ER for risks related to abuse. FDA News Release 2017; https://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm562401.htm. Accessed 6/13, 2017. 149. Endo Pharmaceuticals Inc. Endo Presentations for the March 13-14, 2017 Joint Meeting of the Anesthetic and Analgesic Drug Products Advisory Committee and the Drug Safety and Risk Management Advisory Committee. In:2017: https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM547236.pdf#page=22. Accessed 2017-06-13. 150. Smith HS. Opioid metabolism. Mayo Clinic proceedings. 2009;84(7):613-624. 151. Sapienza FL. Abuse deterrent formulations and the Controlled Substances Act (CSA). Drug and alcohol dependence. 2006;83 Suppl 1:S23-30. 152. Webster LR, Brennan MJ, Kwong LM, Levandowski R, Gudin JA. Opioid abuse-deterrent strategies: role of clinicians in acute pain management. Postgrad Med. 2016;128(1):76-84. 153. Thielking M. Missouri is the only state not monitoring prescription drug use. Will it finally create a database? 2017; https://www.statnews.com/2017/03/07/missouri-prescription-drug- database/ 154. U.S. Department Of Health and Human Services. Opioid abuse in the United States and Department of Health and Human Services actions to address opioid-drug-related overdoses and deaths. J Pain Palliat Care Pharmacother. 2015;29(2):133-139. ©Institute for Clinical and Economic Review, 2017 Page 93 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 155. International Narcotics Control Board. Report 2016. In:2016: https://www.incb.org/documents/Publications/AnnualReports/AR2016/English/AR2016_E_ebo ok.pdf. Accessed 2017-04-12. 156. The White House. Fact Sheet: Obama Administration Announces Public and Private Sector Efforts to Address Prescription Drug Abuse and Heroin Use. 2015; https://obamawhitehouse.archives.gov/the-press-office/2015/10/21/fact-sheet-obama- administration-announces-public-and-private-sector. Accessed 2017-03-13. 157. Cook DJ, Mulrow CD, Haynes RB. Systematic reviews: synthesis of best evidence for clinical decisions. Ann Intern Med. 1997;126(5):376-380. 158. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. International journal of surgery (London, England). 2010;8(5):336-341. 159. Ollendorf D, Pearson S. An integrated evidence rating to frame comparative effectiveness assessments for decision makers. Med Care. 2010;48(6 Suppl):S145-152. 160. Agency for Healthcare Research and Quality. U.S. Preventive Services Task Force Procedure Manual: Publication No. 08-05118-EF. 2008. 161. National Institutes of Health (U.S.), National Heart Lung and Blood Institute. Quality Assessment Tool for Before-After (Pre-Post) Studies With No Control Group. Study Quality Assessment Tools https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk- reduction/tools/before-after. Accessed April 14, 2017. 162. U.S. Food and Drug Administration (FDA). FDA ADVISORY COMMITTEE BRIEFING DOCUMENT RoxyBond immediate-release tablets. 2017; https://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/An estheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM550022.pdf. Accessed May 31, 2017. 163. Peacock A, Degenhardt L, Hordern A, et al. Methods and predictors of tampering with a tamper- resistant controlled-release oxycodone formulation. International Journal of Drug Policy. 2015;26(12):1265-1272. 164. Michna E, Kirson NY, Shei A, Birnbaum HG, Ben-Joseph R. Use of prescription opioids with abuse-deterrent technology to address opioid abuse. Current Medical Research and Opinion. 2014;30(8):1589-1598. 165. Chilcoat HD, Coplan PM, Harikrishnan V, Alexander L. Decreased diversion by doctor-shopping for a reformulated extended release oxycodone product (OxyContin). Drug and alcohol dependence. 2016;165:221-228. 166. Cepeda MS, Fife D, Chow W, Mastrogiovanni G, Henderson SC. Opioid shopping behavior: how often, how soon, which drugs, and what payment method. J Clin Pharmacol. 2013;53(1):112- 117. 167. Hwang CS, Chang HY, Alexander GC. Impact of abuse-deterrent OxyContin on prescription opioid utilization. Pharmacoepidemiology & Drug Safety. 2015;24(2):197-204. 168. Cipriano M. Abuse-Deterrent Opioids: Postmarketing Data Eyed As Development ‘Anchor’. 2017. https://pink.pharmamedtechbi.com/PS120399/AbuseDeterrent-Opioids-Postmarketing-Data- Eyed-As-Development-Anchor. Accessed 2017-04-14. 169. Hooten WM, St Sauver JL, McGree ME, Jacobson DJ, Warner DO. Incidence and Risk Factors for Progression From Short-term to Episodic or Long-term Opioid Prescribing: A Population-Based Study. Mayo Clinic proceedings. 2015;90(7):850-856. ©Institute for Clinical and Economic Review, 2017 Page 94 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 170. Von Korff M, Saunders K, Thomas Ray G, et al. De facto long-term opioid therapy for noncancer pain. Clin J Pain. 2008;24(6):521-527. 171. Hughes A, Williams M, R., Lipari R, N., et al. Prescription Drug Use and Misuse in the United States : Results from the 2015 National Survey on Drug Use and Health. Substance Abuse and Mental Health services Administration;2016. 172. White AG, Birnbaum HG, Rothman DB, Katz N. Development of a budget-impact model to quantify potential cost savings from prescription opioids designed to deter abuse or ease of extraction. Appl Health Econ Health Policy. 2009;7(1):61-70. 173. Winegarden W. Estimating the Net Economic Benefit of Abuse-Deterrent Opioids. EconoSTATS 2015; http://econostats.org/estimating-the-net-economic-benefit-of-abuse-deterrent-opioids/ Accessed 2016-10-04. 174. Kirson NY, S A, White AG, Birnbaum HG, Rami B, Michna E. Societal Economic Benefits Associated with an ExtendedRelease Opioid with AbuseDeterrent Technology in the United States. Pain Medicine. 2014;15:1450-1454. 175. Skinner B, J. Societal cost savings from abuse deterrent formulations for prescription opioids in Canada. Canadian Health Policy. 2017. 176. Health VDo. Annual report - 2015, Vermont Prescription Monitoring Program. 2016. 177. Kaczmarek SJ. The Impact of New Hepatitis C Drug Therapy on Individual Medicare Part D Spending. In: Milliman Client Report, ed. Pharmaceutical Care Management Association,.2014: http://www.natap.org/2014/HCV/partdpremiumstudymilliman.pdf. 178. Hwang CS, Turner LW, Kruszewski SP, Kolodny A, Alexander GC. Primary Care Physicians' Knowledge and Attitudes Regarding Prescription Opioid Abuse and Diversion. Clin J Pain. 2016;32(4):279-284. 179. U.S. Food and Drug Administration (FDA). Quick Minutes of the Joint Meeting of the Anesthetic and Life Support Drugs Advisory Committee and the Drug Safety and Risk Management Advisory Committee October 21 -22 2010. In: U.S. Department of Health & Human Services; 2010: https://wayback.archive- it.org/7993/20170113044232/http:/www.fda.gov/downloads/AdvisoryCommittees/Committees MeetingMaterials/Drugs/AnestheticAndAnalgesicDrugProductsAdvisoryCommittee/UCM236242 .pdf. Accessed 2017-07-24. 180. U.S. Food and Drug Administration (FDA). Letter to Collegium Pharmaceuticals, Inc. Re: Xtampza ER NDA. 2016; https://www.accessdata.fda.gov/drugsatfda_docs/appletter/2016/208090Orig1s000ltr.pdf. Accessed April, 2017. 181. Johnson F, Setnik B. Morphine sulfate and naltrexone hydrochloride extended-release capsules: naltrexone release, pharmacodynamics, and tolerability. Pain Physician. 2011;14(4):391-406. 182. Black R, Coplan P, Cassidy T, et al. Effects of reformulated OxyContin® among patients assessed for substance abuse treatment in the NAVIPPRO sentinel surveillance network. Journal of Pain. 2012;13(4):S58. 183. Butler S, Black R, Kopecky E, Thompson C, Fleming A. Relative abuse of crush-resistant tablets of prescription opioids via alternative oral modes of administration. Postgraduate Medicine. 2016;128:89. 184. Coplan PM, Kadakia A. Changes in diagnosed addiction rates in patients prescribed OxyContin (ERO) or other opioids after introduction of ERO with abuse-deterrent properties. Pharmacoepidemiology and Drug Safety. 2015;24:27. ©Institute for Clinical and Economic Review, 2017 Page 95 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 185. Davis J, Severtson SG, Bartelson BB, et al. Changes in diversion rates following the introduction of a reformulated extended release oxycodone product. Annals of emergency medicine. 2012;60(4):S35. 186. Severtson SG, Bartelson BB, Davis J, et al. Difference in rates of abuse following reformulation of extended release oxycodone using data from the RADARS® system poison center program. Annals of emergency medicine. 2012;60(4):S34-S35. 187. Coplan PM, Green CA, Perrin N, et al. Effects of opioid analgesic tablets resistant to breaking, crushing and dissolving on patient safety outcomes. Pharmacoepidemiology and Drug Safety. 2013;22:88-89. ©Institute for Clinical and Economic Review, 2017 Page 96 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents APPENDICES ©Institute for Clinical and Economic Review, 2017 Page 97 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix A. Search Strategies and Results Table A1. PRISMA 2009 Checklist # Checklist item TITLE Title 1 Identify the report as a systematic review, meta-analysis, or both. ABSTRACT Structured summary 2 Provide a structured summary including, as applicable: background; objectives; data sources; study eligibility criteria, participants, and interventions; study appraisal and synthesis methods; results; limitations; conclusions and implications of key findings; systematic review registration number. INTRODUCTION Rationale 3 Describe the rationale for the review in the context of what is already known. Objectives 4 Provide an explicit statement of questions being addressed with reference to participants, interventions, comparisons, outcomes, and study design (PICOS). METHODS Protocol and registration 5 Indicate if a review protocol exists, if and where it can be accessed (e.g., Web address), and, if available, provide registration information including registration number. Eligibility criteria 6 Specify study characteristics (e.g., PICOS, length of follow-up) and report characteristics (e.g., years considered, language, publication status) used as criteria for eligibility, giving rationale. Information sources 7 Describe all information sources (e.g., databases with dates of coverage, contact with study authors to identify additional studies) in the search and date last searched. Search 8 Present full electronic search strategy for at least one database, including any limits used, such that it could be repeated. Study selection 9 State the process for selecting studies (i.e., screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis). Data collection process 10 Describe method of data extraction from reports (e.g., piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators. Data items 11 List and define all variables for which data were sought (e.g., PICOS, funding sources) and any assumptions and simplifications made. ©Institute for Clinical and Economic Review, 2017 Page 98 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Risk of bias in individual 12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the studies study or outcome level), and how this information is to be used in any data synthesis. Summary measures 13 State the principal summary measures (e.g., risk ratio, difference in means). Synthesis of results 14 Describe the methods of handling data and combining results of studies, if done, including measures of consistency (e.g., I 2) for each meta-analysis. Risk of bias across studies* 15 Specify any assessment of risk of bias that may affect the cumulative evidence (e.g., publication bias, selective reporting within studies). Additional analyses 16 Describe methods of additional analyses (e.g., sensitivity or subgroup analyses, meta-regression), if done, indicating which were pre-specified. RESULTS Study selection 17 Give numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage, ideally with a flow diagram. Study characteristics 18 For each study, present characteristics for which data were extracted (e.g., study size, PICOS, follow-up period) and provide the citations. Risk of bias within studies 19 Present data on risk of bias of each study and, if available, any outcome level assessment (see item 12). Results of individual studies 20 For all outcomes considered (benefits or harms), present, for each study: (a) simple summary data for each intervention group (b) effect estimates and confidence intervals, ideally with a forest plot. Synthesis of results 21 Present results of each meta-analysis done, including confidence intervals and measures of consistency. Risk of bias across studies* 22 Present results of any assessment of risk of bias across studies (see Item 15). Additional analysis 23 Give results of additional analyses, if done (e.g., sensitivity or subgroup analyses, meta-regression [see Item 16]). DISCUSSION Summary of evidence 24 Summarize the main findings including the strength of evidence for each main outcome; consider their relevance to key groups (e.g., healthcare providers, users, and policymakers). Limitations 25 Discuss limitations at study and outcome level (e.g., risk of bias), and at review-level (e.g., incomplete retrieval of identified research, reporting bias). Conclusions 26 Provide a general interpretation of the results in the context of other evidence, and implications for future research. FUNDING Funding 27 Describe sources of funding for the systematic review and other support (e.g., supply of data); role of funders for the systematic review. From: Moher D, Liberati A, Tetzlaff J, Altman DG. The PRISMA Group (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. PLoS Med 6(6): e1000097. doi:10.1371/journal.pmed1000097; * We did not identify any study (ongoing or completed) relevant to this report on clinicaltrials.gov, therefore it was not possible to assess the cumulative evidence for publication bias or selective reporting within studies. ©Institute for Clinical and Economic Review, 2017 Page 99 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table A2. Search Strategy of Medline 1996 to Present with Daily Update, PsycINFO, EBM Reviews - Cochrane Database of Systematic Reviews, EBM Reviews - Cochrane Central Register of Controlled Trials on December 7, 2016 1 Delayed-Action Preparations/ or (extended release or controlled release or slow release or sustained release or delayed release).ti,ab. 2 embeda.ti,ab. 3 (Naltrexone/ and Morphine/) or (naltrexone and morphine).ti,ab 4 1 and 3 5 Morphine/ or morphine.ti,ab. 6 1 and 5 7 2 or 4 or 6 8 (xtampza or OxyContin or oxycodone naltrexone combination).ti,ab 9 Oxycodone/ or oxycodone.ti,ab. 10 1 and 9 11 8 or 10 12 hysingla.ti,ab. 13 Hydrocodone/ or hydrocodone.ti,ab. 14 1 and 13 15 12 or 14 16 targiniq.ti,ab. 17 (Oxycodone/ and Naloxone/) or (oxycodone and naloxone).ti,ab. 18 1 and 17 19 16 or 18 20 7 or 11 or 15 or 19 21 limit 20 to (english language and humans and yr="2000 -Current") [Limit not valid in PsycINFO,PsycTESTS,Books@Ovid,CDSR,ACP Journal Club,DARE,CCTR,Your Journals@Ovid,CLCMR; records were retained] 22 (guidelines or practice guideline or letter or editorial or news or case reports).mp. 23 21 not 22 24 Delayed-Action Preparations/ and Analgesics, Opioid/ 25 (abuse deter* adj5 formulation?).ti,ab. 26 (abuse deter* adj5 opi*).ti,ab. 27 (tamper resist* adj5 formulation?).mp. 28 (tamper resist* adj5 opi*).mp. 29 Analgesics, Opioid/ and Drug Compounding/ 30 Opioid-Related Disorders/pc [Prevention & Control] 31 Prescription Drug Misuse/ 32 30 or 31 33 24 or 25 or 26 or 27 or 28 or 29 34 32 and 33 ©Institute for Clinical and Economic Review, 2017 Page 100 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 35 limit 34 to (english language and humans and yr="2000 -Current") [Limit not valid in PsycINFO,PsycTESTS,Books@Ovid,CDSR,ACP Journal Club,DARE,CCTR,Your Journals@Ovid,CLCMR; records were retained] 36 (guidelines or practice guideline or letter or editorial or news or case reports).mp. 37 35 not 36 38 23 or 37 39 remove duplicates from 38 Table A3. Search Strategy of EMBASE on October 19, 2016 #1 OxyContin #2 Xtampza #3 Troxyca OR oxycodone NEAR/5 naltrexone #4 targiniq OR 'naloxone plus oxycodone' #5 hysingla OR 'hydrocodone bitartrate' #6 Vantrela #7 embeda OR 'morphine sulfate plus naltrexone' #8 Morphabond #9 arymo #10 #1 OR #2 OR #3 OR #4 OR #5 OR #6 OR #7 OR #8 OR #9 #11 naltrexone:ab,ti #12 morphine:ab,ti #13 hydrocodone:ab,ti #14 oxycodone:ab,ti #15 'controlled release formulation'/exp OR 'extended release':ab,ti OR 'controlled release':ab,ti OR 'delayed release':ab,ti OR 'slow release':ab,ti OR 'sustained release':ab,ti #16 'narcotic analgesic agent'/exp OR opi*:ab,ti #17 (abuse NEAR/5 deter*):ab,ti OR (tamper* NEAR/5 resist*):ab,ti #18 #16 AND #17 #19 ('abuse deter*' NEAR/5 formulation?):ab,ti OR ('tamp* resist*' NEAR/5 formulation?):ab,ti #20 #18 OR #19 #21 #15 AND #20 #22 #11 OR #12 OR #13 OR #14 #23 #21 AND #22 #24 #10 OR #23 #25 #24 AND [english]/lim AND [2000-2016]/py #26 #25 AND [medline]/lim #27 #25 NOT #26 #28 #27 AND [humans]/lim AND [animals]/lim #29 #27 AND [animals]/lim #30 #27 AND [humans]/lim #31 #30 NOT #28 NOT #29 ©Institute for Clinical and Economic Review, 2017 Page 101 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents #32 #31 AND ([editorial]/lim OR [erratum]/lim OR [letter]/lim OR [note]/lim OR [review]/lim OR [short survey]/lim) #33 #31 NOT #32 #34 #33 NOT 'case study' NOT 'case report' Figure A1. PRISMA flow Chart Showing Results of Literature Search for Abuse-deterrent Formulations 1,424 potentially relevant references screened 1,291 citations excluded Population: 240 Intervention: 268 Comparator: 69 Outcomes: 418 133 references for full text Study Type: 195 review Duplicates: 101 92 citations excluded (trials using simulated ADFs, non-comparative study designs, conference abstract duplicated peer- reviewed publication) 41 TOTAL • 15 RCTs o 13 publications o 2 conference abstracts/posters • 26 observational studies o 19 publications o 7 conference abstracts/posters • ©Institute for Clinical and Economic Review, 2017 Page 102 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix B. Public and Representative Private Insurer Coverage Policies Table B1. New England Coverage Scan Connecticut Maine Massachusetts New Hampshire Rhode Island Vermont Neighbor Neighbor- Anthem Anthem Tufts Anthem HPHC BCBS -hood HPHC New BCBS hood BCBS MVP (Wellpoint Connecticare (Wellpoint Health (Wellpoint Inc Maine of MA Health Hampshire of RI Health of VT Grp Inc Group) Inc Group) Plan Group) Plan Plan of RI Oxycodone OxyContin (Purdue, 2010) Covered Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes PA* Yes Yes No No Yes No No No No Yes Yes Yes Yes QL** Yes Yes No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Xtampza (Collegium, April 2016) Covered No No No Yes No No No No No No No Yes Yes PA No No No Yes Yes No No No No No No Yes Yes QL No No No Yes Yes No No No No No No Yes Hydrocodone Hysingla (Purdue, 2014) Covered No Yes Yes Yes No No No Yes Yes No Yes Yes Yes PA No Yes Yes No Yes No No No No No Yes Yes Yes QL No NL Yes Yes Yes No No Yes Yes No Yes Yes Yes Morphine Embeda (Pfizer, Approved: 2009; Relaunched: 2015) Covered No Yes Yes Yes Yes Yes Yes No Yes Yes No No Yes PA No Yes Yes No Yes No No No No No No No Yes QL No Yes Yes Yes Yes Yes Yes No Yes Yes No No Yes * QL. Quantity Limits; **PA. Prior Authorization ©Institute for Clinical and Economic Review, 2017 Page 103 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table B2. New England Medicaid Program Coverage Scan Connecticut Maine Massachusetts New Rhode Vermont Hampshire Island OxyContin (Purdue, 2010) Preferred No No No No No No PA* Yes Yes Yes Yes Yes Yes QL** Yes Yes Yes Yes NL Yes Xtampza (Collegium, April 2016) Preferred No No No No No No PA Yes Yes Yes Yes Yes Yes QL Yes Yes Yes NL NL Yes Hysingla (Purdue, 2014) Preferred No No No No No No PA Yes Yes Yes Yes Yes Yes QL Yes Yes Yes NL NL Yes Embeda (Pfizer, Approved: 2009; Relaunched: 2015) Preferred No Yes No No Yes Yes PA Yes No Yes Yes No No QL Yes Yes Yes NL NL Yes * QL. Quantity Limits **PA. Prior Authorization ©Institute for Clinical and Economic Review, 2017 Page 104 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Figure B1. Example of Detailed Coverage Policy from Anthem Maine Hysingla ER (hydrocodone bitartrate extended-release) Override(s) Approval Duration Prior Authorization Initial request: 3 months Quantity Limit Maintenance Therapy: Additional prior authorization required for each additional 6 months Individuals receiving for terminal diagnosis and receiving palliative care/end-of-life therapy: Lifetime Individuals receiving for cancer pain related to active cancer therapy: 1 year Medications Comments Quantity Limits Hysingla ER (hydrocodone bitartrate Non-Preferred 20mg, 30mg, 40mg, 60mg, 120mg: extended-release) 1 tablet per day 80mg, 100mg: 2 tablets per day Quantity Limit Override Criteria For approval of increased quantities of Hysingla ER (hydrocodone bitartrate extended-release), the following criteria must be met: I. Requests for increased quantity can be approved for the diagnosis of cancer related pain. Note: It may be possible in some instances to use a higher strength of the requested medication and take fewer tablets/capsules to achieve the same total daily dosage requested I. Initial requests for Hysingla ER (hydrocodone bitartrate extended-release) may be approved when the following criteria are met: A. Individual is 18 years of age or older; AND B. Individual has a diagnosis of pain severe enough to require daily, around‐the‐clock, long term opioid treatment (please document diagnosis); AND C. Individual has one of the following: a. An inadequate response to alternative treatment options, such as but not limited to non‐opioid analgesics and immediate‐release opioids; OR b. Alternative treatment options would otherwise be inadequate to provide sufficient management of pain; OR c. Individual has contraindications to non‐opioid analgesics (such as NSAID use in individuals with active ulcer condition/gastrointestinal bleeding, renal failure) 1; AND D. Individual is not opioid naïve as noted by the following: ©Institute for Clinical and Economic Review, 2017 Page 105 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents a. Individual has been maintained on a short‐acting opioid analgesic, including use of opioid analgesia as an inpatient for post-surgical pain; OR b. Individual is transitioning from one long‐acting opioid analgesic to another long‐ acting opioid analgesic; AND E. Prescriber has consulted with individual regarding risks of opioid therapy; AND F. Clear treatment goals have been defined and outlined as part of overall plan; OR G. Individual has one of the following: a. Diagnosis of cancer related pain and is actively undergoing cancer therapy; OR b. Diagnosis of terminal illness and is receiving palliative/end‐of‐life care. II. Requests for continuation of Hysingla ER (hydrocodone bitartrate extended-release) may be approved when the following criteria are met: A. Individual has a diagnosis of moderate to severe pain and requires around‐the‐clock long term opioid treatment (please document diagnosis); AND B. Individual has one of the following: a. An inadequate response to alternative treatment options, such as but not limited to non‐opioid analgesics and immediate‐release opioids; OR b. Alternative treatment options would otherwise be inadequate to provide sufficient management of pain; AND C. Therapy with long‐acting opioid has resulted in meaningful improvement in pain AND function; AND D. Risk assessment has been performed including the following: a. Urine drug screens have been obtained within the past year to assess for adherence to therapy; AND b. State prescription drug monitoring program (PDMP) data has been reviewed (where available). OR E. Individual has one of the following: a. Diagnosis of cancer related pain and is actively undergoing cancer therapy; OR b. Diagnosis of terminal illness and is receiving palliative/end-of-life care. III. Requests for Hysingla ER (hydrocodone bitartrate extended-release) may not be approved for the following: A. Individual is requesting or using as an as‐needed analgesic; OR B. Individual has one of the following conditions: a. Significant respiratory depression; OR b. Acute or severe bronchial asthma or hypercarbia; OR c. Known or suspected paralytic ileus. ©Institute for Clinical and Economic Review, 2017 Page 106 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents IV. Requests for Hysingla ER (hydrocodone bitartrate extended-release) must also meet the following criteria (in addition to the above criteria in I.-III.): A. Individual has had a trial and inadequate response or intolerance to two preferred long‐ acting agents; Preferred agents: Fentanyl patch (generic), levorphanol, methadone, methadose, morphine sulfate ER, OxyContin (brand), tramadol ER (generic), oxymorphone ER, hydromorphone ER. OR B. Individual has completed titration and is already maintained on a stable on dose of the requested drug; OR C. The preferred long‐acting opioids are not acceptable due to concomitant clinical situations, such as but not limited to: a. Known hypersensitivity to any ingredient which is not also in the requested non- preferred agent; OR D. Hysingla ER (hydrocodone bitartrate extended-release) abuse-deterrent may be approved if the individual has need for an abuse-deterrent formulation based upon a history of substance abuse disorder OR individual’s family member or household resident has active substance abuse disorder or a history of substance abuse disorder. NOTES: 1. Specific drug therapy and contraindication to therapy should be reported 2. Long‐acting opioid analgesics have a black box warning regarding risk of addiction, abuse and misuse, respiratory depression, risks of accidental exposure and risks for neonatal opioid withdrawal syndrome. Long‐ acting opioid analgesic use can lead to addiction, abuse and misuse which can lead to overdose and death. Individuals should be assessed before prescribing and monitored regularly during therapy for development of these behaviors or conditions. Serious, life‐threatening or fatal respiratory depression may occur while using long‐ acting opioid analgesics. Individuals should be monitored, particularly upon initiation or upon dose increases. Accidental exposure, especially in children, can result in fatal overdose. Prolonged exposure to long‐acting opioid analgesics during pregnancy can result in neonatal opioid withdrawal syndrome. If opioid use is required for prolonged periods of time in a pregnant woman, the individual should be advised of the risk of neonatal opioid withdrawal syndrome and ensure appropriate treatment will be available. Some long acting analgesics (hydrocodone based) may interact with cytochrome P450 3A4 inhibitors, resulting in increased opioid concentration. In addition, discontinuation of a cytochrome P450 3A4 inducer may also result in an increase in opioid concentration. Monitor individuals receiving these opioid analgesics and any cytochrome P450 3A4 inhibitor or inducer. Co‐ingestion with alcohol can increase plasma concentrations of some long‐acting opioid analgesics (i.e., Embeda). This can potentially lead to a fatal overdose. Key References: Clinical Pharmacology [database online]. Tampa, FL: Gold Standard, Inc.: 2016. URL: http://www.clinicalpharmacology.com. Updated periodically. DailyMed. Package inserts. U.S. National Library of Medicine, National Institutes of Health website. http://dailymed.nlm.nih.gov/dailymed/about.cfm. DrugPoints® System (electronic version). Truven Health Analytics, Greenwood Village, CO. Updated periodically. Lexi-Comp ONLINE™ with AHFS™, Hudson, Ohio: Lexi-Comp, Inc.; 2016; Updated periodically. ©Institute for Clinical and Economic Review, 2017 Page 107 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Figure B2. Excerpt of Coverage Policy from Harvard Pilgrim Xtampza ER® (oxycodone extended release capsules) Clinical Information: Dose and Quantity Requested: Date Requested: Length of Treatment (please be specific): Documentation of Medical Necessity: (please check all that apply): 1. Please indicate the patient’s diagnosis: ________________________________________________________ 2. Does the patient have pain severe enough to require daily, around-the-clock long-term opioid treatment? Yes No 3. Has the patient tried and failed OxyContin (brand or generic)? Yes No If no, please provide clinical rationale why OxyContin cannot be used (e.g. ease of administration, difficulty swallowing, etc.) ________________________________________________________________________________________ ________________________________________________________________________________________ 4. For renewal requests: Has the patient experienced improvement while on therapy? Yes No 5. Please provide additional information pertinent to this request: ________________________________________________________________________________________ ________________________________________________________________________________________ ________________________________________________________________________________________ By signing this form, I attest (i) the information is true and accurate to the best of my knowledge and (ii) that the documentation supporting the information provided on this form is recorded in the patient’s medical records. Prescribing Clinician or Authorized Representative Signature: __________________________________________ Date:_______________ ©Institute for Clinical and Economic Review, 2017 Page 108 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Figure B3. Excerpt of Maine Medicaid Prior Authorization State of Maine Department of Health & Human Services MaineCare/MEDEL Prior Authorization Form OPIATE LIMITS PA. Prior authorization is not required for preferred medication for members in a nursing facility, hospice care and members receiving opioids for symptoms of Cancer or HIV/AIDS. Prior authorization will also not be required for members using 30mg or less MSE per day. Please refer to mainecarepdl.org for additional criteria including MSE conversion limitations. Dosage Days’ Supply Drug Name Strength Instructions Quantity Medical Necessity Documentation Required: (Attach copies of supporting office notes.) Why is this medication necessary for this member? (Please include members medical diagnosis) Acute Pain: Have you diagnosed this patient with acute pain? Yes No Has this patient already completed 15 days of opioid medication treatment for acute pain in the last 12 months? Yes No (Please note that if the patient has already received three refills beyond the first 15 days this PA will be denied.) If the PA is for a long acting narcotic, please explain why it is medically necessary to treat short-term acute pain? Chronic Pain: (non-acute only) Have you diagnosed this patient with long-term non-acute (Chronic Pain)? Yes No Have you and this patient established a Pain Management Plan consistent with MaineCare policy Section 80? Yes No Is the patient currently participating in one of the covered treatment options Yes No If yes which one? ©Institute for Clinical and Economic Review, 2017 Page 109 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents If no when is the first appointment? Is this PA intended to authorize opioid medications for treatment of headache, back pain, neck pain or fibromyalgia? Yes No If yes, please attach second opinion note recommending that opioids be used as part of a Pain Management Plan for this patient. If this PA request is for more than 300mg of morphine sulfate equivalent (MSE) per day please state the timeframe for tapering down to less than 300mg of morphine sulfate equivalent ©Institute for Clinical and Economic Review, 2017 Page 110 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix C. Previous Systematic Reviews and Technology Assessments We identified one systematic review evaluating the impact of abuse-deterrent formulation on abuse and other abuse related outcomes. California Health Benefit Review Program. Analysis of California Assembly Bill AB 623 Abuse-Deterrent Opioid Analgesics The California Health Benefit Review Program (CHBRP) assessed the medical and public health impacts of ADFs as part of a broader evaluation of a new bill to be enacted on the use of abuse- deterrent opioid analgesics in the state of California. The review examined the impact of ADFs on opioid abuse, including a possible shift of abuse to other prescription opioids, other routes of administration, or to illicit drugs (e.g., heroin). Although the introduction of ADFs was shown to reduce some forms of abuse of the reformulated drug (particularly those related to inhaling or injecting), some of the studies reviewed by CHBRP suggested that there was a shift to other routes of administration or abuse of other opioid analgesics and/or to illicit drugs (such as heroin) following the introduction of ADFs. The authors concluded that the impact of ADFs on abuse is ambiguous and further epidemiologic surveillance and study is required to ascertain its effectiveness. ©Institute for Clinical and Economic Review, 2017 Page 111 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix D. Ongoing Studies We did not identify any ongoing clinical trials registered on ClinicalTrials.gov that evaluated the abuse and/or abuse potential of either approved or investigational abuse-deterrent formulations of opioids. A review of publicly-available correspondence between FDA and drug manufacturers, however, provided some information on the postmarketing reporting required for all agents of focus. The FDA has required ADF manufacturers to conduct studies assessing whether the properties intended to deter the misuse and abuse of each ADF result “in a meaningful decrease in misuse and abuse, and their consequences, addiction, overdose, and death, in post-approval settings.”180 In addition, the FDA has required several manufacturers to submit nationally representative descriptive studies analyzing ADF utilization (and that of select comparators) as well as the scope and patterns of abuse in diverse populations. The proposed final report submission dates for such studies are summarized in Table D1. Table D1. Final Report Submission Dates for Required Postmarket Reporting ADF* Descriptive study of utilization and Study to evaluate impact on misuse/abuse, if abuse any, attributable to the abuse-deterrent properties Xtampza ER 06/2018 06/2021 Troxyca ER 10/2018 10/2018 Targiniq ER N/A 01/2020 Hysingla ER N/A 04/2020 Vantrela ER 03/2019 03/2022 Embeda N/A 04/2020 Morphabond N/A 02/2021 Arymo ER 03/2019 03/2022 *Detailed reporting requirements for OxyContin were not identified ©Institute for Clinical and Economic Review, 2017 Page 112 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix E. Comparative Clinical Effectiveness Supplemental Information Methods: Supplemental Information Screening for Study Inclusion Subsequent to literature search and removal of duplicate citations using both online and local software tools, study selection was accomplished through two levels of screening, at the abstract and full-text level. Three reviewers screened the titles and abstracts of all publications identified; all three reviewers worked together to resolve any issues of disagreement through consensus. No study was excluded at abstract-level screening due to insufficient information. Citations accepted during abstract-level screening were retrieved in full text for review. Reasons for exclusion were categorized according to the PICOTS (Population, Interventions, Comparisons, Outcomes, Timing, and Setting) elements during both title/abstract and full-text review. We used criteria published by the US Preventive Services Task Force (USPSTF) to assess the quality of pre-market RCTs, using the categories “good,” “fair,” or “poor”.160 Guidance for quality ratings using these criteria is presented below, as is a description of any modifications we made to these ratings specific to the purposes of this review. Good: Meets all criteria: Comparable groups are assembled initially and maintained throughout the study; reliable and valid measurement instruments are used and applied equally to the groups; interventions are spelled out clearly; all important outcomes are considered; and appropriate attention is paid to confounders in analysis. In addition, intention to treat analysis is used for RCTs. Fair: Studies were graded "fair" if any or all of the following problems occur, without the fatal flaws noted in the "poor" category below: Generally comparable groups are assembled initially but some question remains whether some (although not major) differences occurred with follow-up; measurement instruments are acceptable (although not the best) and generally applied equally; some but not all important outcomes are considered; and some but not all potential confounders are addressed. Intention to treat analysis is done for RCTs. Poor: Studies were graded "poor" if any of the following fatal flaws exists: Groups assembled initially are not close to being comparable or maintained throughout the study; unreliable or invalid measurement instruments are used or not applied equally among groups (including not masking ©Institute for Clinical and Economic Review, 2017 Page 113 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents outcome assessment); and key confounders are given little or no attention. For RCTs, intention to treat analysis is lacking. We used the National Institutes of Health (NIH) “Quality Assessment Tool for pre-post Studies with no Control group” presented below as guidance criteria to assess the quality of the postmarket studies.161 Table E1. Criteria for Assessing Pre-Post Studies with No Control Group Criteria for assessing pre-post studies with no control group 1. Was the study question or objective clearly stated? 2. Were eligibility/selection criteria for the study population prespecified and clearly described? 3. Were the participants in the study representative of those who would be eligible for the test/service/intervention in the general or clinical population of interest? 4. Were all eligible participants that met the prespecified entry criteria enrolled? 5. Was the sample size sufficiently large to provide confidence in the findings? 6. Was the test/service/intervention clearly described and delivered consistently across the study population? 7. Were the outcome measures prespecified, clearly defined, valid, reliable, and assessed consistently across all study participants? 8. Were the people assessing the outcomes blinded to the participants' exposures/interventions? 9. Was the loss to follow-up after baseline 20% or less? Were those lost to follow-up accounted for in the analysis? 10. Did the statistical methods examine changes in outcome measures from before to after the intervention? Were statistical tests done that provided p values for the pre-to-post changes? 11. Were outcome measures of interest taken multiple times before the intervention and multiple times after the intervention (i.e., did they use an interrupted time-series design)? 12. If the intervention was conducted at a group level (e.g., a whole hospital, a community, etc.) did the statistical analysis take into account the use of individual-level data to determine effects at the group level? Using these criteria, in general terms, a "good" study has the least risk of bias, and results are considered to be valid. A "fair" study is susceptible to some bias deemed not sufficient to invalidate its results. The fair quality category was broad, so studies with this rating will vary in their strengths and weaknesses and a "poor" rating indicates significant risk of bias. ©Institute for Clinical and Economic Review, 2017 Page 114 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix F. Evidence Summary Tables Table F1. Pre-Market Studies Author & Year of Study design Intervention and dosing Inclusion and Patient Outcomes Publication (Duration of schedule Exclusion Criteria characteristics (Trial name) Quality follow up) rating OxyContin (oxycodone extended release) Harris J of Clin Pharm RCT double-blind 1) Finely crushed OxyContin Healthy adults 18-55 Mean age, yrs (SD): Overall drug liking VAS, Emax Mean 2014 49 positive- and reformulated (ORF-F) years old, with a 32.1 (8.99) (SD) placebo- 2) Coarsely crushed OxyContin history of nonmedical 1) 69.7 (29.4) Fair Quality controlled five- reformulated (ORF-C) use of opioids via Male: 86.7% 2) 61.1 (25.8) treatment 3) Original formulation intranasal route 3) 87.4 (22.2) crossover study OxyContin (OC) White: 86.7% 4) 84.8 (18.9) 4) Oxycodone powder (Oxy API) Exclusion: Objective 5) 48.9 (14.8) 5) OC Placebo Opiate withdrawal BMI, range: 19- scale (OOWS)≥3 29.7kg/m2 Take drug again VAS, Emax Mean N= 30 following naloxone (SD) challenge test, self- Recreational use of 1) 64.0 (38.2) Subjects self-administered reported drug other psychoactive 2) 52.8 (37.4) intranasal doses of 30mg doses dependence (past 2 drugs 3) 89.6 (20.7) of placebo, ORF-F, ORF-C, finely years), or a positive 1) Cannabinoids- 4) 86.6 (23.5) crushed OC, and Oxy API in a urine drug screen or 86.7% 5) 28.2 (24.3) randomized crossover fashion, breath alcohol test 2) Stimulants- with a washout period of at least 53.3% Subjective drug value, Emax Mean 48 hours between treatments 3) Depressants- (SD) 30% 1) $17.01 ($16.39) 2) $17.25 ($17.93) ©Institute for Clinical and Economic Review, 2017 Page 115 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Prior to the treatment phase 4) Dissociative 3) $27.95 ($16.03) subjects go through: anesthetics- 30% 4) $27.30 ($17.40) 1) A screening phase -This 5) Hallucinogens- 5) $0.37 ($0.60) includes a Naloxone challenge 23.3% test to determine physical *All p values for 1, 3 & 4 ≤0.003 vs. dependence placebo except for 2 which did not 2) A qualification phase: Use of differ from placebo on drug liking 30mg Oxy API & lactose powder * All p values for 1 & 2 ≤0.002 vs. 3 & in a randomized crossover, with 4 24hrs interval; subjects enter * 1 & 2 did not differ significantly double-blind treatment phase if from each other except in drug liking 30mg Oxy API is tolerated where p value= 0.043 Xtampza® ER (oxycodone extended release) Kopecky, 2016 43 RCT, double- 1) DETERx Intact (HFHC) Healthy adult Mean age, yrs Drug liking VAS, Emax Mean (SD) blind, active- and 2) DETERx Chewed (HFHC) nondependent (range): 26.2 (18- 1) 68.6 (13.1) Fair Quality placebo- 3) DETERx Intact (Fasted) recreational 46) 2) 70.8 (11.5) controlled, triple- 4) DETERx Chewed (Fasted) opioid users aged 18- 3) 68.8 (13.0) dummy, single- 5) Crushed IR Oxycodone 55 years who have Male: 66% 4) 73.4 (13.9) dose, 6-way 6) Placebo previously taken and 5) 81.8 (11.5) crossover, tolerated 40-mg of White: 87% 6) 54.9 (8.4) hypothesis- N=38 (completer population) oxycodone driven study hydrochloride All subjects Overall drug liking VAS, Emax Mean HFHC=Taken after a high-fat, reported a history (SD) high-calorie meal Exclusion: lifetime of recreational 1) 68.5 (16.5)* Fasted=Taken after overnight history of drug or opioid use 2) 69.8 (17.4)α fast of at least 10 hours alcohol dependence, (hydrocodone, 3) 69.4 (15.3)α heavy use of tobacco oxycodone, 4) 74.2 (14.4) Prior to the treatment phase products morphine, 5) 76.2 (16.4) subjects go through: buprenorphine, 6) 54.4 (10.1)* codeine, ©Institute for Clinical and Economic Review, 2017 Page 116 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 1) A screening phase -This oxymorphone, or Take drug again VAS, Emax Mean includes a Naloxone challenge heroin) (SD) test to determine physical 1) 70.6 (18.1) dependence 2) 69.3 (18.9) 2) A qualification phase: subjects 3) 70.2 (16.0) received oral IR oxycodone 4) 73.7 (14.9) 20mg or placebo in a double- 5) 75.4 (16.8) blind, crossover with 24-hr 6) 52.7 (13.4) washout between. *p<0.0001 vs. IR oxycodone Intact study drug (oxycodone α p<0.05 vs. IR oxycodone DETERx or placebo) was administered first with 50 mL of Mean ARCI Score (SD) solution (IR oxycodone or 1) 4.1 (4.8)* placebo) followed by chewed 2) 4.0 (4.3)* study drug. 3) 4.3 (5.0)* 4) 5.3 (5.0)ε 5) 7.1 (5.6) 6) 1.4 (2.7)α *p<0.01 vs IR oxycodone ε p<0.05 vs IR oxycodone α p<0.0001 vs IR oxycodone Webster L Pain RCT, double- 1) Crushed DETERx 40 mg IN + Inclusion: N=39 Overall drug liking (mm) Medicine 2016 50 blind, double- Intact PBO-ER PO Men or nonpregnant, 1) 48.42 dummy, positive- 2) Crushed PBO-ER IN + Intact nonlactating women; Mean age, yrs 2) 62.20 Fair Quality and placebo- DETERx 40 mg PO aged 18 to 55 yrs; (SEM): 3) 71.78 controlled, single- 3) Crushed OXY-IR 40 mg IN recreational opioid 26.77 (1.07) dose, four-phase, (active control) + Intact PBO-ER users (use of opioids Take drug again (mm) four-treatment, PO for nonmedical Male, N (%): 1) 47.77 crossover study purposes on ≥10 28 (71.8) 2) 58.98 ©Institute for Clinical and Economic Review, 2017 Page 117 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 4) Crushed PBO-ER IN (PBO occasions during the 3) 71.25 control) + Intact PBO-ER PO past year and ≥once White, N (%): in the 12 wks prior to 33 (84.6) ACRI Scores IN=intranasal; PO=oral; screening); required 1) 1.34 ER=extended release, to have a history of IN Mean weight, kg 2) 3.10 IR=immediate release opioid use ≥3 times (SEM): 3) 5.93 OXY=oxycodone powder within past year 77.35 (2.81) Exclusion: Mean height, cm 1) A screening phase -This Physical dependence (SEM): includes a Naloxone challenge or tolerance to 174.75 (1.32) test to determine physical opioids, alcohol, or dependence other drugs Mean BMI, kg/m^2 2) A drug discrimination test: (excepting caffeine (SEM): each subject received either a and nicotine); positive 25.27 (0.81) single IN dose of crushed OXY-IR urine drug screen 20 mg or a single IN dose of (excluding THC) and crushed PBO-IR and was later alcohol breath test; crossed-over to the other significant unstable treatment after 24 hours wash medical condition or out; subjects were excluded if chronic disease; they could not discriminate positive for infectious between OXY-IR opioid and PBO disease; contraindication to opioid; heavy smokers unable to abstain from smoking for ≥5 hours during day Troxyca® ER (oxycodone hydrochloride and naltrexone hydrochloride) ©Institute for Clinical and Economic Review, 2017 Page 118 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Setnik, 201551 Randomized, 1) Placebo sugar sphere Healthy adults aged Mean age, yrs (SD): Drug liking VAS, Emax Mean (95% CI) double-blind, (crushed) 18–55 years with 35.1 (8.4) 1) 51.0 (47.7, 54.3) Fair Quality placebo- and 2) ALO-02 30 mg/3.6 mg body weight ≥50 kg 2) 60.5 (57.2, 63.8)*,ǂ active-controlled, (crushed) and BMI 17.5–30.5 Male: 86% 3) 51.3 (48.0, 54.6) 4-way crossover 3) Placebo lactose tablet kg/m2 who were 4) 92.8 (89.5, 96.1)ǂ study (crushed) nondependent White: 96% 4) Oxycodone IR 30 mg recreational opioid Take drug again VAS, Emax Mean (crushed) users; intranasal use Mean weight (SD): (95% CI) of opioids ≥3 times 78.6 (11.3) kg 1) 48.2 (39.4, 57.0) N=28 (completer population) within year 2) 58.9 (50.1, 67.8)α 4 treatment periods separated before screening visit Recreational drug 3) 46.9 (38.1, 55.8) by ≥5 days. use in last 12 4) 88.4 (79.6, 97.2)β Exclusion: substance months: Patients underwent naloxone use disorder and/or 1) oxycodone- Overall drug liking VAS, Emax Mean challenge to determine signs of dependence; heavy 46.4% (95% CI) withdrawal (COWS method) use of tobacco; 2) OxyContin- 1) 50.6 (44.4, 56.8) followed by 0.6mg. This was positive urine drug 46.4% 2) 60.2 (54.0, 66.4)α,β followed by drug discrimination screen (excluding 3) Percocet-35.7% 3) 51.6 (45.3, 57.8) phase: participants randomly tetrahydrocannabinol) 4) Cannabinoids- 4) 85.4 (79.1, 91.6)β received in double-blind manner 85.7% either crushed oxycodone IR 5) Alcohol-82.1% *p≤0.0001 versus oxycodone IR 30 30mg or crushed placebo lactose 6) Stimulants- mg. tablets intranasally for 2 60.7% ǂp≤0.01 versus corresponding consecutive days. Patients placebo. excluded if they can’t distinguish α p≤0.01 versus oxycodone IR 30 mg. placebo from oxycodone. β p≤0.05 versus corresponding placebo. Setnik Pain Medicine Randomized, 1) ALO-02, 40 mg (crushed) Inclusion: N=41 Drug liking, Emax VAS scores 2016 44 double-blind, 2) ALO-02, 60 mg (intact) Healthy, nondependent 1) 70.2 placebo-/active- 3) ALO-02, 60 mg (crushed) recreational opioid White: 78 percent 2) 59.3 ©Institute for Clinical and Economic Review, 2017 Page 119 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Fair Quality controlled, 6-way 4) Oxycodone IR, 40 mg users (user of opioids 3) 74.5 crossover study, (crushed) for nontherapeutic Mean age, yrs 4) 85.5 with naloxone 5) Oxycodone IR, 60 mg purposes on ≥10 (SD): 37.8 (9.3) 5) 89.8 challenge, drug (crushed) occasions within the 6) 51.6 discrimination, 6) Placebo previous year and ≥8 Mean body p≤0.05, drug vs placebo group and treatment weeks before the weight, kg (SD): phases -Screening: standard medical screening visit); aged 78.1 (9.4) Drug high, Emax (VAS) evaluation 18 to 55 years; a BMI 1) 46.5 -Naloxone challenge: received IV between 17.5-30.5 Mean BMI, 2) 22.5 naloxone (0.2 mg followed by an kg/m^2 kg/m^2 (SD): 3) 52.8 additional 0.6 mg if no signs of 25.6 (2.3) 4) 78.6 withdrawal were observed Exclusion: 5) 85.7 within the first 30 seconds); Diagnosis of substance Common opioids 6) 10.2 withdrawal was assessed using and/or alcohol used in previous p≤0.05, drug vs placebo group the Clinical Opiate Withdrawal dependence or 12 months: Scale (COWS); a score of <5 on treatment for -Oxycodone: 50% Take drug again, Emax (VAS) COWS were eligible substance and/or -OxyContin: 1) 58.1 -Drug discrimination phase: alcohol-related 31.3% 2) 48.7 participants had to distinguish disorders; positive -Percocet: 18.8% 3) 72.5 between orally administered urine drug screen or 4) 83.7 crushed Oxycodone HCL IR 40 alcohol breath test; any 5) 81.5 mg and placebo; this was condition where an 6) 46.1 defined as ≥15-point peak opioid is p≤0.05, drug vs placebo group increase on the drug liking and contraindicated; take drug again visual analog evidence or history of Overall drug liking, Emax scale, and ≥30-point peak clinically significant 1) 64.4 increase on the high VAS within disease; history of 2) 53.3 2 hours unresolved sleep apnea 3) 74.3 in last 5 years; other 4) 80.9 severe acute or chronic 5) 81.8 ©Institute for Clinical and Economic Review, 2017 Page 120 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents medical or psychiatric 6) 51.1 condition or laboratory p≤0.05, drug vs placebo group abnormality Hysingla® ER (hydrocodone bitartrate) Harris, 2016 52 Single-center, 1) PBO Healthy, moderately Mean age, yrs (SD): Drug liking VAS, Emax Mean (SD) double-blind, 2) Hysingla (HYD) coarse experienced opioid 38.9 (10.21) 1) 50.6 (0.5) Fair Quality positive- and particles 60 mg users aged 18-55 2) 65.4 (18.4)* placebo- 3) HYD fine particles 60 mg years with a history of Male: 90.3% 3) 66.8 (18.4)* controlled, 4) Hydrocodone powder 60 mg intranasal opioid 4) 90.4 (13.2) randomized, abuse; ≥10 use White: 64.5% four-treatment N=25 (completer population) opioids in previous Overall drug liking VAS, Emax Mean crossover study year, ≥3 use 12 wks Mean BMI (SD): (SD) at 12/24 hrs Treatment administrations were prior to screen, ≥3 25.3 (2.42) kg/m2 1) 50.2 (0.5) / 50.0 (0.4) separated by a washout period intranasal use in past 2) 60.8 (16.4) / 52.8 (21.4) of five to seven days year; BMI 18.0 - 29.9 Recreational drug 3) 59.4 (24.4) / 55.8 (22.5) kg/m2 and a weight of experience: 4) 88.2 (13.4) / 83.4 (19.0) Treatment periods followed ≥50 kg; dose of opioid 1) Cannabinoids- 1) A naloxone challenge: a equivalent to ≥40 mg 77.4% Take drug again VAS, Emax Mean subset of subjects completed a hydrocodone by any 2) Stimulants-71% (SD) at 12/24 hrs dose selection phase to identify route of 3) Hallucinogens- 1) 0.0 (0) / 2.0 (10.0) appropriate intranasal dose of administration at 38.7% 2) 33.1 (40.2) / 29.5 (38.3) hydrocodone powder. 2) A least once in the past 4) Depressants- 3) 36.4 (38.1) / 34.6 (36.4) Qualification phase: eligible year 16.1% 4) 84.7 (25.1) / 83.9 (25.6) subjects self-administered 60mg 5) Dissociative of hydrocodone powder and Exclusion: heavy use anesthetics-12.9% Subjective drug value, Emax Mean $ placebo powder intranasally in of tobacco; history of (SD) at 12/24 hrs double-blind crossover design, drug/alcohol 1) 0.25 (0.0) / 0.25 (0.0) with 24-hr washout between. dependence; past or 2) 7.8 (13.2) / 7.4 (13.0) Subjects had to distinguish planned abdominal 3) 12.0 (15.7) / 11.0 (15.8) ©Institute for Clinical and Economic Review, 2017 Page 121 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents between them on “at this surgery; history of 4) 27.7 (14.9) / 27.8 (14.4) moment” drug liking VAS and asthma or airway overall liking/feeling high VAS disease; history of *p<0.001 vs. hydrocodone powder hypotension Harris, 2016b 39 single-center, 1) Placebo Healthy, moderately Mean age, yrs (SD): Drug liking VAS, Emax Mean (SD) double-blind, 2) HYD intact 60 mg experienced 36.3 (9.2) 1) 52.3 (7.14) Fair Quality positive- and 3) HYD chewed 60 mg recreational opioid 2) 63.3 (16.0)* placebo- 4) HYD fine particles users age 18 – 55 Male: 82.5% 3) 69.0 (17.5)* controlled, 5) Hydrocodone solution 60 mg years, weight ≥ 50 kg 4) 89.2 (14.0)ǂ randomized, five- and BMI 18.0 - 29.9 White: 72.5% 5) 94.0 (10.2) treatment N=35 (completer population) kg/m2; chewed an crossover study opioid ≥3 times for Mean BMI (SD): Overall drug liking VAS, Emax Mean The following treatments were recreational oral 25.2 (3.02) kg/m2 (SD) at 12/24 hrs administered orally: 1) HYD 60 abuse/misuse during 1) 48.2 (13.1) / 48.1 (13.0) mg tablet intact; 2) HYD 60 mg previous 12 months; Recreational drug 2) 53.3 (16.8) / 54.9 (22.2) tablet chewed (2-3 minutes); 3) used 60 mg experience: 3) 57.6 (28.3) / 56.8 (28.1) HYD 60mg fine particles; 4) hydrocodone 1) Cannabinoids- 4) 83.7 (18.0) / 80.1 (22.4) hydrocodone solution 60 mg; 5) equivalent or higher 87.5% 5) 83.0 (19.2) / 84.1 (19.7) placebo solution. At each opioid dose at least 2) Stimulants- treatment visit, subjects once during lifetime; 77.5% Take drug again VAS, Emax Mean received an intact tablet, milled negative urine screen 3) Hallucinogens- (SD) at 12/24 hrs tablet, chewed tablet, and oral (except cannabinoids 32.5% 1) 3.9 (15.9) / 2.2 (12.8) solution. All treatments were and benzodiazepines) 4) Depressants- 2) 19.5 (33.7) / 32.6 (35.5) separated by a washout period 32.5% 3) 41.3 (40.7) / 43.0 (41.2) of five to seven days. Exclusion: heavy use 5) Dissociative 4) 82.6 (29.7) / 77.0 (31.5) of tobacco; history of anesthetics-20.0% 5) 84.6 (25.7) / 86.7 (22.8) Patients underwent naloxone drug/alcohol challenge. Qualification phase: dependence, dental Subjective drug value, Emax Mean $ oral solution of hydrocodone work or clinically (SD) at 12/24 hrs 60mg and matching placebo relevant dental issues 1) 0.5 (1.4) / 0.5 (1.6) ©Institute for Clinical and Economic Review, 2017 Page 122 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents solution in double-blind 2) 6.8 (14.6) / 8.8 (14.5) crossover fashion, separated by 3) 11.4 (14.8) / 13.7 (16.5) 24-hr washout period. Subjects 4) 24.2 (17.0) / 25.9 (16.5) required to distinguish 5) 22.9 (17.1) / 25.8 (16.8) treatment and placebo on “at this moment” drug-liking, overall *p<0.001 vs. hydrocodone solution drug liking, high VAS. ǂp-0.015 vs. hydrocodone solution Vantrela® ER (hydrocodone bitartrate) Bond Drg and alc dep Randomized, 1) Placebo (n=42) Inclusion: Overall drug liking VAS, Emax Mean 2015 46 double-blind, 2) Hydrocodone IR (n=39) Age 18 to 50 years; (SD) triple-dummy, 3) Hydrocodone ER intact (n=41) BMI between 18 and 1) 51.1 (7.6) Conference Abstract placebo- 4) Hydrocodone ER crushed 32 kg/m^2; history of 2) 75.0 (16.8) controlled, (n=42) recreational opioid 3) 49.2 (11.0) crossover study use ≥10 times in last 4) 59.0 (19.9) consisted of 3 Split into three phases: year and ≥1 within 12 p≤0.0022 in comparison with phases Phase A – screening wks; not physically hydrocodone IR Phase B – qualification phase: dependent on Returned for a randomly assigned in double- opioids; negative Take drug again VAS, Emax Mean follow-up visit blind crossover fashion 60mL of urine drug screening (SD) ~48 to 72 hours noncarbonated flavored and alcohol breath 1) 47.2 (15.5) after discharge beverage (placebo) and 45mg of test (except THC); 2) 75.2 (17.3) form the study hydrocodone IR bitartrate women must be 3) 46.4 (18.3) center powder reconstituted in 60mL of surgically sterile, 2 4) 58.7 (21.5) a noncarbonated flavored years p≤0.0022 in comparison with beverage with 48-hr washout postmenopausal, or hydrocodone IR between. Subjects had to using a medically distinguish between placebo and acceptable Drug liking VAS, Emax Mean hydrocodone on a drug-liking contraceptive 1) 53.2 and overall drug-liking bipolar 2) 85.2 VAS. Exclusion: 3) 53.9 ©Institute for Clinical and Economic Review, 2017 Page 123 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Phase C – treatment phase, Any clinically 4) 66.9 subjects received, in random significant p≤0.0022 in comparison with sequence, separated by a ≥14 uncontrolled medical hydrocodone IR day washout condition or abnormalities; history of drug or alcohol abuse; history of hypersensitivity or idiosyncratic reaction to hydrocodone or hydromorphone Bond 2015b 41 Randomized, 1) IN oral hydrocodone ER, 45 Inclusion: Drug Liking VAS, Emax placebo- mg Participants able to 1) 57.3 Conference Abstract controlled, 2) IN hydrocodone API, 45 mg tolerate a 45 mg 2) 80.2 double-blind, 5- 3) intact oral hydrocodone ER, intranasal dose of 3) 72.8 (p=0.004 vs 2 and 3); period crossover 45 mg hydrocodone API (p<0.001 vs 1 and 5) 4) IN manipulated Zohydro API, powder; be able to 4) 83.2 45 mg discriminate effects of 5) 58.6 5) Placebo hydrocodone from placebo Overall Drug Liking VAS, Emax N=45 1) 57.8 IN=intranasal; API=active 2) 77.1 pharmaceutical ingredient 3) 68.5 (p=0.004 vs 2 and 3); (p≤0.001 vs 1 and 5) 34 participants were evaluable 4) 79.8 for pharmacodynamic 5) 57.7 assessments performed through 48 hours after administration of study drug Embeda® (morphine sulfate and naltrexone hydrochloride) ©Institute for Clinical and Economic Review, 2017 Page 124 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Johnson F Pain RCT, 4-way 1) Crushed pellets from 2 ER Inclusion: Drug liking VAS, Emax Mean (SD) Physician 2011 181 crossover, morphine sulfate with a Aged 18 to 55 years; 1) 68.1 (17.5) double-blind, sequestered naltrexone core nondependent 2) 67.6 (13.1) Fair Quality triple-dummy, (MS-sNT) 60 mg capsules healthy opioid users; 3) 89.5 (12.6) placebo- 2) Two intact MS-sNT 60 mg had experience in 4) 52.2 (4.5) controlled capsules non-therapeutic use 3) Morphine sulfate solution of opioids on at least Cole/ARCI Stimulation-Euphoria, 12 weeks (MSS), 120 mg 10 occasions within Emax (SD) 4) Placebo last year and at least 1) 10.8 (11.2) once in the 12 wks 2) 11.9 (11.3) N=32 prior to screening 3) 18.4 (11.6) 4) 6.90 (8.2) Participants received 4 Exclusion: treatments, one per session, History or presence of with each session separated by clinically significant washout period of 14 to 21 days disease; history of allergic or adverse response to the study drugs or related drugs; used an over the counter medication within seven days prior to first dose of study medication Setnik Pain Res & Randomized, 1) Placebo, 100 mg tablets Inclusion: Mean age, yrs (SD): Drug liking VAS, Emax Mean Man 2013 53 double-blinded, 2) Crushed EMBEDA – morphine Aged 18 to 55 years; 35.2 (10.01) 1) 50.9 placebo- sulfate/naltrexone healthy 2) 69.6 Fair Quality controlled, hydrochloride, 30 mg/1.2 mg ER nondependent Male: 85 percent 3) 87.6 single-dose, recreational opioid ©Institute for Clinical and Economic Review, 2017 Page 125 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents three-way 3) Crushed morphine sulfate user; must have White: 85 percent p<0.001 for crushed EMBEDA and crossover study crushed (CR), 30 mg tablet experience with crushed morphine sulfate CR vs. intranasal drug Mean weight, kg placebo 16 weeks N=33 administration (≥3 (SD): 79.18 (8.86) occasions within last Overall drug liking VAS, Emax Mean Eligible participants underwent a year) Mean BMI, kg/m^2 1) 50.9 naloxone challenge test, an (SD): 25.62 (2.75) 2) 60.8 intravenous 0.2 mg naloxone Exclusion: 3) 83.8 HCL bolus followed by an Diagnosis of p<0.001 for crushed EMBEDA and assessment for signs of opioid substance and/or crushed morphine sulfate CR vs. withdrawal. If no signs within 30 alcohol dependence; placebo seconds, an additional naloxone participated or 0.6 mg bolus dose was seeking treatment for Take drug again VAS, Emax Mean administered and observed for 5 substance use 1) 42.2 mins. This was followed by dose disorder; has a 2) 60.6 selection phase: crushed condition in which an 3) 84.9 morphine sulfate (MS) (30mg) opioid is p<0.001 for crushed EMBEDA and and placebo administered contraindicated; crushed morphine sulfate CR vs. intranasally and double-blinded allergy or history of placebo crossover fashion to the first hypersensitivity to cohort of four eligible opioids participants; dose escalated to 60mg and 90mg in up to two cohorts. Dose determination based on drug liking bipolar VAS. Drug discrimination phase followed with MS and placebo intranasally, ability to discriminate on drug liking VAS and unipolar high VAS. ©Institute for Clinical and Economic Review, 2017 Page 126 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Setnik Pain Medicine Randomized, 1) Placebo (single-dose, 2 x Inclusion: Mean age, yrs (SD): Drug liking VAS, Emax Mean (SD) 2013 47 single center, microcrystalline cellulose) Healthy 24.2 (3.7) 1) 51.7 double-blind, 2) MS Contin (single-dose, 2 x 60 nondependent 2) 80.8 Fair Quality placebo- mg morphine sulfate whole recreational opioid Males: 91 percent 3) 65.2 controlled, three- tablets manually crushed) user way crossover 3) EMBEDA (single-dose, White: 97 percent Overall drug liking VAS, LS Mean study solution 2 x 60 mg morphine Exclusion: 1) 50.5 sulfate with sequestered 2.4 Has a history or Mean body weight, 2) 69.8 naltrexone hydrochloride whole current diagnosis of lb. (SD): 170.2 3) 58.6 capsules manually crushed) substance (29.9) dependence Take drug again VAS, LS Mean N=33 (excluding caffeine Mean BMI, kg/m^2 1) 49.5 and nicotine); seeking (SD): 23.9 (3.5) 2) 70.7 A naloxone challenge test treatment for 3) 57.7 consisted of an IV bolus dose of substance and/or naloxone hydrochloride, 0.2 mg alcohol related and if there was no evidence of disorders; history or withdrawal within 30 seconds presence of clinically and additional 0.6 mg bolus significant illness; dose was injected. females who are pregnant, lactating, or In the drug discrimination phase, planning to become participants received either 120 pregnant during the mg of morphine sulfate or study; allergy or placebo in solution (150ml). history of Patient eligibility was based on hypersensitivity to the ability of the subject to opioids distinguish morphine from placebo ©Institute for Clinical and Economic Review, 2017 Page 127 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Treatment phase occurred from days 3-21 and comprised three visits, each with a 2-night confined stay; each treatment period was separated by a minimum of 4 days not to exceed 14 days between dosing Stauffer Clin Drug Randomized, 1) Placebo, 120 mg Inclusion: Mean age, yrs (SD): Drug liking VAS, Emax Mean (SD) Inves 2009 48 double-blind, 2) ALO-01 whole, 2 x 60 mg Aged 18 to 55 yrs; 35.0 (7.6) 1) 52.2 (4.5) placebo- capsules healthy with a BMI of 2) 67.6 (13.1) Fair Quality controlled, triple- 3) ALO-01 crushed, 2 x 60 mg 21-31 kg/m^2 and Male: 81 percent 3) 68.1 (17.5) dummy, four-way capsules weight <55 kg; 4) 89.5 (12.6) crossover study 4) Morphine sulfate solution nondependent opioid White: 69 percent (MSS), 120 mg users; previously used Subjective drug value, Emax Mean opioids non- Bodyweight, kg ($Can) N=32 therapeutically for (SD): 1) 2.73 (7.08) psychoactive effects 82.4 (11.0) 2) 14.22 (15.46) Subjects were screened for on ≥10 occasions 3) 13.72 (16.98) eligibility through whether they within previous yr and BMI, kg/m^2 (SD): 4) 28.85 (14.55) could tolerate a single dose of ≥1 in last 12 wks; a 26.4 (2.8) morphine 120 mg and positive drug test at ARCI, 0-51 scale (SD) distinguish between morphine screening was allowed 1) 9.4 (9.8) and placebo if it was negative at 2) 13.4 (12.5) qualifying session and 3) 15.7 (13.5) all treatment 4) 23.0 (12.8) sessions; women must have negative p<0.001, all comparisons except ALO-01 pregnancy test and whole vs ALO-01 crushed (p=NS), ALO-01 not lactating crushed vs placebo (p=0.002) and ALO- 01 whole vs placebo (p=NS) ©Institute for Clinical and Economic Review, 2017 Page 128 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Exclusions: History of substance use disorder (alcohol included); opioid addiction or dependence; current psychiatric illness or significant medial or neurological conditions; positive for HIV/hepatitis B&C Morphabond® ER (morphine sulfate) Webster Pain med Randomized, 1) Placebo Inclusion: Mean age, yrs (SD): Drug liking VAS, Emax LS Mean (SE) 54 2016 double-blind, 2) crushed intranasal ER Aged 18 to 55 yrs; 25.4 (6.57) 2) 84.79 double-dummy, morphine (60 mg) + intact oral nondependent opioid 3) 71.13 (p<0.0001 vs. 2 & 4) Fair Quality placebo- placebo users who used for Male: 85.2 percent 4) 67.03 controlled, four- 3) crushed intranasal Morphine nontherapeutic Overall drug liking VAS, p value way crossover ARER (60 mg) + intact oral purposed ≥10 White: 96.3 (vs. crushed intranasal ER morphine) placebo occasions within the percent 3) 0.007 4) crushed intranasal placebo + last year and at least 4) 0.0025 intact oral Morphine ARER (60 once in 12 wks prior Mean BMI, kg/m^2 mg) to screening; must (SD): Take drug again VAS, p value have ≥3 experiences 24.9 (3.89) (vs. crushed intranasal ER morphine) Each treatment is separated by a with insufflating drugs 3) 0.0341 minimum seven-day washout within last year 4) 0.0103 period Exclusion: ARCI N=25 Participated in, were (Intact oral morphine ARER vs participating in, or crushed intranasal ER morphine): seeking treatment for 0.0003 ©Institute for Clinical and Economic Review, 2017 Page 129 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Qualification phase consisted of substance use a three-night inpatient, double- disorders; presence of blind session; the naloxone drug or alcohol challenge test was an initial dose dependence; except of 0.2 mg of naloxone THC, patients were hydrochloride through excluded if positive intravenous bolus. If no urine drug screen; evidence of withdrawal occurred history or presence of within 30 seconds, 0.6 mg of clinically significant naloxone hydrochloride was disease; any condition given and the subject was in which an opioid observed for 5 mins; for the was contraindicated drug discrimination test, subjects received a single, intranasal dose each of morphine sulfate IR and placebo and subjects had to distinguish the morphine from the placebo. Arymo® ER (morphine sulfate) Smith M Pain Randomized, 1) Placebo Inclusion: Male, n (%): Drug liking VAS, Emax Median (SD) Medicine 2016 40 double-blind, 2) Morphine-ADER-IMT (60 mg, Aged 18 to 55 yrs; 28 (73.7) 1) 50 triple-dummy, intact) [abuse-deterrent experienced, 2) 62 Fair Quality active- and extended release, injection nondependent, White, n (%): 3) 67 (p=0.007) placebo- molded tablet] recreational opioid 35 (92.1) 4) 74 controlled, four- 3) Morphine-ADER-IMT (60 mg, users; recreational way crossover, manipulated) user has a history of Mean age, yr (SD): Overall drug liking VAS, Emax Mean single-center 4) Morphine ER (60 mg, nonmedical use of 24.3 (4.2) (SD) manipulated) opioids with ≥10 2) 57.0 (p<0.001) occasions within the Mean weight, lb 3) 63.5 (p=0.13) past year and ≥1 in (SD): 159.9 (27.2) 4) 67.5 ©Institute for Clinical and Economic Review, 2017 Page 130 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Everyone received 1 dose of 12 wks before each oral agent in crossover screening Mean BMI, kg/m^2 Take drug again VAS, Emax Mean fashion separated by ≥5 days (SD): (SD) Exclusion: 24.3 (3.9) 2) 56.0 (p<0.001) The qualification phase History of substance 3) 61.5 (p=0.05) consisted of a naloxone and/or alcohol 4) 68.0 challenge to exclude participants dependence; any who were opioid dependent, condition in which and a drug discrimination test to opioids are exclude participants who could contraindicated; not tolerate 30 mg morphine or presence of hepatitis distinguish its positive subjective B/C or HIV; history of effects from placebo sleep apnea in the past 5 yrs that hasn’t been corrected or resolved Webster L Pain Single-center, 1) Intranasal low volume (IN LV) Inclusion: Male, N (%): Drug liking VAS, Emax Median (SD) Medicine 2016 55 randomized, manipulated morphine ER, 60 Aged 18 to 55 yrs; 36 (78.3) 1) 77.5 double-blind, mg (n=46) experienced 2) 62.0 (p<0.0001) Fair Quality double-dummy, 2) IN manipulated high volume nondependent White, N (%): 3) 52.5 (p<0.0001) active- and (HV) morphine abuse-deterrent, recreational opioid 44 (95.7) 4) 68.0 (p=0.0001) placebo- injection molded tablets (ADER- user with experience 5) 51.0 (p<0.0001) controlled five- IMT), 60 mg (n=46) of IN opioid Mean age, yrs (SD): P values related to manipulated ER way crossover 3) IN manipulated LV morphine- administration (≥3 28.1 (8.1) ADER-IMT, 60 mg (n=46) occasions within the Overall drug liking VAS, Emax 4) Oral morphine-ADER-IMT, 60 yr before screening); a Mean weight, lb Median (SD) mg (n=46) recreational user is a (SD): 1) 71.0 5) Placebo nonmedical opioid 161.8 (26.0) 2) 51.0 user with ≥10 3) 50.5 occasions within past 4) 59.0 ©Institute for Clinical and Economic Review, 2017 Page 131 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents After screening, participants year and ≥1 in the 12 Mean BMI, kg/m^2 5) 50.0 entered a naloxone challenge to wks before screening (SD): p<0.0001 relative to manipulated ER exclude opioid-dependent 24.0 (2.9) participants Exclusion: Take drug again VAS, Emax Median History of substance (SD) Then a drug discrimination test and/or alcohol 1) 73.0 (received IN placebo or dependence 2) 50.0 morphine, 30 mg IR in a RCT, (excluding caffeine 3) 50.0 double-blind, double-dummy and nicotine); any 4) 56.0 (p=0.0003) manner) to exclude participants condition in which an 5) 50.0 who couldn’t tolerate 30 mg IR opioid is p<0.0001 for other arms relative to morphine. contraindicated; manipulated ER history of sleep apnea After IN administration of in past 5 yrs that has manipulated high-volume not been resolved or morphine-ADER-IMT, been corrected participants were randomized 1:1:1:1 to receive IN manipulated LV morphine ER, IN manipulated LV morphine- ADER-IMT, intact oral morphine- ADER-IMT, and placebo in crossover fashion ©Institute for Clinical and Economic Review, 2017 Page 132 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table F2. Post-Market Studies Author & Year Study Design Intervention (Inclusion & Exclusion Outcomes of Publication (Study sites/ N Criteria) (Trial name) Duration of study) Patient characteristics Quality Rating OxyContin (OxyContin extended release) Black R JPain Observational design 1) OxyContin NAVIPPRO System Changes in past 30 day abuse patterns of OxyContin and 182 2012 comparing the 2) ER Oxymorphone comparator opioids before and after introduction of ORF prevalence, by 3) ER morphine Patient characteristics: Percent of product-specific abuse among Rx opioid abusers: POSTER routes, of past 30- Mean age, yrs. (SD): pre- / post- / %change (p-value) ABSTRACT day abuse of ORF in N= 104,630 (all 33.9 (11.6) 1) 23.84% / 11.91% / -50 (<0.0001) the period after its respondents) 2) 1.95% / 4.55% / +134 (<0.0001) introduction to that White (%): 3) 5.25% / 4.54% / -14 (0.0302) of OC before ORF 54,737 (54.1) introduction Percent of product-specific abuse among all individuals Male (%): 1) 4.30% / 2.35% / -45 (<0.0001) Uses the ASI-MV, a 68,496 (65.5) 2) 0.36% / 0.89% / +145 (<0.001) computer- 3) 0.96% / 0.89% / -8 (0.2330) administered Reported chronic medical interview based on problem (%): Changes in average number of days per month reported the Addiction Yes: 31,747 (30.3) abusing OxyContin and comparator opioids before and after Severity Index for No: 72,612 (69.4) introduction of ORF treatment planning Unknown/NR: 271 (<1) pre- / post- / %change (p-value) for adults 1) 11.0 days / 7.3 days / -33.3% (<0.001) Self-reported pain problem 2) 5.2 days / 7.3 days / +40.83% / (0.0023) Pre-introduction: (%): 3) 9.3 days / 9.3 days / 0.01% / 0.0 days (0.9983) Jun 2009 – Aug 2010 Yes: 32,791 (31.3) No: 71,648 (68.5) Post introduction: Unknown/NR: 191 (<1) ©Institute for Clinical and Economic Review, 2017 Page 133 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Aug 2010 – Jun 2011 Percent of abuse via specific ROA for OxyContin, ER Oxymorphone and ER morphine before and after introduction of ORF among those who reported abuse Inject (pre- / post-): 1) 34.3% / 15.6% 2) 8.4% / 11.6% 3) 32.6% / 35.8% Snort (pre- / post-): 1) 58.4% / 27.4% 2) 70.2% / 76.5% 3) 31.1% / 28.5% Oral (pre- / post-): 1) 54.9% / 77.0% 2) 37.3% / 27.1% 3) 48.7% / 44.2% Butler S Journal Cross-sectional, 1) Crush resistant Patients characteristics Abuse prevalence of CRTs and comparators by any oral of Pain 2016 183 observational study tablets (CRT) ER opioid N= 364,329 mode of administration involving product manipulation using sentinel category (OxyContin Male: 56.3% among oral abusers POSTER surveillance system reformulated, Opana (prevalence of abuse per 100 abusers) ABSTRACT of substance abuse ER reformulated, and Age 1) 41.5* evaluations Nucynta ER) 21-34: 64.0% 2) 34 2) Non-CRT versions of 35-44: 20% 3) 34.3 Uses the ASI-MV, a tablets category >45: 15.9% 4) 35.3 standard 3) Original or generic 5) 32 computerized clinical oxycodone ER Region 6) 36 interview for 4) All morphine ER South: 50.4% *p<0.003 evaluation and triage (excluding EMBEDA) West: 24.0% Non-tampering abuse/abuse by chewing/abuse by in substance use 5) Original or generic Midwest: 19.2% dissolving and drinking of CRTs and comparators by Oxymorphone ER Northeast: 6.4% ©Institute for Clinical and Economic Review, 2017 Page 134 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents disorder treatment 6) Oxycodone IR SE swallowing whole among oral abusers (prevalence of abuse settings per 100 abusers of product by any oral route) Reported abuse 1) 79/35.5/4.1 Jan 2009 – Mar 2015 -any Rx opioid: 76,108 (20.9%) 2) 90/30/5.2 -target drugs: 28,107 (7.7%) 3) 90/30.4/5.3 -by oral route: 18,135 (5.0%) 4) 82/31/4.9 5) 88.5/26/4.1 6) 87.5/31/5 *note these are estimates taken from a graph Butler, 2013 58 Time-series 1) ER oxycodone Included sites that collected Changes in rates of abuse: pre-ORF/post-ORF/% change (p observational study (OxyContin pre-ORF & data for OxyContin and ORF in value pre- vs. post-) Fair ORF post-ORF) both the pre- and post ORF A) Prevalence of past 30-days abuse among all individuals Samples obtained 2) ER Oxymorphone periods assessed by ASI-MV from 357 centers in 3) ER morphine 1) 4.06/2.41/-41 (p<0.0001) United States and Patient characteristics: 2) 0.32/1.11/+246 (p<0.0001) part of the ORF=Reformulated ER Age 3) 0.92/0.95/ +2 (0.6634) NAVIPPRO oxycodone Over 55yrs-3.7% B Prevalence of past 30-days abuse among prescription surveillance system 35 to 54 yrs-32.6% opioid abusers assessed by ASI-MV N=140,496 21 to 34 yrs-53.9% 1) 23.69/ 12.12/ -49 (p<0.0001) 14-months pre- under 21yrs-9.9% 2) 1.87/ 5.54/ +196 (p<0.0001) release of 3) 5.37/ 4.7/-12 (0.0209) reformulated ER Male: 55.6% oxycodone (ORF) & Changes in routes of administration (ROA): Pre/post (p value 20-months post White: 66.2% pre- vs. post-) release of ORF 1) Oral: 54.5% / 76.1% (p<0.0001) Abuse of any prescription Snorting: 52.7% / 25.4% (p<0.0001) opioids: Smoking: 6.4% / 4.2% (p=0.0373) 18.8% Injection: 35.7% / 15.9% (p=0.0002) 2) Oral: 38.2% / 30.1% (p=0.0056) ©Institute for Clinical and Economic Review, 2017 Page 135 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Past 30-day ER oxycodone Snorting: 61.8% / 68.8% (p=0.0162) abuse Smoking: 0.2%/1.9% (no statistically significant change) 1) Pre-ORF: 2,894 (24%) Injection: 8.6% /15.6% (p=0.0124) 2) Post-ORF: 1,705 (12.1%) 3) Oral: 46.7%/45.9% Snorting: 25.3%/25.6% Smoking: 0.9%/1.7% Injection: 45.7%/46.0% (No statistically significant change) Frequency of abuse in Past 30 Days Reported, mean days: pre-ORF/post-ORF/% change (p value pre- vs. post-) 1) 10.75 days / 7.48 days / -30.44% (p<0.0001) 2) 5.11 days / 7.78 days /+52.23% (p<0.0001) 3) 9.11 days / 10.07 days / +10.55% (p=0.0909) Cassidy Pain Observational, cross- 1) All Rx opioids Patient characteristics From 2008-2011 Medicine 2014 64 sectional; time-series 2) ER opioids Majority of individuals were Trends in prevalence of 30-day abuse of opioid analgesics analysis study 3) IR opioids between age 18-35: 57.3 and other drug categories Fair percent Total average quarterly percent change (QPC) 30 days; the 4) Oxycodone ER Abuse Prevalence: individuals assessed 5) Oxymorphone ER Median age: 32 years 1) +0.70% were from the 6) Morphine ER 2) +0.97% Addiction Severity Males: 64.5 percent 3) +0.18% Index – Multimedia N=232,874 p<0.05 compared to IR opioids Version (ASI-MV) White: 54.2 Percent Adjusted Abuse Prevalence (per million prescriptions): from January 1, 1) -0.38% 2008, to December Prescription opioids as Patient population: 2) -2.33% (p<0.0001) 31, 2011 a class, all immediate- -Traditional substance use 3) -0.87% (p<0.05) release (IR) opioids disorder treatment (55%) 437 facilities and all extended -involvement in the criminal release (ER) opioids as justice system (24.7%) ©Institute for Clinical and Economic Review, 2017 Page 136 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents separate categories as -DUIs (7.3%) Change in prevalence of past 30-day abuse of opioid well as specific -Other settings (13.0%) analgesics after opioid ADF introduction: pre- /and post-ADF prescription opioid period/ pre-post relative percent change (p value): compounds and other West: 41.4% Rates per 100 assessments drugs of abuse South: 39.2% 1) 16.94/18.36/+8.3% (p<0.0001) (heroin, cocaine, Midwest: 15.2% 2) 12.57/12.88/+10.5% (p<0.0001) amphetamine) Northeast: 4.1% 3) 9.70/10.72/+2.5% (p=0.192) 4) 6.49/5.08/-21.7% (p<0.0001) 5) 0.32/0.94/+190.9% (p<0.0001) 6) 1.13/1.10/NR (p=0.677) Change in prevalence of past 30-day abuse of opioid abuse compounds among abuser route of administration: pre- /post-ADF/pre-post RR: Oral: 4) 12.88/17.49/0.74 5) 0.17/0.41/2.47 6) 1.27/1.29/1.02 Snort: 4) 55.35/33.71/0.61 5) 3.42/13.16/3.85 6) 3.58/2.55/0.71 Inject: 4) 50.86/41.49/0.82 5)0.68/5.30/7.82 6) 14.78/25.21/1.71 Chilcoat H Drg Open cohort study (# of pre/post IMS LRx database with 150 Changes in doctor-shopping rates for brand ER oxycodone and Alc Dep covering >150 million patients) million unique patients relative to changes for comparator opioid analgesics and 2016 165 patients and 65% of 1) ER oxycodone benzodiazepines pre- and post-reformulation of brand ER retail U.S. (849,860/2,130,955) oxycodone: (pre-/post-period rate/ pre-to post % change): ©Institute for Clinical and Economic Review, 2017 Page 137 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Fair prescriptions from 2) IR hydromorphone 1) 0.25%/0.12%/-50 IMS LRx longitudinal (620,444/1,949,226) 2) 0.09%/0.06%/-25 data 3) IR oxycodone APAP 3) 0.13%/0.10%/-23 (9,335,562/25,167,86 4) 0.15%/0.13%/-13 6-month calendar 3) 5) 0.18%/0.16%/-9 intervals before and 4) IR hydrocodone 6) 0.09%/0.09%/4 after the APAP 7) 0.34%/0.36%/5 introduction of (26,479,737/74,140,8 8) 0.09%/0.15%/66 reformulated brand 39) ER oxycodone 5) benzodiazepines (15,519,660/43,160,2 31) 6) ER morphine (663,514/2,164,569) 7) IR oxycodone SE (1,527,554/6,420,004) 8) ER Oxymorphone (103,559/400,809) APAP=acetaminophen SE=single entity Cicero NEJM Self-administered N=2,566 Patients entering treatment Primary drug (%) 201263 surveys completed programs in U.S. for whom a (4Q2009-3Q2010)/3Q2010/1Q2011/4Q2011/1Q2012: anonymously by 1) OxyContin prescription opioid was 1) 35.6*/28/26/15/12.8 Poor independent cohorts 2) Hydrocodone primary drug of abuse 2) 24/32/29/29/29 3) Other opioids 3) 20.1**/17/21/33/32.3 Additional subset 4) Other oxycodone 4) 20/21/22/22/26 voluntarily agreed to 5) Heroin Drug used to get high in the last 30 days (%) online/telephone 1) 47.4*/48/42/38/30.0 interviews for 2) 65/73/66/68/65 ©Institute for Clinical and Economic Review, 2017 Page 138 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents qualitative 3) 61/62/65/68.5/66 information 4) 60/60.5/61/52/51 gathering 5) 9/5/5/15/17 *p<0.001 compared to 1Q2012 **p=0.005 compared to 1Q2012 Cicero Pain 2016 Survey study from 1) Original OxyContin Aged above 18; entering their Rates of abuse over time (pre- and post-rates) 62 Survey of RADARS (n=966) substance use disorder 1) 44.2%/ 25% Key Informants’ 2) Opana ER (n=128) treatment program; primary 2) 5.5%/ 7.6% Fair Patients (SKIP) diagnosis of opioid abuse (as *total SKIP respondents (n=12,124) program and an 85% defined by DSM criteria) response rate was Routes of administration for pre- and post- OxyContin and attained Patient characteristics: Opana ER: (pre-/post-/% change (p value)) [%] Mean age, yrs *RAPID participants; OxyContin (n=117) and Opana ER Subset of 1) 32.8 (n=35) respondents (25.5%) 2) 29.9 Any non-oral said that they were 1) 91.5/47.9/ 47.7 (p<0.001) willing to give up Male, % 2) 94.3/77.1/18.2 (p=0.06) their anonymity and 1) 56.2 Inject participate in a 2) 58.4 1) 42.7/21.4/50.0 (p=0.001) follow-up study, 2) 60.0/51.4/14.3 (p=0.471) dubbed Researchers White, % Snort and Participants 1) 75.4 1) 78.6/28.2/64.1 (p<0.001) Interacting Directly 2) 84.9 2) 80.0/37.1/53.6 (p<0.001) (RAPID) Smoke Regions 1) 17.9/7.7/57.1 (p=0.022) Midwest: 27.9% 2) 20.0/2.9/85.7 (p=0.052) Northeast: 15.5% Any oral South: 33.6% 1) 63/94/-49.2% (p<0.0001) West: 23.0% 2) 20/21/-5.0% (p=0.808) ©Institute for Clinical and Economic Review, 2017 Page 139 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Cicero T JAMA Survey study using 1) Original (pre-2010) SKIP program consisted of key Past month abuse of OxyContin, % Psychiatry 2015 data from the OxyContin informants from more than -Prior to reformulation (Jan-June 2009): 45.1 61 ongoing Survey of formulation 150 public and privately -Post-reformulation (Jul-Dec 2012): 26 (p<0.001) Key Informants’ 2) Reformulated funded treatment centers in -Post reformulation rate reached a plateau at 25 to 30%, Fair Patients (SKIP) OxyContin 48 states with no further decline from 2012 to 2014 program (n=10,784), part of the Mean age, yrs (SEM) Past month abuse of heroin, % Researched Abuse, SKIP/RAPID respondents: -Prior to reformulation (Jan-June 2009): 25 Diversion and 34.1 (0.1)/35.9 (0.6) -Post-reformulation (Jul-Dec 2012): 40 (p<0.001) Addiction-Related Surveillance Male, % Interview of RAPID participants (RADARS) system 50.6/46.4 Residual Abuse (N=153, RAPID) that collects and 51 (33.3%) ADF had no effect on drug selection, continue analyzes post- White, % OxyContin abuse marketing data on 78.4/90.4 51 (33.3%) replaced OxyContin because of ADFs misuse and diversion 5 (3.3%) stopped abusing drugs because of ADFs of prescription opioid % of participants from each 46 (30.1%) didn’t abuse enough to be influenced analgesics and heroin region: Route of Administration (N=244, RAPID) Midwest: 28.5% 38 (43%) switch from injecting/inhaling to swallowing RAPID data come Northeast: 16.9% 30 (34%) defeated ADF and continued injecting/inhaling from respondents South: 31.7% 20 (23%) had been swallowing before ADF formulation and from the SKIP survey West: 22.9% ADF had no effect on their continued oral use that were willing to Transition to Other Drugs among those that replaced give up their OxyContin (n=55) anonymity and 70% changed to heroin participate in the 25% changed to other oxycodone interview-based RAPID program (n=244)*82.0% ©Institute for Clinical and Economic Review, 2017 Page 140 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents response rate from the SKIP survey January 2009 – June 2014 Coplan Clin 10 studies which 1) Original Oxycodone NAVIPPRO: patients in 1,000 Rates of OxyContin overdose diagnoses, per 100 person- Pharm and Ther examine from 1 year 2) Reformulated different substance use years 2016 60 before to 3 years Oxycodone disorder treatment centers in 1) 0.42, 2) 0.28 after OxyContin 3) Other Schedule II 36 states Fair reformulation opioids RADARS OTP: patients in 70 Changes 1 year before to 3 years after reformulation conducted as part of 4) IR Oxycodone different public methadone Misuse (Radars Poison Center) (%) required FDA (single-entity) maintenance clinics 1-2) -43 postmarketing 5) IR Oxycodone- RADARS SKIP: patients at 3) -6 program. A 6 month Acetaminophen private substance use transition period 6) ER Morphine disorder treatment centers Abuse, Radars PC /NPDS/NAVIPPRO/SKIP/OTP/Kentucky (3Q2010-4Q2010) 7) ER Oxymorphone University of Kentucky: study (%) was excluded from 8) IR Hydrocodone- abusers of OxyContin in rural 1-2) -55/-55/-48/-30/-43/-85 calculation to allow Acetaminophen Kentucky 3) -7/-4/-3/16/9/53 for original OC to be 9) Methadone MarketScan: patients depleted. 10) IR commercially insured Doctor shopping, IMS prescription data (%) Hydromorphone RADARS Drug Diversion: 1-2) -50 Studies included: patients involved in law 7) 66 National Poison Data enforcement cases regarding System (NPDS); drug diversion Change in Overdose using population rates, Rate of RADARS System Diagnosed Events/Adverse Event Database (%) Poison Center, 1-2) -34/-65 System Outpatient 6) 17/NR Treatment Program (OTP), Study of Key Change in death, Adverse Event Database (%) Informants’ Patients 1-2) -60 ©Institute for Clinical and Economic Review, 2017 Page 141 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents (SKIP), Drug Diversion Program; a Opioid prescription rates, prior-to-reformulation/post- study of individuals reformulation in rural Kentucky 1-2) 3.6/2.9 conducted by U of 4) 5.1/7.7 Kentucky; 5) 19.2/18.5 MarketScan; Fatal 6) 3.1/3.4 adverse events 7) 0.4/0.5 reported to 8) 66.4/65.0 manufacturer; IMS 9) 2.2/2.0 Health Prescription database; National Addictions Vigilance Intervention and Prevention Program (NAVIPPRO) Coplan Pharm Retrospective cohort 1) OxyContin Aged 18 to 64 years; incident Changes in Rates of Diagnosed Addiction/Dependence per and drug safe study from 2) ER Morphine or prevalent users of 100 Person-Years of Opioid Use in Individuals Dispensed One 2015 184 MarketScan 3) ER Oxymorphone OxyContin or 4 comparator Opioid (from 1 yr before to 3 yrs after introduction of commercial database 4) IR oxycodone opioids; separate cohorts reformulated OxyContin) POSTER single-entity were included for each drug. 1) -25% ABSTRACT Study period (divided 5) IR hydromorphone Opioid use defined as 2) +21% in three times duration of continuous use: 3) +13% around introduction ≤15 days between 4) +7% of reformulated prescriptions plus 15 days end 5) +31% OxyContin): 1 year of last prescription before (August 2009 Rates of Diagnosed Addiction/Dependence per 100 Person- – July 2010); Years of Opioid Use in Individuals Dispensed One Opioid, 2011-2013 1) 3.00 ©Institute for Clinical and Economic Review, 2017 Page 142 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 3 months’ transition 2) 3.18 period (August 2010 3) 5.95 – October 2010); 4) 5.58 3 years’ after 5) 4.89 (November 2010 – October 2013) Changes in rates of diagnosed addiction/dependence per 100 person-years of opioid use in individuals dispensed 1 August 2009-October opioid, % 2013 1) -25 2) 21 Diagnosed event of 3) 13 interest: ICD-9 CM 4) 7 diagnostic codes of 5) 31 304.0x and 304.7x codes Coplan Pharm National Poison Data 1) ERO OxyContin Data from the NPDS, which Changes in number of ERO exposures per quarter from 1 and drug safety System (NPDS) 2) Other single-entity captures 99.5% of poison year before to 2 years after reformulations: 2013 67 covering all US (SE) oxycodone exposures reported to all Average per quarter intentional abuse, pre-/post-/%change poison centers was 3) Heroin poison centers in the USA (p value): Fair used to measure 1) 130.3 / 83.3 / -36 % (p<0.0001) changes in exposures Exposures reported to poison 2) 228.5 / 273.4 / +20% (p<0.0001) in the year before centers are classifieds into 3) 355.8 / 505.1 / +42% (p<0.0001) versus the 2 years reasons: intentional abuse, after introduction of unintentional therapeutic Average per quarter intentional misuse, pre-/post- reformulated errors, unintentional general /%change (p value): extended release exposures, and adverse 1) 51.3 / 40.4 / -21% (p=0.0076) oxycodone (ERO) reactions 2) 104.0 / 119.6 / +15% (p=0.0172) OxyContin 3) 46.5 / 60.4 / +30% (p=0.0025) ©Institute for Clinical and Economic Review, 2017 Page 143 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 1 year preceding (3Q2009-2Q2010) to the 2 years following (4Q2010-3Q2012) Davis Annals of Data from RADARS 1) ER Oxycodone Number of diversion events prior-to- Emer Med 2012 Drug Diversion 2) Other prescription reformulation/Percentage of 2000 US population covered 185 Program opioids by RADARS database 2008-Q4 POSTER Approximately 300 1) 466/37.3 ABSTRACT drug diversion agents 2) 4310/37.3 in 50 states and Puerto Rico submit 2010-Q3 data quarterly on the 1) 488/38.4 number of 2) 3586/38.4Number of diversion events post-re- documented drug formulation/Percentage of 2000 US population covered by diversion cases RADARS within their 2010-Q4 jurisdiction for 1) 306/36.3 specific prescription 2) 3282/36.3 drugs 2012-Q1 Original OxyContin 1) 177/45.7 (October 2008 – 2) 3488/45.7 September 2010) Average ER oxycodone diversion population rate after compared to ADF reformulation is 53% lower than average population rate reformulation before reformulation (p<0.001) (October 2010 – March 2012) ©Institute for Clinical and Economic Review, 2017 Page 144 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents DeVeaugh-Geiss Survey data from The 1) OxyContin – One Past year initiation of nonmedical use of OxyContin per Post Med 2016 National Survey on year before 10,000 population 66 Drug Use and Health reformulation (2009) 2009 2011 2012 2013 2014 (NSDUH) were used; 2) OxyContin – Each 19 15 12 14 9 POSTER NSDUH is designed to year post % Change -19% -38% -28% -51% ABSTRACT provide estimates of reformulation from compared with the prevalence of 2011 - 2014 2009 nonmedical drugs in the US household Past year initiation of nonmedical use of OxyContin per population age 12+ 10,000 prescription OxyContin dispensed years and assesses 2009 2011 2012 2013 2014 drug use from a 868 746 635 773 551 sample of 60,000 % Change -14% -27% -11% -36% individuals per year compared with 2009 Looks at data from 2008-2014 (two Past month nonmedical use of OxyContin per 10,000 years before population reformulation and 2009 2011 2012 2013 2014 four years after) 17 14 11 16 11 % Change -16% -31% -6% -33% compared with 2009 Past month nonmedical use of OxyContin per 10,000 OxyContin prescription dispensed 2009 2011 2012 2013 2014 753 671 611 873 654 ©Institute for Clinical and Economic Review, 2017 Page 145 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents % Change -11% -19% 16% -13% compared with 2009 Degenhardt L Methods include: 1) Original OxyContin, Opioid use in the NOMAD cohort, pre-and post-OxyContin Drug and data on 80 mg formulation Alcohol pharmaceutical 2) Reformulated Pre-introduction Dependence opioid sales; drug use OxyContin, 80 mg % used in past month 2015 71 by people who inject 3) Morphine 1) 56, 2) N/A, 3) 65, 4) 64 drugs regularly 4) Heroin % injected past month NOMAD (PWID); client visits 1) 55, 2) N/A, 3) 63, 4) 64 to the Sydney N=606 (Jan-March % chewed, snorted, dissolved or smoked past month Poor Medically Supervised 2014) 1) 3, 2) N/A, 3) 4, 4) NR Injecting Centre (MSIC); and last drug N=547 (May-August Post-introduction injected by clients of 2014) % used in past month inner-Sydney needle- 1) 16, 2) 8, 3) 44, 4) 49 syringe programs %injected past month (NSPs) 1) 15, 2) 3, 3) 42, 4) 48 %chewed, snorted, dissolved, or smoked past month Conducted in 1) 0.2, 2) 1, 3) 4, 4) NR Australia 2009-2014 Past month accidental overdoses reported by the NOMAD cohort Pre-introduction of reformulated OxyContin 4% (25) of total sample overdosed in past month 3% (17) involved heroin 0.3 % (2) involved morphine 0.3% (2) involved oxycodone Post-introduction of reformulated OxyContin 3% (17) of total sample overdosed in past month ©Institute for Clinical and Economic Review, 2017 Page 146 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 2.4% (13) involved heroin 0.2% (1) involved morphine 0.2% (1) involved morphine % strongly agreed that they would tamper with drug in future, prior to reformulation/post-reformulation 1) 74/73, 2) NA/20*, 3) 71/73 *p<0.05 compared to prior-to-reformulation % of those who injected strongly agreed that it is easy to dissolve/cut up/inject, prior to reformulation 1) 27/34/7 2) NR/NR/NR 3) 23/28/10 % of those who injected strongly agreed that it is easy to dissolve/cut/inject, post-reformulation 1) 43*/52*/3 2) 10*/21/50 3) 40/51/6 *p<0.05 compared to prior to reformulation Havens J Drug Structured interviews 1) ER Oxycodone Patient characteristics: Differences in prevalence and frequency of abuse (through Alcohol Depend assessing opioid 2) IR Oxycodone N=189 any route) 2014 69 abuse were 3) Reformulated ER Prior to reformulation prevalence – Overall completed by 190 Oxycodone Male: 54.5 percent 1) 0.99, 2) Ref., 3) N/A, 4) N/A Poor individuals recruited 4) Any ER Oxycodone Post-reformulation prevalence – Overall from rural Perry Median age: 32 1) 0.62, 2) Ref., 3) 0.34, 4) 0.76 County, Kentucky Prior to reformulation frequency – Overall between December White: 97.9 percent 1) 1.05, 2) Ref., 3) N/A, 4) N/A Past abuse of opioids: Post-reformulation frequency – Overall ©Institute for Clinical and Economic Review, 2017 Page 147 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 2010 and September Original ER Oxycodone: 100% 1) 0.35, 2) Ref., 3) 0.01, 4) 0.43 2011 Reformulated: 51.3% *values are relative rate Hydrocodone: 97.9% Past 30-day abuse Heroin: 31.2% Prevalence for IR/ER Oxycodone and retrospectively (pre-/post-reformulation) reported abuse prior 1) 74% / 33% to the reformulation 2) 74% / 96% in August 2010 Frequency for IR/ER Oxycodone (pre-/post-reformulation) 1) 13.4/1.9 2) 12.8/19.5 Hwang C Segmented time- 1) IR Oxycodone 38,000 retail stores, 119 mail Annual prescription growth rate (%) Pharmacoepide series analysis using 2) IR Hydrocodone service pharmacy outlets, and August 2009-July 2010 miology Drug the IMS Health 3) ER Opioids about 820 long term care 1) 14.7, 2) 5.8, 3) -1.1, 4) 19.2, 5) -10.3, 6) 6.0 Saf 2015 167 national prescription 4) IR Opioids facilities, which captured over Audit, a nationally 5) OxyContin 70% of all prescription activity August 2010-July 2011 Fair representative 6) ER oxycodone in the USA 1) 6.7, 2) 4.9, 3) 3.9, 4) 12.8, 5) -7.0, 6) -24.9 source of (OxyContin + generic prescription activity formulations) Monthly change in the number of prescriptions dispensed (in in the USA thousands) August 2009-July 2010 12 months prior to 1) 45.3, 2) 51.6, 3) -0.2, 4) 39.2, 5) -5.2, 6) 3.4 and following August 2010 August 2010-July 2011 1) 25.9, 2) 45.7, 3) 0.7, 4) 34.2, 5) -3.7, 6) -15.9* *p<0.01 Jones C Clin J Data was from the OxyContin Civilian, non-institutionalized Past-year nonmedical use of OxyContin among US overall Pain 2016 65 National Survey on population aged 12 years and population age 12 or older (%) Drug Use and Health; older 2006/2007/2008/2009/2010/2011/2012/2013 Fair state-based sampling 0.5/0.6/0.6/0.7/0.7*/0.6/0.6/0.5 ©Institute for Clinical and Economic Review, 2017 Page 148 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents design with independent, Past-year nonmedical use of OxyContin among past-year multistage area nonmedical users of pain relievers, (%) probability samples 10.5*/11.4/12.3/13.5/15.4/14.6/11.8/13.0 within each state and District of Columbia Past-year nonmedical use of OxyContin among people with pain reliever abuse or dependence, past-year heroin users, Multivariable logistic and people with a history of drug injection, (%) regression was used Pain reliever abuse: to identify individual 20.1*/31.4/32.5/32.6/37.7/37.1/31.1/31.8 characteristics Heroin abuse: NR/NR/34.2/38.3/42.0/44.4/43.3/37.6 associated with past- Ever inject: 7.3/6.3/8.6/10.2/9.1/9.0/9.9/8.6 year OxyContin nonmedical use prior *p<0.05 compared to 2013 to and after reformulation 2006 through 2013 Kadakia A Retrospective cohort 1) ER morphine Aged 18 to 64 years; incident Opioid overdose/poisoning diagnosis rate in patients Pharm and drug study using data 2) ER Oxymorphone or prevalent users of dispensed one opioid, by type of opioid, Percent change Saf 2015 68 from Truven 3) IR hydromorphone OxyContin or 4 comparator (pre-/post reformulation) MarketScan 4) IR oxycodone opioids 1) +17% ABSTRACT commercial database single-entity (SE) 2) 0% POSTER 5) OxyContin 3) +10% August 2009 – 4) -1% October 2013 *separate cohorts 5) -34% were included for each opioid Opioid addiction/dependence diagnosis rate among patients dispensed one opioid, Percent change (pre-/post- reformulation) 1) +21% ©Institute for Clinical and Economic Review, 2017 Page 149 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 2) +13% 3) +31% 4) +7% 5) -25% Opioid abuse diagnosis rate among patients dispensed one opioid, Percent change (pre-/post-reformulation) 1) +44% 2) +236% 3) -6% 4) +36% 5) -35% Difference in abuse change per 100 person years from baseline for OxyContin vs. comparator opioids (pre-/post- /%change (p-value)) 1) 0.29 / 0.42 / +44% (0.004) 2) 0.16 / 0.55 / +236% (0.027) 3) 0.64 / 0.60 / -6% (0.158) 4) 0.58 / 0.79 / +36% (0.000) 5) 0.49 / 0.31 / -35% Michna E Curr Truven MarketScan 1) Reformulated ER Inclusion: Following the introduction of reformulated ER oxycodone Med Res Opin pharmacy and oxycodone (n=15,162) Commercially insured Primary drug post-reformulation of ER oxycodone (%) 2014 164 medical claims data 2) Reformulated ER patients, age 18-64; 1) 10,520 (69.4) in a 6-month period Oxymorphone continuous use of ER/LA 3) 3230 (21.3) Fair prior to the (n=2285) opioids, at least 120 days’ 4) 1412 (9.3) introduction of the 3) Other ER/LA opioid supply; primary ER/LA opioid respective with abuse-deterrent in the 6-month period from Primary drug post-reformulation of ER Oxymorphone (%) reformulation technology Feb to Aug 2010 (prior to 1) 157 (6.9) 4) No ER/LA opioid reformulation of OxyContin) 2) 1149 (50.3) ©Institute for Clinical and Economic Review, 2017 Page 150 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Patient extended was ER oxycodone – primary 3) 581 (25.4) release (ER)/long LA=long acting opioid defined as ER/LA 4) 398 (17.4) acting (LA) opioid opioid that accounted for at utilization in 6-month least 70% of days’ supply of all Patients diagnosed with abuse during 15-month study period from Nov ER/LA opioids in 6-month period 2010 to May 2011 period. ER oxycodone patients observed. Rate of Abuse 1) 3.5% Assessed whether ER 3) 6.7% oxycodone patients 4) 10.9% switched to reformulated ER ER Oxymorphone patients oxycodone with ADF, Rate of Abuse switched to non-ADF 1) 2.5% ER/LA opioids, or 2) 2.1% discontinued ER/LA 3) 2.6% opioids; also, 4) 5.0% evaluated rates of abuse (ICD-9-CM claims) between ADF/non-ADF switch populations Peacock Intl J Prospective cohort 1) Original OxyContin Inclusion: ≥18 years old; Attractiveness by NOMAD participants who injected original Drug Policy study 2) Original OxyContin, English language proficient; and reformulated oxycodone, % agree with statement 2015163 after the release of extra-medical pharmaceutical I would definitely tamper with the oxycodone product: Participants recruited reformulation opioid use on ≥monthly basis 1) 84, 2) 79, 3) 53 Fair through Needle- 2) Reformulated in the last 6 months; reported The oxycodone product is unpleasant to use (tamper): Syringe Programs, OxyContin injecting, snorting, chewing, 1) 16, 2) 5, 3) 50 snowballing and smoking, and/or dissolving The oxycodone product is difficult to inject: word-of-mouth, N=606 and drinking a pharmaceutical 1) 0, 2) 0, 3) 47 ©Institute for Clinical and Economic Review, 2017 Page 151 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents opioid substitution opioid in the last month and The oxycodone product is painful to inject: therapy on a monthly or more 1) 0, 2) 11, 3) 40 clinics/prescribers, frequent basis in the past 6 The oxycodone product contains fillers that cause safety community months. issues: pharmacies, and 1) 74, 2) 63, 3) 93 advertisements in Exclusion: Not a resident of The oxycodone product is easy to cut up: media across the city/state for the 6 1) 79, 2) 79, 3) 21 Australia. months prior to interview; The oxycodone product is easy to dissolve: had been in prison for the 1) 67, 2) 74, 3) 14 Participants past month; had only completed tampered with an opioid structured computer- substitution therapy assisted interviews medication; reported only (Phase 1: Jan-March using opioid medication as 2014 prior to release per a doctor’s instructions. of reformulated oxycodone. Phase 2: Mean age: May-August 2014 Only tampered with original following formulation: 41 reformulation). Tampered with original and reformulated: 39 % male: 69 Rossiter JME Time-series Prior to reformulation Commercially insured patients Abuse among continuous ERO users following introduction 201487 observational study ER oxycodone Abusers (n=2532)/Non- of reformulated ER oxycodone, commercial/Medicare- abusers (n=61,421) eligible/Medicaid: Poor Measure reductions Post-reformulation ER Age: 47.9/51.2 Continuous users of EROs, prior to reformulation (%) in rates of diagnosed oxycodone Male, %: 47.2/43.9 3.6/1.2/6.2 opioid abuse Continuous users of reformulated ER oxycodone, post- following ER Medicare eligible patients reformulation (%) ©Institute for Clinical and Economic Review, 2017 Page 152 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents oxycodone Abusers (n=272)/Non-abusers 2.8/1.3/5.1 reformulation; used (n=19,564) Relative change in abuse rates, (%) medical and Age: 74.1/76.9 -22.7*/6.1/-18.0** pharmacy claims for Male, %: 30.5/33.3 *p<0.001 continuous ERO **p=0.034 (extended-release Medicaid patients opioids). Abusers (n=548)/Non-abusers Reduction in number of abusers, commercially- (n=7,770) insured/Medicare-eligible/Medicaid/Uninsured Used 2009-2011 Age: 46.9/48.9 Diagnosed abusers: Truven Health Male, %: 37.8/40.2 3,673/0/1,371/3,079 Analytics (Truven) Undiagnosed abusers: de-identified medical Patients classified as 18,364/0/6,856/15,394 and pharmacy claims diagnosed abusers if they had data for patients with medical claims with ICD-9-CM ≥1 Rx drug claim for diagnosis codes for opioid an opioid during this abuse or dependence. period. Opioid includes both extended- and immediate-release opioids. Sankey JOM Noninterventional, 1) OxyContin Patients included if Mean (SD) per-patient incidence rate of oxycodone-positive 2016 70 multicenter, 2) Oxycodone methadone maintenance UDS during baseline/transition/post-OxyContin periods, %: prospective historical 3) Heroin therapy (MMT) and 22.4 (27.1)/13.8 (21.5)/10.5 (19.6) Poor chart review 4) Morphine maintained with diagnosis of 5) Hydrocodone opioid dependency; if Mean (SD) decrease in incidence rates from baseline in 2014-November entrance into methadone oxycodone-positive UDS from baseline to transition/baseline 2015 Completer population, treatment program no later to post-OxyContin period, %: Prospective historical n=250 than March 1, 2011; -8.7 (20.4)/-11.9 (24.1) chart review continued treatment up to P<0.001 for both ©Institute for Clinical and Economic Review, 2017 Page 153 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents conducted in opioid- Questionnaire and including until December dependent patients completed, n=177 31, 2012; had at least one Self-reported opioid use in Baseline/Transition/Post- on methadone oxycodone-≥1 positive UDS OxyContin periods, % maintenance therapy (urine drug screen) for 1) 25.7/14.5/7.4 in 3 Canadian oxycodone during baseline. 2) 88.6/79.3/71.3 centers. Two-part 3) 4.0/4.1/3.7 study included chart Completer patients had ≥36 review and self- UDS completed and ≥10 Self-reported OxyContin/OxyNEO sourcing, as indicated by reported physician visits during questionnaire, %: questionnaire. baseline and ≥30 UDS and ≥8 Bought from dealer: 74.2/59.3 physician visits during Prescription from 1 doctor: 38.3/14.8 Baseline period: transition/post-OxyContin Prescription from >1 doctor: 5.5/1.9 March 1, 2011 to periods. UDS visits conducted Bought from friends/family: 34.4/35.2 February 29, 2012 ≥1/wk and physician visits Free from friends/family: 20.3/16.7 ≥1/month. Free from stranger/dealer: 9.4/5.6 Transition period: Patient characteristics: Stolen: 4.7/1.9 March 1, 2012 to Male, 55.0% August 31, 2012 Overall incidence rate of oxycodone-positive UDs: 19.5% [represents Canadian Mean age: 33.9 years OxyNeo release] Overall incidence rate of morphine-related-positive UDSs: Route of administration/self- 10 % Post-OxyContin reported nonmedical drug use period: September 1, history, % Ratio of the incidence rates of overall oxycodone-positive to 2012 to December Heroin: 16.4 morphine-related positive UDSs: 31, 2012 OxyContin oral: 26.0 1.96 OxyContin intranasal: 36.0 OxyContin IV: 13.2 Ratio of oxycodone-positive UDS incidence rate to morphine- OxyContin other: 6.0 related positive UDS incidence rate by study period OxyContin not specified: 12.4 Mean ratio (SD) Baseline: 5.49 (9.124) ©Institute for Clinical and Economic Review, 2017 Page 154 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Oxycodone not brand specific: Transition: 1.88 (3.766) 65.6 Post-OxyContin: 1.02 (1.028) Change from baseline in ratio (SD)) Preferred route of Transition: -1.99 (5.878) administration for opioids Post-OxyContin: -3.17 (6.181) self-reported in questionnaire, % Self-reported opioid use in methadone-maintained patients Oral route: 80.8 by study period (Baseline / Transition / Post-OxyContin) n Intranasal: 67.8 (%) Chewing: 64.4 1) 52(25.7) / 21 (14.5) / 8 (7.4) Injection: 44.6 2) 179 (88.6) / 115 (79.3) / 77 (71.3) 7.6% had comorbid chronic 4) 23 (11.4) / 12 (8.3) / 11 (10.2) pain, 1.2% comorbid acute 5) 1 (0.5) / 0 (0.0) / 0 (0.0) pain, ≥90% pain not reported Sessler Pharm & Time series OxyContin pre- and Frequently involved a person Changes in the number of ER oxycodone fatality reports per Drug Saf 2014 73 observational study; post-reformulation aged 18-64 years quarter received by the manufacturer from pre-to-post data pulled from introduction of reformulated ER oxycodone: Fair 3Q2009-3Q2013 1. All fatal cases Patient characteristics: *n=236 manufacturer’s (n=326) PRE-R/POST-R, male % adverse event 2. Subset of fatal cases 1. 63/66 Mean number of fatality per quarter (%change) reporting database of overdose (n=240) 2. 65/68 All fatal reports submitted to national (1 year pre/1 year post/2 year post/3 year post) drug-regulatory Prior to reformulation: PRE-R/POST-R, age authorities (PRE-R) distribution 13 to <18 yrs, % All fatal reports: 32.8 /30.5 (-7) /12.5 (-62) /5.8 (-82) 3Q2009-2Q2010 1. 5/6 Individual case report 2. 6/9 Fatality reports for ER oxycodone versus all oxycodone: narratives were Post-reformulation: (1 year pre/1 year post/2 year post/3 year post) categorized as (POST-R) PRE-R/POST-R, age % (n/N): mentioning an opioid 3Q2010-2Q2013 distribution 18 to <65 years, % 21 (131/637) /22 (122/551) /8 (50/616) /10 (12/120) overdose-related 1. 69/68 P<0.0001 ©Institute for Clinical and Economic Review, 2017 Page 155 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents event and/or drug 3rd year post- 2. 77/71 abuse-related reformulation: PRE-R/POST-R, Oxycodone and other opioid mentions, % behavior (3POST-R) PRE-R/POST-R, case reporter Group 1 Group 2 3Q2012-2Q2013 region, % OxyContin 52/52 44/41 Northeast: 1) 17/20, 2) 15/18 Oxycodone* 48/48 54/57 Midwest: 1) 16/19, 2) 17/20 Other opioid 30/18 37/24 South: 1) 39/30, 2) 40/29 Illicit** 18/16 22/21 West: 1) 18/17, 2) 18/18 *not specified formulation, although implied to be Missing: 1) 10/13, 2) 9/14 OxyContin ER because submitted to manufacturer **Marijuana, cocaine, amphetamines, and heroin % change in number of oxycodone ER fatality reports per quarter from PRE-R to 3POST-R 1. -82 2. -87 Ratio (%) of the number of fatalities involving ER oxycodone reported to manufacturer relative to fatalities with any oxycodone as suspect drug reported to FDA PRE-R/first 6 months of 3POST-R: 21/10, p<0.0001 Severtson Time series OxyContin original Patient characteristics: Average ER oxycodone abuse exposure mention population Annals of Emer observational study formulation and Percent of the 2000 US rate after reformulation: 38% (95% CI: 31-45%, p<0.001) Med 2012 186 from October 2008 reformulation population covered by Poison lower than the average population mention rate prior to to March 2012 Center Program: reformulation POSTER ABSTRACT 4 year follow up 2008Q4 – 2010Q1: 85.0% Mean ER oxycodone prior to reformulation events/Other prescription opioids events: ©Institute for Clinical and Economic Review, 2017 Page 156 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Data from the 2010Q2 – 2010Q4: 85.5% 152/1,440 RADARS System Poison Center 2011Q1 – 2011Q2: 86.0% Mean ER oxycodone post-reformulation events/Other Program; prescription opioids events: Quarterly rates of 2011Q3 – 2011Q4: 89.9% 98/1,535 poison center calls citing abuse of ER 2012Q1: 90.0% oxycodone before reformulation of OxyContin were compared to rates after introduction of reformulation Severtson S Time series 1) OxyContin RADARS Database: multiple Population adjusted baseline (2010-Q2) rate of abuse and Drug Alcohol observational study; 2) Other opioids (IR programs with post-market diversion /projected rate of abuse and diversion in 2015-Q2, Depend 2016 57 5-year span following OxyContin, IR and ER: surveillance of prescription per 100,000 population OxyContin hydrocodone, medication abuse Poison Center program: Fair reformation morphine, Poison Center Program: 1) 0.056/0.014, 2) 0.387/0.260 hydromorphone, recorded the substances Drug Diversion Program Analyzed post- tramadol, involved in poison center 1) 0.195/0.021, 2) 1.344/0.983 market surveillance oxymorphone, and cases classified as intentional Opioid Treatment Program of abuse and tapentadol) abuse 1) 0.574/0.100, 2) 0.986/0.670 diversion, poison Drug Diversion Program: Survey of Key Informants’ Patients center data, legal recording drugs involved in 1) 0.265/0.122, 2) 0.475/0.441 cases, drug abuse cases opened by law treatment programs, enforcement drug diversion Change in rate of abuse and diversion after reformulation of and drug street price investigators OxyContin (projected for 2015-Q2), population adjusted, % data Opioid Treatment Program Poison Center Program: and the Survey of Key 1) -75.0* , 2) -32.8 Informants’ Patients Program: Drug Diversion Program: ©Institute for Clinical and Economic Review, 2017 Page 157 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents query new patients entering 1) -89.4* , 2) -26.8 substance-abuse treatment Opioid Treatment Program: about medications abused 1) -82.6* , 2) -32.0 StreetRx Program: utilizes Survey of Key Informants’ Patients crowdsourcing website that 1) -53.9* , 2) -7.2 gathers street price data for *p<0.0001 compared with Other Opioids group drugs (publicly accessible) Route of administration, oral route/non-oral route change in rate of abuse from pre to post reformulation (%) -71.0/-86.7 (p=0.006) Severtson S J Surveillance data 1) ERO (extended RADARS System Poison Center Number of Events for ERO and Other Prescription Opioids Pain 2013 59 collected from the release oxycodone) and Drug Diversion Pre-reformulated ERO Abuse, 2008Q4/2010Q3 Fair RADARS System 2) all prescription 1) 158/183 Poison Center and opioids 2) 1497/1588 Drug Diversion programs were used Post-reformulated ERO abuse 2010Q4/2012Q1 to estimate rates of 1) 101/79 abuse exposures, 2) 1353/1610 unintentional Drug Diversion Program of RADARS – Number of Events, therapeutic error Pre-Reformulated ERO, 2008Q4/2010Q3 exposures, and 1) 466/488 diversion for ERO 2) 4,310/3,586 ©Institute for Clinical and Economic Review, 2017 Page 158 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents manufactured by Purdue Pharma LP Drug Diversion Program of RADARS – Number of Events, and other opioids in Post-Reformulated ERO, 2010Q4/2012Q1 aggregate in the 1) 306/177 periods before and 2) 3,282/3,488 after the introduction of reformulated ERO 2008-2012 Coplan P Pharm Patient safety 1) OxyContin Change in patient outcomes, 1-year before to 2/2.5 years Drug Saf 2013 187 outcomes were 2) Oxycodone single- after ERO Reformulation assessed in 4 post- entity Overdose fatalities (Adverse Event Reports) [%] POSTER marketing studies: 2) All Rx opioids 1) -64 ABSTRACT RADARS System p<0.0001 Poison Control study, All fatalities (Adverse Event Reports) [%] National Poison Data 1) -50* System (NPDS), 2) 15 adverse events or P<0.0001 fatalities and/or abuse reported to Adverse reactions (NPDS) [%] the manufacturer, 1) -34* and Kaiser 2) 15 Permanente P=0.0005 Northwest and Northern California Prescribing (IMS NPA) [%] study of opioid 1) -8.9 overdoses among patients prescribed opioids ©Institute for Clinical and Economic Review, 2017 Page 159 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 1 year preceding reformulated ERO vs. 2 or 2.5 yrs after reformulation LaRochelle M Interrupted time 1) all opioids Aged 18 to 64 years; enrolled Opioid dispensing rate, mg morphine equivalent dose per JAMA 2015 72 series study design 2) ER oxycodone in a commercial health plan; member per quarter; using an open cohort 3) other long acting used Optum data which estimated instantaneous change in overdose rate in the first Fair from a large national opioids contains all impatient, post-reformulation change quarter (2011Q1) compared US health insurer 4) propoxyphene outpatient, and pharmacy with expected rate based on baseline trend claims 5) other IR opioids claims from a large US health 1) -14.8 insurer with member in all 50 2) -4.56 Segmented N=31,316,598 states 3) 1.09 regression to analyze Patients could enter and exit 4) -12.2 changes in outcomes the cohort over the 10-year 5) NA from 30 quarters period on a rolling basis before to 8 quarters Ratio of prescription opioid overdose to total prescription after the 2 Patient characteristics: opioid dispensing, episodes per million g MED per quarter: interventions No. of members/male 0.31 / -0.005 / 0.0002 / NA / -0.0067 sex/pop. of white people, Intercept/linear trend/quadratic trend/level change) January 2003 to millions (by quarter): December 2012 2003: 7.2/3.6/5.1 Result of sensitivity analysis: 2005: 7.6/3.8/5.3 2007: 8.1/3.9/5.5 Scenario 1: 30 Quarter Baseline, Quadratic Model 2009: 8.1/4.0/5.4 Opioid dispensing rate, mg MED per member per quarter 2011: 7.8/3.8/5.2 (Relative change, %) Q42012: 7.7/3.8/5.1 1) -19 2) -39 3) 11 Age, n in millions 4) -100 ©Institute for Clinical and Economic Review, 2017 Page 160 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents 18-24: 0.9/1.0/10/1.0/1.0/ 1.0 5) -16 25-34: 1.7/1.7/1.8/1.8/1.7/ 1.7 Scenario 2: 8 Quarter Baseline, Linear Model 35-44: 2.0/2.0/2.0/2.0/1.8/ 1) -17 1.8 2) -41 45-54: 1.7/1.8/1.9/2.0/1.9/ 3) 4 1.8 4) -100 55-64: 0.9/1.1/1.2/1.3/1.4/ 5) -11 1.4 *organized by year (‘03/’05/’07/’09/’11/’12) ©Institute for Clinical and Economic Review, 2017 Page 161 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix G. Cost-Benefit and Budget Impact Supplemental Information Table G1: Massachusetts ER Opioid Market share – February 2016 to January 2017 ADF Opioids OxyContin 93.57% Embeda 1.80% Hysingla ER 4.24% Xtampza ER 0.39% Non-ADF Opioids Avinza Brand 0% Avinza Generic 0.35% Duragesic Brand 0.12% Duragesic Generic 17.54% Exalgo Brand 0.09% Exalgo Generic 0.85% Kadian Brand 0.04% Kadian Generic 2.22% Methadone HCl Generic 11.96% Morphine ER Generic 40.44% MS Contin 0.07% Nucynta 3.87% Opana ER Brand 1.61% Opana ER Generic 1.23% Oxycodone ER Generic 19.10% Zohydro ER Brand 0.50% Source: IMS Health8 ©Institute for Clinical and Economic Review, 2017 Page 162 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table G2: Opioid Strength and Number of Daily Doses to Reach 90mg Morphine Equivalent Dose (MED) Opioid Number of doses per day to reach 90mg MED ADF Opioids OxyContin 20mg 3 Embeda ER 30mg 3 Hysingla ER 30mg 3 Xtampza 18mg 3 Non-ADF Opioids Avinza ER 30mg 3 Avinza 30mg Generic 3 Duragesic 12mcg/hr 3 Fentanyl Patch 12mcg/hr 3 Generic Exalgo ER 8mg 3 Hydromorphone 8mg Generic 3 Kadian 30mg 3 Kadian 30mg Generic 3 Methadone 5mg 4 Morphine ER 30mg Generic 3 MS Contin 30mg 3 Nucynta 50mg 4 Opana ER 10mg 3 Oxymorphone ER 15mg 6 Oxycodone ER 20mg 3 Zohydro ER 30mg 3 Number of doses based on calculation for each opioid to reach 90mg MED Table G3: ICD-9 Opioid Diagnosis Codes for Identifying Opioid Abuse Patients Used by Rossiter et al., 2014 • 30400: Opioid Dependence-Unspecified • 30470: Opioid Other Dep-Unspecified • 30550: Opioid Abuse-Unspecified • 96500: Opium Poisoning Source: Rossiter et al., 201487 ©Institute for Clinical and Economic Review, 2017 Page 163 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table G4: All-Cause Mortality Death Prob. Male Death Prob. Female Death Prob. Pop. Population Age (in years) 55% 45% 37 0.001774 0.001038 0.001440592 38 0.001861 0.001113 0.001522156 39 0.001967 0.001196 0.001617737 40 0.002092 0.001287 0.001727335 41 0.00224 0.001393 0.001856309 42 0.002418 0.001517 0.002009847 Source: Social Security Administration88 Table G5: Incremental Costs of Diversion and Percentage Decrease in ADF Opioid Diversion Required to Achieve Cost-Neutrality Base diversion with non-ADF opioids Percentage decrease in diversion with ADF 125% (1.25:1) 100% (1:1) 75% (0.75:1) opioids 0% 0% $533,119,214 0% $533,119,214 0% $533,119,214 10% 10% $382,400,682 10% $412,544,389 10% $442,688,095 20% 20% $231,682,150 20% $291,969,563 20% $352,256,976 30% 30% $80,963,618 30% $171,394,738 30% $261,825,857 40% 35% $0 40% $50,819,912 40% $171,394,738 50% 44% $0 50% $80,963,618 60% 59% $0 70% 80% 90% 100% Source: ICER Calculation ©Institute for Clinical and Economic Review, 2017 Page 164 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table G6: Massachusetts Opioid Overdose Death Rate State Death rate per 100,000 population Massachusetts 17.0 Source: The Henry J. Kaiser Family Foundation96 Table G7: New England State-Specific ER Prescription Opioid Prevalent Users for Non-Cancer Pain in 2015 State ADF opioids Non-ADF opioids Total Massachusetts 60,222 113,045 173,267 Source: IMS Health8 Table G8: Commonwealth of Massachusetts Health Policy Commission Health Care Resource Utilization Claims Analysis Methods Inclusion Criteria • 18 years and over • ADF and non-ADF opioid prescriptions of 90 or more days Exclusion Criteria • Those receiving opioids as part of their cancer treatment were excluded; • Those that, within the calendar year, used both ADF and Non-ADF opioids; • Those with opioid pharmacy claims in the year, but without at least one medical claim. Source: Commonwealth of Massachusetts Health Policy Commission 91 ©Institute for Clinical and Economic Review, 2017 Page 165 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Table G9: ICD-9 Codes Used for Identifying Abuse Cases in Claims Analysis Conducted by Commonwealth of Massachusetts Health Policy Commission • 30400: Opioid Dependence-Unspecified • 30401: Opioid Dependence-Continuous • 30402: Opioid Dependence-Episodic • 30403: Opioid Dependence, In Remission • 30470: Opioid Other Dep-Unspecified • 30471: Opioid Other Dep-Continuous • 30472: Opioid Other Dep-Episodic • 30473: Opioid Other Dep-In Remission • 30550: Opioid Abuse-Unspecified • 30551: Opioid Abuse-Continuous • 30552: Opioid Abuse-Episodic • 30553: Opioid Abuse-In Remission • 96500: Opium Poisoning • 96509: Poisoning by Other Opiates and Related Narcotics • E8502: Accidental Poisoning by Other Opiates and Related Narcotics • E9352: Other Opiates and Related Narcotics Causing Adverse Effects in Therapeutic Use • 96501: Heroin Poisoning* • E8500: Accidental Poisoning by Heroin* • E9350: Adverse Effects of Heroin* *Only 11 patients matched these criteria and were hence merged with the opioid abuse patient groups Source: Commonwealth of Massachusetts Health Policy Commission 91 ©Institute for Clinical and Economic Review, 2017 Page 166 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix H. Conflict of Interest Disclosures for Expert Reviewers Name Title Disclosures Alan White Managing Principal I have worked on a number of projects on Analysis Group, Inc behalf of pharmaceutical companies, for which Analysis Group, Inc. has received compensation. Paul Gileno President Receipt or potential receipt of anything of U.S. Pain Foundation monetary value, including but not limited to, salary or other payments for services such as consulting fees or honoraria in excess of $5,000 Status or position as an officer, board member, trustee, owner or employee of a health care company, or an organization which receives more than 25% of its funding from health care companies • The U.S. Pain Foundation receives grants from health care companies to fund educational programming Lewis Nelson Professor and Chair, No conflicts to disclose. Emergency Medicine Rutgers New Jersey Medical School Richard Dart Director, Rocky Mountain I receive no personal payment for any activity Poison and Drug Center for any entity except Denver Health and Denver Health and Hospital Hospital Authority (DHHA). On behalf of DHHA, Authority I direct the RADARS System, which is supported Professor, University of Colorado School of by subscriptions from several pharmaceutical Medicine; Denver, Colorado manufacturers as well as the US FDA. No outside party participates in the design, collection, processing or reporting of the data. ©Institute for Clinical and Economic Review, 2017 Page 167 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Appendix I. Oral Public Comments This section includes summaries of the public comments prepared for the New England CEPAC Public Meeting on July 20, 2017 in Boston, MA. These summaries were prepared by those who delivered the public comments at the meeting and are presented in order of delivery. A video recording of these comments can be found on our site here at minute 1:25:00. Conflict of interest disclosures are included at the bottom of each statement for all participants who do not represent a manufacturer in this review. Gwendolyn Niebler, DO Senior Vice President, Clinical Research and Medical Affairs; Egalet Corporation Summary of Comments Delivered by Egalet at the July 20, 2017 Meeting of the New England CEPAC to Review ICER’s Evidence Report on Abuse-Deterrent Opioids According to the 2015 National Survey on Drug Use and Health, approximately 12.5 million people misused prescription pain relievers and 54% reported obtaining them from friends or relatives. This highlights the importance of diversion in the opioid epidemic. Solutions to this public health crisis must balance the need for access to effective analgesics for patients living with pain while minimizing the risk of opioid misuse and abuse that occurs in our communities. Opioid abuse is a chronic disease. In medicine, chronic illnesses are approached not only from a treatment perspective, but also from a preventive perspective. Abuse-deterrent formulations of opioids (ADFs) can play an important role as part of a multi-faceted preventative approach to the public health crisis of opioid abuse. While currently available ADFs cannot prevent all forms of abuse, they can play a substantial role in deterring abuse, especially with the more dangerous non- oral routes of abuse. Except for OxyContin®, the extended-release ADFs have either only recently become commercially available, like Arymo ER®, or access challenges have resulted in low market penetration, which accounts for the lack of “real world” effectiveness data for these products. Scott Gottlieb, FDA commissioner, recently remarked that the FDA strongly supports the transition from conventional opioids to a market in which the majority of opioids have meaningful abuse-deterrent properties. We encourage all stakeholders, including healthcare providers, insurance companies and legislators, to align with the FDA’s position on ADFs in order to have an impact on our country’s opioid epidemic. ©Institute for Clinical and Economic Review, 2017 Page 168 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Sunny Cho, PharmD Director, Medical Affairs; Daiichi Sankyo, Inc Daiichi Sankyo Public Comment Summary Daiichi Sankyo is thankful to ICER for bringing attention to the problem of opioid abuse and for recognizing the value of abuse-deterrent opioid formulations. We believe the baseline analysis of ICER’s report is incomplete as it does not consider the impact of abuse deterrent opioid formulations on family, friends and others in the community who are misusing or abusing diverted supplies of prescription opioids. Data show that opioid diversion is widespread, with more than half of abused prescription pain relievers obtained from a friend or relative and for those who initiated abuse, the most common source of pain relievers was also from a friend or relative. By reducing the potential for diversion and progression to advanced forms of abuse, we believe that abuse-deterrent formulations (ADFs) could also reduce the burden of associated costs. We support taking proactive steps to build a scientific base and framework to assess the intended impact of ADFs. This information will build on the existing tamper resistance and abuse liability data and help guide therapeutic decisions. Daiichi Sankyo stands behind the value of abuse deterrent formulations of opioid medications as part of a comprehensive approach to addressing the opioid epidemic. These medications give healthcare providers additional valuable tools in the fight against the potential for misuse and abuse while at the same time helping to provide patients relief from their pain. Tracy Mayne, PhD Head of Medical Affairs Strategic Research; Purdue Pharma, L.P. Did not submit comments for inclusion in this report. Shaina Smith Director of State Advocacy & Alliance Development; U.S. Pain Foundation U.S. Pain Foundation supports abuse deterrent formulations (ADFs) as an important step in deterring abuse and creating safer opioids. ADFs present a step forward in removing the problematic effect of opioid use and thus its appeal for abuse. We were discouraged to learn that the Institute for Clinical and Economic Review (ICER) did not view ADF technology as a way to help put a stop to the abuse of critical medications. ©Institute for Clinical and Economic Review, 2017 Page 169 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents Diversion mechanisms such as ADFs takes away the unintended and current issue of penalizing those who have legitimate need for pain relief. The draft evidence report failed to recognize the direct impact such findings will have towards both stigmatized populations: those with addiction disease and those with chronic pain disease. The report also failed to recognize the benefits of reducing diversion. As a mother with a rare, incurable chronic pain condition, I may be advised to consider opioid therapy. As a legitimate chronic pain patient who’d like to remain a contributing member of society, it’s only ethical I be provided with all available options, including ADFs. It’s my right to keep my family safe. It’s my right to manage my chronic pain. Should the final report remain status quo, insurers will interpret the document as law of the land, arguing it doesn’t need to cover ADFs for enrollees, despite the technology’s promising outcomes. Coverage and access to ADFs is not a privilege, it’s a right. Conflict of Interest Disclosures: Receipt or potential receipt of anything of monetary value, including but not limited to, salary or other payments for services such as consulting fees or honoraria in excess of $5,000; Status or position as an officer, board member, trustee, owner or employee of a health care company, or an organization which receives more than 25% of its funding from health care companies. The U.S. Pain Foundation receives grants from health care companies to fund educational programming. Sponsors listed from website include: Amgen, Abbott, AbbVie, Genentech, Pfizer, AstraZeneca, Teva, Purdue, Collegium, Depomed, Endo, McNeil, UCB, Shinogi, Daiichi Sankyo, GlaxoSmithKline, Johnson and Johnson, Mallinckrodt, Pernix Therapeutics, Kaleo, PhRMA Edmund Pezalla CEO; Enlightenment Bioconsult, LLC Thank you for this opportunity to address this important issue. I am the lead author on a soon to be published paper on this topic. With regards to a model examining ADF of long acting opioids: can the societal and payer views be reconciled? I believe so. ©Institute for Clinical and Economic Review, 2017 Page 170 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents A societal view is a broad one that includes both concerns for individual patients and for those around them. Just as in a contagion model, what happens to one person can influence others. In this case the influence is access to opioids through diversion. The impact of ADF opioids on someone already using, and perhaps abusing, opioids may be limited. However, limited opportunities for abuse of the drug may reduce illicit use by others who have not yet reached the point of abuse. At first glance it would appear that payers would be concerned solely with the individual patients and less concerned with those around them. However the usual unit of commercial insurance is the family, including children and young adults to age 26. Additionally, a given payer may have a high concentration of members in a given geographic area due to employment, etc. This exposes the payer to public health issues at the community level. In conclusion, diversion of opioids beyond the individual who has been prescribed the drug is both a community and a payer issue, and one that is in part addressed by abuse deterrent formulations and should therefore be considered in any economic or policy discussion. Conflict of Interest Disclosures: Receipt or potential receipt of anything of monetary value, including but not limited to, salary or other payments for services such as consulting fees or honoraria in excess of $5,000; Manufacturer support of research in the clinical area of this meeting in which you are participating. I am a consultant to Purdue on a regular basis advising on payer strategy and payment/access. Purdue has supported the article that I have authored along with Dr. Tracy Mayne on ADF opioids and modeling of impact. Dan Cohen Chair, Abuse Deterrent Coalition Abuse-deterrent formulations (ADFs) is the most substantive contribution to the fight against prescription drug abuse that innovators can offer. Having different understanding of the terms being used will yield a faulty decision and not appropriately address the questions before ICER. Yet the economic analysis being used today is incomplete and disconnected from reality as well. What should not under consideration today is ADF capabilities as an Abuse Prevention Formulation, or APF. Products with ADF technology do not, and are not, expected to prevent abuse of Scheduled products – only to lower the abuse potential of the products. ICER’s measure of the abuse ©Institute for Clinical and Economic Review, 2017 Page 171 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents prevention abilities in its cost/benefit analysis in factoring abusers deterred from continuing their abuse is a misapplication of the technology. The development of ADF is part of a multifactorial effort to reduce the risk of abuse and diversion. And we achieve effective therapies for patients while making abuse, misuse and diversion of important medications as difficult as possible within the bounds of known science. ADF is getting more effective – but we cannot get to future innovation by failing to pay for approve current discovery. Deterrence technologies is best understood as a technology the reduces the risk of misuse and diversion – by focusing to the point of exclusivity on the opiate naive and early stage recreational abusers. ADF is NOT a technology capable of effectively deterring a professional at manipulation, a desperate addict or experienced abuser. ADF is an early stage intervention that is one tree in the forest of effort of prescription drug abuse reduction. Conflict of Interest Disclosures: EVP, Government and Public Relations, KemPharm Inc. Richard Dart, MD, PhD Director, Rocky Mountain Poison and Drug Center; Denver Health and Hospital Authority It is important for the panel to understand the progression of prescription drug abuse from simple ingestion of intact tablets to snorting, injection and the outcomes of overdose, addiction and death. Abuse deterrent formulations (ADF) are often portrayed as affecting only the advanced abuser to decrease injection. More importantly, ADF impair abuse of an opioid when manipulation (chewing or crushing) is attempted. Thus they can decrease the progression from oral to riskier types of abuse. A systematic review of the effect of ADFs in the community showed that the reformulation of OxyContin is the only currently marketed drug with adequate data to evaluate real world effects. The introduction of reformulated OxyContin was followed by marked decreases in its abuse in Australia (2014), Canada (2012), and the United States (2010). At least 44 studies are available that show the effect was specific to the formulation, had substantial effect size; was consistency, specific, temporally related to decreased abuse with with no apparent or plausible alternative explanation. Various measures of abuse decreased 27% to 85% after reformulation. During the first 12 to 18 months after the reformulation, the sales and abuse of other opioid analgesics increased in contrast to OxyContin’s decrease. The observed changes were consistent across many different regions, investigators and 3 different countries and time frames. In the end, a prescriber of opioid ©Institute for Clinical and Economic Review, 2017 Page 172 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents analgesics needs the tools like prescribing guidelines, opioid analgesics with abuse deterrent properties and prescription drug monitoring plans to appropriately prescribe and monitor persons treated with opioid analgesics. Conflict of Interest Disclosures: Manufacturer support of research in the clinical area of this meeting in which you are participating; I am director of the RADARS System, which is operated by the Denver Health and Hospital Authority (governmental). The program is supported by subscription fees from multiple parties including many pharmaceutical companies and US government. All funds go the institution. No personal financial relationships are allowed. ©Institute for Clinical and Economic Review, 2017 Page 173 Final Evidence Report—Abuse-deterrent Formulations of Opioids: Effectiveness and Value Return to Table of Contents