HHS-OIG Data Snapshot National Review of Opioid Prescribing in Medicaid Is Not Yet Possible OEI-05-18-00480 August 2019 What OIG Did Historically, national Medicaid data—a collection of data submitted by all States—have Key Takeaway not been complete, accurate, and timely. These data have not yet been adequate for A national review of national analysis and oversight, even though some States’ data have been sufficient for opioid prescribing individual State analysis.1 Because of concerns with the quality and completeness of in Medicaid using the national Medicaid claims database—the Transformed Medicaid Statistical T-MSIS is not yet Information System (T-MSIS)—we assessed the completeness of variables needed to possible because monitor national opioid prescribing in Medicaid. We assessed variables needed to not all at-risk identify (1) beneficiaries at risk of opioid misuse or overdose (i.e., variables needed to beneficiaries and calculate beneficiaries’ total opioid dosage, and diagnosis codes to exclude patients providers can be for whom higher doses of opioids may be appropriate) and (2) the National Provider identified. Identifiers (NPIs) of providers that ordered and dispensed opioids (i.e., prescribers and pharmacies, respectively). Results Limitations of T-MSIS data impede identification Nineteen States were missing pharmacy or of individual beneficiaries for national opioid prescriber NPI—most frequently, pharmacy NPI. analysis. A given Medicaid beneficiary can have multiple IDs within a State (e.g., if he/she disenrolled and later re-enrolled) and across States. (Because Medicaid IDs are assigned at the State level, a beneficiary gets a new ID if he/she moves to another State.) If a beneficiary has multiple IDs, the prescriptions dispensed to those IDs would erroneously appear to be for multiple people, not one person. As a result, beneficiaries’ total opioid dosages would be NOTE: Three States were missing both types of NPIs. undercounted. Source: OIG analysis of T-MSIS data, 2019. Thirty-two States were missing NPI, diagnosis Examples of reasons States were missing data: code, or quantity.  Did not require NPI to be collected  Have NPI in their State data but (1) submit it to the wrong field or file in T-MSIS or (2) are not able to transmit it to T-MSIS because they are operating with an outdated system  Do not report diagnosis codes for all services NOTE: Five States were missing NPI AND diagnosis code. even though the variable is required Source: OIG analysis of T-MSIS data, 2019. Data Snapshot can be found at oig.hhs.gov/oei/reports/oei-05-18-00480.asp 1 Why This Matters Until T-MSIS data are complete in all States and limitations across States are addressed, it will not be possible to conduct a national evaluation of Medicaid beneficiaries at risk of opioid misuse or overdose. According to CMS, Medicaid covered over 31 million prescriptions for opioids in 2017.2 The data we reviewed are critical for nationally quantifying the opioid crisis’s impact on Medicaid and for monitoring the crisis, as well as for conducting general program integrity efforts across States.  Without a unique beneficiary ID, it is not possible to identify all at-risk beneficiaries in need of opioid-related treatment and to take appropriate action, or to monitor utilization of services to protect beneficiaries from poorly coordinated care.  Without a provider NPI, it is not possible to identify all providers who may be overprescribing opioids and take appropriate action, or to identify providers for investigations of fraud, waste, or abuse.  Without a diagnosis code, it is not possible to exclude all patients with cancer diagnoses for whom higher doses of opioids may be appropriate, or to identify patients’ medical conditions to determine medical necessity for services. In August 2018, the Centers for Medicare & Medicaid Services (CMS) announced that all States were submitting T-MSIS data, that CMS was prioritizing T-MSIS data quality, and that CMS would have research-ready files available in 2019.3 Since then, CMS has been working with States to improve the quality of their data submissions, so some States’ data may have improved since we pulled our data in December 2018. Also, in May 2019, CMS added three of the variables we reviewed—diagnosis code, drug quantity, and pharmacy NPI—to its priorities for data quality.4 What OIG Recommends To ensure the identification of at-risk beneficiaries and providers who may be overprescribing, CMS should: Work to ensure that individual Ensure the correct submission of prescriber NPIs beneficiaries can be uniquely identified CMS should prioritize State reporting of prescriber NPIs. at a national level using T-MSIS CMS recently prioritized completeness of pharmacy NPIs, CMS should work with States to address but not that of prescriber NPIs. instances in which a single beneficiary has CMS concurred with this recommendation and will prioritize more than one Medicaid ID within a State. completeness of prescriber NPIs. Additionally, CMS should work to ensure that in cases in which a beneficiary was enrolled in more than one State over time, claims for individual beneficiaries can be linked across States. Clarify requirements for diagnosis codes CMS should issue guidance to clarify which services require CMS concurred with this recommendation and a diagnosis code. CMS recently prioritized completeness of will implement a process to enable IDs to be diagnosis codes, but some States are unsure whether linked. CMS will also issue guidance to States a diagnosis code is required for all services. on assignment and coding of unique IDs, and to data users on identification of individual CMS concurred with this recommendation and will revise beneficiaries at the national level. guidance to clarify requirements. Data Snapshot: National Review of Opioid Prescribing in Medicaid Is Not Yet Possible 2 Data: Percentage missing for each variable, by State NOTE: Cells highlighted in blue indicate variables for which 100 percent of the corresponding data were missing. Cells highlighted in gray indicate variables for which the percentage of corresponding data missing was greater than 10 percent but less than 100 percent. *We averaged the percentage missing for beneficiary ID across all claims files. Source: OIG analysis of T-MSIS data, 2019. Data Snapshot: National Review of Opioid Prescribing in Medicaid Is Not Yet Possible 3 Methodology We assessed the completeness of the T-MSIS variables that would be needed to monitor national opioid prescribing in Medicaid: prescription information, provider NPIs, and diagnosis codes.  We did not assess the completeness of the 46 variables that can be used to identify hospice care.  Review period: We reviewed the December 2018 data for claims with dates of service between January 2017 and March 2018.  We categorized variables as “missing” in a State if the corresponding data fields were blank for more than 10 percent of the State’s claims in our review period.  We excluded certain claims: o denied claims, o financial transactions (i.e., capitation payment, supplemental payment, and service tracking claims), and o claims in which the variables we selected were not required (e.g., we excluded claims that do not require a diagnosis code, such as claims for durable medical equipment, laboratory services, and transportation services)  We conducted interviews about missing data with 10 States’ Medicaid officials and T-MSIS staff to find out why they were missing data. Standards We conducted this study in accordance with the Quality Standards for Inspection and Evaluation issued by the Council of the Inspectors General on Integrity and Efficiency. Acknowledgments Hilary Slover served as the team leader for this study, and Joy Rooney served as the lead analyst. Others in the Office of Evaluation and Inspections who conducted the study include Nicole Hrycyk and Lisa Minich. Office of Evaluation and Inspections staff who provided support include: Clarence Arnold, Randi Hall, Althea Hosein, Kevin Manley, Christine Moritz, Kayla Phelps, and Keirsha Thompson. We would also like to acknowledge the contributions of other OIG staff, including Mandy Brooks and Jessica Swanstrom. This report was prepared under the direction of Thomas Komaniecki, Regional Inspector General for Evaluation and Inspections in the Chicago regional office, and Laura T. Kordish and Kelly Waldhoff, Deputy Regional Inspectors General. To obtain additional information concerning this report or to obtain copies, contact the Office of Public Affairs at Public.Affairs@oig.hhs.gov. Sources 1 OIG, Opioids in Ohio Medicaid: Review of Extreme Use and Prescribing, OEI-05-18-00010, July 2018; OIG, Status Update: T-MSIS Data Not Yet Available for Overseeing Medicaid, OEI-05-15-00050, June 2017; OIG, Early Outcomes Show Limited Progress for the Transformed Medicaid Statistical Information System, OEI-05-12-00610, September 2013; OIG, MSIS Data Usefulness for Detecting Fraud, Waste, and Abuse, OEI-04-07-00240, August 2009. 2 CMS, Medicaid Opioid Prescribing Mapping Tool, April 23, 2019. Accessed at https://www.cms.gov/Research-Statistics-Data-and- Systems/Statistics-Trends-and-Reports/Medicare-Provider-Charge-Data/OpioidMap_Medicaid_State.html on June 14, 2019. 3 CMS, State Health Official letter SHO # 18-008, August 10, 2018. Accessed at https://www.medicaid.gov/Federal-Policy- Guidance/downloads/SHO18008.pdf on February 5, 2019. 4 CMS, CMS Guidance: Overview of Data Quality Top Priority Items, May 2019. Accessed at https://www.medicaid.gov/medicaid/data- and-systems/macbis/tmsis/tmsis-blog/?entry=51423 on May 21, 2019. Data Snapshot: National Review of Opioid Prescribing in Medicaid Is Not Yet Possible 4 Agency Comments Data Snapshot: National Review of Opioid Prescribing in Medicaid Is Not Yet Possible 5 4 Data Snapshot: National Review of Opioid Prescribing in Medicaid Is Not Yet Possible 6 4