Background Paper November 27, 2006 Medicare and Mental Health: The Fundamentals Christopher W. Loftis, PhD, Research Associate Eileen Salinsky, Principal Research Associate OVERVIEW — This background paper provides a review of mental health coverage in the Medicare program. It examines the prevalence of mental disorders among Medicare beneficiaries, treatment available through Medi- care, and the cost of such treatment. A brief summary of relevant policy issues is provided, including Medicare’s outpatient mental health limita- tion and the potential effect of the prescription drug benefit on the provi- sion of mental health services. NATIONAL HEALTH POLICY FORUM FACILITATING DIALOGUE. FOSTERING UNDERSTANDING. Background Paper November 27, 2006 CONTENTS INTRODUCTION ................................................................................ 3 MENTAL DISORDERS AND MEDICARE BENEFICIARIES ....................... 4 Mental Disorders Among Disabled Beneficiaries ............................ 6 Figure 1: Prevalence of Mental Disorders ...................................... 6 Figure 2: Disabled Medicare Beneficiaries Under 65: Medicaid Eligibility and Health Care Costs, 1995 .......................... 7 Mental Disorders Among Aged Beneficiaries................................. 7 Figure 3: Distribution of Medicare Beneficiaries with Mental Disorders, by Eligibility Class ............................................. 7 MEDICARE POLICIES RELATED TO MENTAL HEALTH SERVICES .......... 9 Inpatient Hospital Services .......................................................... 10 Partial Hospitalization ................................................................. 12 Outpatient Services ..................................................................... 13 Prescription Drugs ...................................................................... 17 Figure 4: Average Annual Drug Spending Among Medicare Beneficiaries ................................................................ 18 Table 1: Percentage of Medicare Beneficiaries Using Psychotropic Drugs by Therapeutic Category and Eligibility, 2002 .................................................................... 19 Figure 5: Sources of Prescription Drug Spending for Medicare Beneficiaries, 2002 ................................................. 20 Care Management ..................................................................... 21 MEDICARE SPENDING FOR MENTAL HEALTH SERVICES .................. 22 Figure 6: Mental Health Spending by Payer Class, 2001 .............. 23 Figure 7: Psychotropic Drug Expenses as a Percentage of National Health Policy Forum Overall Mental Health Spending, 1987–2001.............................. 24 Facilitating dialogue. Fostering understanding. CONCLUSION ................................................................................. 24 ENDNOTES ..................................................................................... 25 2131 K Street NW, Suite 500 Washington DC 20037 202/872-1390 202/862-9837 [fax] nhpf@gwu.edu [e-mail] www.nhpf.org [web] Judith Miller Jones Director Sally Coberly Deputy Director Monique Martineau Publications Director National Health Policy Forum | www.nhpf.org 2 Background Paper November 27, 2006 Medicare and Mental Health: The Fundamentals The Medicare prescription drug benefit has drawn attention to the mental health care needs of Medicare beneficiaries. This visibility is likely to be- come more pronounced once the initial tab for psychotropic drugs under Part D is tallied. Press coverage highlighted the transition problems faced by mentally ill beneficiaries dually eligible for Medicare and Medicaid, as well as the high prices paid by Medicare drug plans for psychotropic drugs. Although psychotropic drugs are an important dimension of mental health treatment, providing appropriate mental health services to Medicare’s disabled and aged beneficiaries is a broader challenge than simply ensuring access to medications. Many mental health professionals contend that Medicare’s benefit structure is not optimally designed to sup- port the kinds of mental health services demonstrated to be most effective for disabled and older Americans. As the prescription drug utilization and spending patterns of Medicare beneficiaries unfold over time, a broader examination of Medicare’s mental health coverage and payment policies may occur. When Medicare was first established in 1965, its benefit package was closely modeled after the typical commercial health insurance product of the time—a product that offered very limited coverage for mental health services. Traditional health insurance policies often singled out mental health services for more restricted benefits because of concerns that prevailing treatments were of questionable efficacy and that subjective diagnostic standards made the benefit particularly vulnerable to overuse. In creating the Medicare pro- gram, few policymakers saw little impetus to depart from this conventional wisdom regarding the structure of mental health benefits. Over the past 40 years, a number of changes have taken place in the financing and delivery of mental health services. Science has become increasingly ad- ept at explaining how behavioral manifestations of mental illness are linked to biological mechanisms that disrupt brain and nervous system functions. As this evidence base has expanded, the diagnosis and treatment of mental health disorders has improved. Scientific and clinical advances, combined with legislative mandates and a growing understanding of how mental disorders influence general health status, have also increased insurers’ willingness to pay for mental health treatments. A variety of employer- sponsored health insurance products have evolved to provide more extensive coverage for mental health while managing utilization in more sophisticated ways. Despite these changes, Medicare’s coverage for mental health services has evolved only modestly since the program’s inception. National Health Policy Forum | www.nhpf.org 3 Background Paper November 27, 2006 Medicare’s benefit structure has, until recently, con- cealed the impact of mental illness on beneficiary A Word of Caution health, service use, and overall program costs. Prevalence estimates for mental disorders Mental disorders are at least as prevalent among within the Medicare population are likely to beneficiaries as they are in the general population, be understated. Prevalence of mental disor- yet Medicare spending on mental health has not ders in the general adult population has been matched that of other payers. Program policies that determined through methodologically robust do not support preferred models of care have likely epidemiological studies that relied on standard affected beneficiaries’ treatment and outcomes for clinical diagnostic criteria to identify persons some time. Outpatient mental health coverage with mental disorders and assessed the sever- restrictions, originally intended to minimize inap- ity of confirmed cases. In contrast, prevalence propriate service utilization, have the potential to estimates for the Medicare population rely foster an excessive reliance on pharmacological solely on service utilization history to identify interventions and could contribute to higher utili- persons with mental health diagnoses. In reality, zation of other, more expensive health care services, people who may not actually require treatment such as hospitalizations. receive services—and people who need care do New spending on psychotropic drugs and demo- not receive it—yielding imprecise estimates of graphic changes in the beneficiary population will disease prevalence as a result of both over- and soon collide to raise the prominence of mental undercounting of cases. Rigorous studies of the health issues within Medicare. The number of Medi- prevalence of mental disorders in the general care beneficiaries with mental illness is expected to population indicate that undercounting domi- grow substantially as a result of a variety of factors, nates, suggesting that prevalence estimates for including the large cohort of “baby boomers” ag- Medicare beneficiaries are conservative. ing into the program, longer life expectancies and the increasing prevalence of dementia and other cognitive disorders that come with advanced age, and an increase in the number of Americans qualifying for Social Security benefits (and thus for Medicare) on the basis of disability due to mental illnesses.1 At the same time, Part D–related expenses for psychotropic drugs are likely to more than double current mental health spending by Medicare. The convergence of these trends suggests a large increase in mental health spending over the next 15 years. Managing the cost and quality of care for chronic diseases in a largely fee-for-service context has long proved to be a difficult struggle for Medi- care—and this struggle is abundantly apparent with respect to mental disorders. The diverse mix of services, providers, and care modalities inherent in mental health treatment, along with the cognitive limitations of people with mental disorders, combine to make utilization management and care coordination especially challenging. Mental disorders AND Medicare Beneficiaries Medicare beneficiaries appear as likely to experience mental health problems as the general adult population.2 Over a 12-month period, ap- proximately 26 percent of Medicare beneficiaries report having some type of mental disorder,3 compared to 26.2 percent of all noninstitutionalized National Health Policy Forum | www.nhpf.org 4 Background Paper November 27, 2006 Growth in Disability Awards Related to Mental Disorders Beneficiaries with disabilities account for a relatively small proportion of all Medicare ben- eficiaries, but this group has grown rapidly in SSDI Awards to Disabled Workers recent years. The number of disabled Medicare 800,000 beneficiaries more than doubled between 1990 and 2005, with increasing numbers of disability awards due to mental illness being a major source of this 600,000 growth. Disabled beneficiaries currently represent nearly 14 percent of all persons who are covered by Medicare. Growth in the number of beneficiaries 400,000 with disabilities has been attributed to a number Total SSDI of policy changes rather than changes in underly- Awards ing disease prevalence. Policy changes clarified 200,000 Social Security’s disability determination criteria, Mental Disorder resulting in more approvals and fewer termina- 0 Awards tions. Several of these changes focused specifically Years: 1960 – 2004* on liberalizing the disability determinations for mental disorders, such as tying determinations *Data were not available for the years 1964, 1965, 1966, 1974, 1979, to functionality assessments and mandating re- and 1980. The trends are shown as continuous, but the data have been extrapolated. view by a specialty mental health provider in all appealed denials. People judged to be disabled Source: Social Security Administration, “Annual Statistical Report on the Social Security Disability Insurance Program, 2004,” March 2006, and awarded Social Security Disability Insurance p. 98; available at www.ssa.gov/policy/docs/statcomps/di_asr/2004/index. benefits are eligible for Medicare on the basis of html#editions. their disability. adults over the age of 18.4 These prevalence estimates reflect the inclusive definition of mental disorder used in the Diagnostic and Statistical Manual of Mental Disorders (DSM), which identifies a wide variety of mental health diagnoses, including mood disorders (such as depression), anxiety dis- orders, psychoses, cognitive impairments (such as Alzheimer’s disease), and substance abuse. The severity of these disorders can vary significantly, both within and across diagnostic categories. As is the case in the general adult population, a relatively small proportion of all Medicare beneficiaries experience severe mental disorders (such as schizophrenia or major depression) that result in significant functional impairment. About 9 percent of Medicare beneficiaries have a severe mental disorder, compared to 6 percent of the general adult population (Figure 1, following page).5 However, the overall prevalence rate of severe illness among Medicare beneficiaries obscures important differences in the burden of disease in the aged beneficiary population relative to those who are eligible for Medicare due to disability. National Health Policy Forum | www.nhpf.org 5 Background Paper November 27, 2006 Mental Disorders Among Disabled Beneficiaries Severe mental illness is highly prevalent among Medicare beneficiaries under the age of 65 who qualify for Medicare because of a disability. Ap- proximately 37 percent of disabled Medicare beneficiaries have a severe mental disorder,6 compared to only 4 percent of aged beneficiaries.7 The high rate of mental disorders in the disabled Medicare population is pri- marily driven by the large proportion of disabled beneficiaries who qualify for Medicare because of a long-lasting and disabling mental illness. Since 1987, both the number and proportion of disability awards due to mental illness have increased dramatically. Mental disorders were the leading reason disabled workers received Social Security Disability Insurance (SSDI) in 2004.8 People who qualify for Medicare based on a physical disability may also have comorbid, potentially severe, mental disorders, and the prevalence of all types of mental disorders appears high for this population. An estimated 59 percent of all disabled beneficiaries have some type of mental disorder, compared to 21 percent of aged beneficiaries.9 The rate of severe mental illness among the physically disabled is not well documented. Even with Medicare’s limited benefits, care for disabled beneficiaries with mental disorders is costly. Excluding persons with end-stage renal disease, figure 1 Prevalence of Mental Disorders Sources: William E. Narrow et al., “Revised Prevalence Estimates of Mental Disorders in the United States,” Archives of General Psychiatry, 59, Feb 2002; Ronald C. Kessler et al., “Prevalence and Treatment of Mental Disorders, 1990 to 2003,” New England Journal of Medicine, 352, no. 24 (June 15, 2005), pp. 2515–2523; Henry J. Kaiser Family Foundation, Medicare Chartbook (Menlo Park, CA: Henry J. Kaiser Family Foundation, 3rd ed., Summer 2005), available at www.kff.org/medicare/upload/ Medicare-Chart-Book-3rd-Edition-Summer-2005-Report.pdf.; Ronald C. Kessler et al., “Estimating the Prevalence and Correlates of SMI in Community Epidemiological Studies,” chap. 12 in Mental Health, United States, 2002, Ronald W. Manderscheid and Marilyn J. Henderson, (Eds.), DHHS Publication No. SMA04-3938 (Rockville, MD: U.S. Department of Health and Human Services, 2002), available at http://mentalhealth.samhsa.gov/publications/allpubs/sma04-3938/Chapter12.asp; and Sandra M. Foote and Christopher Hogan, “Disability Profile And Health Care Costs Of Medicare Beneficiaries Under Age Sixty-Five,” Health Affairs, 20, no.6 (November/December, 2001), pp. 242–253, available at http:// content.healthaffairs.org/cgi/reprint/20/6/242.pdf. National Health Policy Forum | www.nhpf.org 6 Background Paper November 27, 2006 figure 2 Disabled Medicare Beneficiaries Under 65: Medicaid Eligibility and Health Care Costs, 1995 Medicaid Eligible Beneficiary Type Medicare Cost Per Person Per Year 43% All Disabled $ 3,832 DISABLED WITH: 62% Any Mental Disorder $ 4,689 73% Mental Retardation $ 3,872 53% Severe Mental Illness $ 4,454 59% Dementia $11,488 Source: Sandra M. Foote and Christopher Hogan, “Disability Profile And Health Care Costs Of Medicare Beneficiaries Under Age Sixty-Five,” Health Affairs, 20, no.6 (November/December, 2001), pp. 245; available at http://content.healthaffairs.org/cgi/reprint/20/6/242.pdf. . Medicare’s per capita spending on disabled beneficiaries is figure 3 somewhat lower than its spending on aged beneficiaries.10 Distribution of Medicare However, disabled beneficiaries with mental disorders are Beneficiaries with Mental Disorders, more costly for the Medicare program than the typical dis- by Eligibility Class abled beneficiary: Per capita spending on disabled benefi- ciaries with mental disorders is over 22 percent higher than Any Mental Serious Mental average spending per disabled beneficiary.11 (See Figure 2 for Disorder 100% Illness Medicare spending on disabled beneficiaries across types of mental disorders.) 80 44% Aged Mental Disorders Among Aged Beneficiaries Aged 65% Aged beneficiaries are far less likely to exhibit mental dis- 60 orders than disabled beneficiaries but, because of the size of the aged population, most of the Medicare beneficiaries with 40 mental disorders are over the age of 65. Prevalence of serious mental illness is more equally divided across aged and dis- abled Medicare beneficiaries. Nearly half of all beneficiaries 56% Disabled 20 with severe mental illness and about two-thirds of those Disabled 35% with any form of mental disorder are aged. (See Figure 3 for distribution of Medicare beneficiaries with mental disorders.) 0% Significant cognitive impairments, such as dementia, clearly Source: Authors’ calculations based on disease prevalence rates contribute to the prevalence of severe mental illness among cited in this paper and population statistics in CMS, “Medicare Enrollment: National Trends—Medicare Aged and Disabled aged beneficiaries, but the extent to which other severe psy- Enrollees by Type of Coverage,” 2005; available at www.cms.hhs. chiatric conditions influence disease prevalence has not been gov/MedicareEnRpts/Downloads/HISMI05.pdf. Continued on p. 9  National Health Policy Forum | www.nhpf.org 7 Background Paper November 27, 2006 Disability, Medicare, and Medicaid Workers who become severely disabled before age 65 and can no longer work are eligible for Social Security Disability Insurance (SSDI) payments after five months of unemployment. SSDI recipients are subsequently eligible for Medicare coverage after they qualify and receive SSDI benefits for 24 months. Approximately 34 percent of all such disabled adults with mental disorders are Medicare beneficiaries. For the Social Security Administration (SSA) “disability” is the inability to work at a job that pays at least $900 per month (in 2007) because of a medically determinable physical or mental impairment that is expected to result in death or to continue for at least 12 months. Not all physical and mental conditions Source of Health Insurance Coverage are included. For example, drug addiction and for Persons Disabled by Mental Disorders alcoholism are not qualifying conditions. Fur- ther, people with serious conditions meet the 54% Medicaid 35% Medicare criteria only after their disease is in an advanced stage. Maintaining disability status can be par- ticularly difficult for mentally ill persons who 13% Medicaid 12% can be stabilized and able to work when they & Medicare Medicare are receiving appropriate medical treatment. 10% Dually eligible — Many low-income ben- Medicare & Private 41% eficiaries who qualify for Medicare due to Medicaid a disability also qualify for Medicaid. Most states provide Medicaid coverage for aged, blind, and disabled individuals with limited income who qualify for Supplemental Security 6% 2% 15% Uninsured Other Private Income (SSI) from SSA. There is no wait time for SSI eligibility. Approximately 35 percent Source: Richard G. Frank and Sherry Glied, “Changes In Mental Health of SSI recipients qualify as a result of a mental Financing Since 1971: Implications For Policymakers And Patients,” Health Affairs, 25, no.3 (May/June, 2006), p. 606. disorder. Approximately 13 percent of all dis- abled adults with mental disorders are dually eligible for Medicare and Medicaid. Nearly 40 percent of disabled Medicare beneficiaries with mental disorders are also eligible for Medicaid, compared to 30 percent of disabled beneficiaries who do not have a mental disorder. Eligible for Medicaid but NOT Medicare — Some disabled people are ineligible for SSDI and Medicare because they do not have a sufficient work history but they may be eligible for SSI and Medicaid, depend- ing on their income level and their state’s eligibility criteria. This group also includes disabled children and people who receive SSDI but have not yet met the two-year waiting requirement for Medicare benefits. About 40 percent of all disabled adults with mental disorders are eligible for Medicaid but not Medicare. Eligible for Medicare but NOT Medicaid — Disabled persons with a work history who have fulfilled the two year wait period will be eligible for Medicare, but some of these individuals will be ineligible for Medicaid because their income or assets are too high. Approximately 21 percent of all disabled adults with mental disorders are eligible for Medicare but not Medicaid. Ineligible for Medicare and Medicaid — Some disabled people will not qualify for Medicaid or Medicare. This group includes those in the process of fulfilling the two-year waiting period, as well as those who do not meet the SSA criteria for disability determination. Approximately 22 percent of disabled adults with mental disorders are ineligible for either Medicare or Medicaid. National Health Policy Forum | www.nhpf.org 8 Background Paper November 27, 2006 Continued from p. 7 fully documented. Available data suggest that one-quarter to one-half of severe mental disorders among the aged are due solely to cognitive limi- tations. Beneficiaries with dementia may have co-occurring psychiatric disorders, such as depression. The prevalence of other serious mental disorders, such as schizophrenia, appears to be lower in the aged popula- tion, most likely because people with these conditions have high rates of premature mortality and life-spans that are, on average, 20 years shorter than the U.S. average.12 Although access concerns often center on the needs of vulnerable individu- als with severe mental disorders, mental disorders among aged benefi- ciaries are more frequently mild to moderate in severity. The three most common mental disorders experienced by Medicare beneficiaries over the age of 65, in order of prevalence, are anxiety, dementia or other cognitive impairments, and depression. The prevalence of dementia doubles every 5 years beyond age 65, becoming the most common mental health diagnosis after age 80.13 The severity of more common disorders, such as dementia, can vary considerably, and only a small proportion of those afflicted are likely to face significant functional impairments. While the most common types of mental disorders are not usually severe and disabling, people with these mild to moderate mental health diagnoses tend to have high rates of other illnesses. The relationship between mental health disorders and physical disorders such as diabetes, heart disease, and cancer is not fully understood. In some cases, a single disease process may be fundamentally responsible for both mental and physical symptoms. In other cases, underlying mental disorders may undermine adherence to medical protocols (for example, medication management) and healthy lifestyle choices, such as diet and exercise, which are known to influence the onset and management of other chronic diseases. In still other cases, preexisting physical disorders may lead to decreased functionality and increased social isolation, which may subsequently trigger mental health complications. Left untreated, mild to moderate mental disorders can escalate into more serious mental conditions, complicate the treatment of physical health conditions, compromise patient outcomes, and increase the cost of care. Medicare Policies Related to Mental Health Services Medicare covers a wide range of mental health services, but special restric- tions and limitations are often applied to differentiate psychiatric benefits from comparable physical health services. The impact of Medicare’s cov- erage and payment policies on access to mental health services, quality of care, and patient outcomes is frequently unclear and sometimes hotly debated. The following reviews key issues related to Medicare’s coverage of and payment for mental health services. National Health Policy Forum | www.nhpf.org 9 Background Paper November 27, 2006 Inpatient Hospital Services The acute care services needed to diagnose and stabilize individuals in a mental health crisis have traditionally been provided in inpatient settings. Follow-up treatment planning and referrals for monitoring or additional treatment in the community may also begin in an inpatient facility. Care is delivered through the combined expertise of multidisciplinary teams that include, among others, nurses, psychiatrists, psychologists, and social workers. The relative contributions and mix of disciplines appears to vary considerably across inpatient settings and geographic areas, depending upon care models, patient mix, and workforce characteristics. Many mental health experts believe that most The Medicare statute imposes a 190-day life- inpatient psychiatric admissions could be avoided if appropriate community-based ser- time limit on services received in specialty vices were utilized properly. However, such psychiatric hospitals. services are often in short supply or difficult to access. The nature of mental disorders can lead patients to delay, resist, or refuse necessary treatment until a crisis precipitates intervention. In some cases, patients can be hospitalized involuntarily if a court determines that their mental illness has resulted in an immediate danger to themselves or others. Involuntary hospitalizations typically require admission to a locked, secure psychiatric unit. Life-time limit on inpatient psychiatric services — Under Part A, Medi- care pays for inpatient psychiatric care provided in both freestanding psychiatric hospitals and in dedicated psychiatric units of general acute hospitals. These providers are called inpatient psychiatric facilities (IPFs). Medicare will also cover inpatient psychiatric services provided in nonspe- cialized, medical/surgical beds of general hospitals (commonly referred to as “scatter beds”). Medicare pays for 100 percent of inpatient psychiatric services, subject to a deductible and copayment for extended stays. Medicare does not cover long-term custodial care for any condition and imposes a 90-day limit on all inpatient hospital coverage for each episode of illness (known as a “benefit period,” which begins on the day the ben- eficiary enters a hospital or skilled nursing facility and ends when the beneficiary has been out of the hospital for 60 consecutive days).14 This limit applies to both psychiatric and general inpatient services. For general hospital services, there are no coverage limits on the total number of benefit periods a beneficiary is entitled to in his or her lifetime. However, Medicare limits lifetime treatment in freestanding IPFs to 190 days.15 Stays in general hospitals for psychiatric services are not subject to the 190-day limit. The lifetime psychiatric limit was intended to limit the federal government’s role in paying for long-term custodial support of the mentally ill. The limit was not imposed on hospital-based units because these facilities have only rarely been used for long-term care purposes. The extent to which the 190-day limit hinders access to inpatient psychi- atric services is not clearly understood. Although few beneficiaries reach National Health Policy Forum | www.nhpf.org 10 Background Paper November 27, 2006 the 190-day lifetime limit, the policy is most likely to affect beneficiaries disabled by mental illness. These beneficiaries have more frequent psy- chiatric hospitalizations in a given year and are more likely to be enrolled in Medicare for a longer period of time than aged beneficiaries. Although beneficiaries who reach the 190-day limit could still be covered for psychi- atric care provided in general hospitals, consumer advocates and provid- ers are concerned that the limit may undermine choice of physician and jeopardize access to care, particularly in communities where the general hospitals lack psychiatric capacity. Prospective payment — Medicare recently revised its reimbursement policy for IPFs, moving in 2005 from a cost-based system, with limits, to a per diem prospective payment rate, referred to as the inpatient psychiatric facility prospective payment system (IPF-PPS).16 The per diem rate varies, depending on the patient’s psychiatric diagnosis-related group—which accounts for cost differences across patients that are due to their diagno- ses, procedures, and other factors—and on whether the patient has any of 17 coexisting general medical conditions that are associated with higher average costs. The per diem rate is higher in the early days of a psychiatric stay, in recognition of the costs associated with stabilizing a psychiatric patient.17 Care provided in scatter beds is reimbursed under the per case hospital inpatient prospective payment system (PPS). The impact of the IPF-PPS on access to care is uncertain at this time, since it is still in its three-year phase-in period. Psychiatric providers are cautiously monitoring its effect on a market that has experienced significant fluctuation in inpatient psychiatric capacity over the past two decades. Some worry that the IPF-PPS may encourage general hospitals (which tend to have higher av- erage cost structures than freestanding facilities) to further reduce inpatient psychiatric capacity in favor of more reliably profitable medical services or to shift beds from specialized units to scatter beds, which will not be reim- bursed under IPF-PPS. In 1995, nearly one-third of all Medicare inpatient psychiatric discharges were from scatter beds; this proportion has increased as dedicated psychiatric capacity in both freestanding facilities and general hospitals has decreased over the last decade. Some experts have expressed concerns that the care delivered through scatter beds is not as effective as the care delivered in more specialized facilities. Adding to their concern is the inability of most scatter beds to accommodate involuntary admissions because they lack the security measures required to contain patients who have been hospitalized against their will. Others believe that the payment system may be advantageous to freestand- ing hospitals that have lower cost structures and more flexibility to target services to patient needs. Some believe that, while less efficient providers might eliminate psychiatric services, organizations that have developed the expertise and economies of scale to deliver psychiatric services in the most efficient manner could expand their operations, resulting in little overall change in capacity levels. National Health Policy Forum | www.nhpf.org 11 Background Paper November 27, 2006 Partial Hospitalization Partial hospitalization programs, used in lieu of inpatient hospitalizations, provide structured, intensive outpatient services. These programs, some- times referred to as day programs, offer a range of medically appropriate services “wrapped around the particular needs of the patient.”18 The services generally reflect a multidisciplinary team approach to patient care under the direction of a physician and are typically much more intensive than those offered through the private practice of an individual mental health provider, such as a psychiatrist or psychologist.19 Partial hospitalization programs do not provide around-the-clock nursing care, but supervised residential facilities may be offered through affiliated organizations. Some advocates question the effectiveness of partial hospitalization pro- grams. The lack of around-the-clock supervision may make these programs inappropriate for patients who have not been stabilized. At the same time, the institutional nature of these programs hinders the development of behavioral skills and coping strategies in a real-world, community-based context. Furthermore, partial hospitalization programs often do not include the kinds of psychosocial services (such as vocational training and hous- ing assistance) that are known to support recovery from mental disorders because such programs tend to focus solely on clinical services that are reimbursable through Medicare and commercial insurers. Regulatory concerns — Partial hospital services were added as an explicit Medicare-covered benefit through the Omnibus Budget Reconciliation Act of 1987. Previously, mental health services delivered through hospital- based partial hospitalization programs were sometimes covered as a Part A (inpatient) benefit, and freestanding partial hospitalization programs were sometimes covered as a Part B (outpatient) benefit, depending on the policies of Medicare contractors charged with administering claims. Partial hospitalization coverage and reimbursement policies have been the source of considerable controversy over the past decade. Providers com- plained that the Centers for Medicare & Medicaid Services (CMS) was slow to issue a national review standard when the benefit was first established, allowing Medicare contractors to implement variable review procedures that limited access in some areas and inadequately regulated services in others. After payments to freestanding partial hospitalization programs rose sharply between 1993 and 1997, increasing nearly 500 percent (from $60 million to $349 million), a series of investigative reports revealed that several providers billing for this care delivered care that failed to include the mandated set of partial hospitalization services and was not sufficiently intensive and therapeutic in nature. 20 Much of this substandard care was delivered to patients who did not meet the medical necessity criteria for partial hospitalization services. Once apprised of fraud and abuse prob- lems, CMS intensified scrutiny and decertified many providers across the nation. Also, prospective payment for partial hospitalization services, implemented in 2000, established a fixed payment amount for the bundle National Health Policy Forum | www.nhpf.org 12 Background Paper November 27, 2006 of services offered through the program21 and minimized opportunities and incentives for fraud. Outpatient Services Outpatient mental health can include a diverse mix of clinical services, including diagnostic testing, psychotherapy, targeted case management, medication management, and psychosocial services. These services are provided by primary care physicians, nurse practitioners, and physician assis- Under Medicare, outpatient psychotherapy tants, as well as by specialty mental health services are subject to a 50 percent coinsurance. providers, such as psychiatrists, clinical psychologists, psychiatric nurse specialists, and social workers. The spe- cific scope of practice for these professional disciplines is dictated by state licensure laws and can vary substantially from state to state. Medicare Part B pays for professional services delivered in hospital and outpatient settings and, since 1989, has allowed billing from a broad range of nonphysician mental health service providers subject to state licensing parameters.22 However, most mental health services covered by Medicare are provided by primary care physicians. More than half (51.6 percent) of patients treated for major depression are seen in the general medical sec- tor and are cared for exclusively by primary care or other nonpsychiatrist physicians.23 It is also estimated that 67 percent of psychopharmacological drugs are prescribed by primary care physicians.24 Coverage limitations — Medicare has an outpatient mental health limi- tation that imposes a 50 percent coinsurance on psychotherapy services (including individual, family, and group psychotherapy).25 Other mental health services (such as inpatient therapy services, diagnostic testing, psychological evaluation, and brief visits to monitor the efficacy of pre- scribed medications) are subject to the standard 20 percent coinsurance.26 Psychosocial services are generally not covered by Medicare. The higher coinsurance level for the most common forms of outpatient mental health treatment was originally established to deter inappropriate utilization. Policymakers believed that these services were particularly vulnerable to overutilization for a variety of reasons, including the nonin- vasive nature of such services, the lack of objective biomarkers to validate diagnoses of mental disorders, and biases regarding the legitimacy of mental disorders and mental health treatment.27 Differential cost sharing for mental health services was once common practice in commercial insurance prod- ucts, but many private health insurance plans have moved away from this approach as a result of the adoption of more effective, targeted utilization management techniques; the enactment of state laws requiring parity for mental health coverage; and a growing recognition that untreated mental illness can fuel overall health care spending. The Office of the Inspector General (OIG) within the Department of Health and Human Services (DHHS) recently found that Medicare carriers have National Health Policy Forum | www.nhpf.org 13 Background Paper November 27, 2006 adopted inconsistent policies regarding the application of the limitation, which has led to wide geographic variation in beneficiary copayment obligations.28 Among the 57 carriers studied, nine different policies for the application of the limitation were identified. In over one-half of the service areas, carriers subjected evaluation and management services to patients with Alzheimer’s disease to the limi- tation, contrary to federal regulations. MedPAC has recommended that the outpatient Although CMS has issued guidance mental health limitation be eliminated. related to the limitation, the OIG found that this guidance was not sufficiently detailed to ensure uniform imple- mentation across carriers and recommended that new guidance be issued. Patient advocates have expressed concern that more onerous cost-sharing obligations for outpatient mental health benefits under Medicare discourage beneficiaries from seeking necessary services and may significantly limit access to treatment for beneficiaries who lack the financial resources or supplemental insurance to fulfill the 50 percent coinsurance requirement.29 The restriction applies only to psychotherapy services but, especially given inconsistent policies among carriers, concerns have been raised that both patients and providers may believe the higher copayment requirement applies to all outpatient mental health services, discouraging use of diag- nostic and other services not subject to the higher copayment. Access barriers are perhaps most significant for those dually eligible for Medicare and Medicaid.30 While Medicaid is responsible for covering the Medicare coinsurance of dual eligibles, state Medicaid programs can elect to withhold such payment if the amount paid by Medicare meets or exceeds the Medicaid payment rate for that service. Because Medicaid re- imbursement rates for professional services are generally very low relative to Medicare, mental health service providers may be unable to recoup the coinsurance payment from either Medicaid or low-income dually eligible beneficiaries. Although this dynamic also holds true for outpatient treat- ment of physical health conditions, the 50 percent effective “discount” on mental health providers’ service fees is far more substantial than the 20 percent “discount” imposed on physical health services. The Medicare Payment Advisory Committee (MedPAC) has recommend- ed that the outpatient mental health limitation be eliminated,31 finding that the modest increase in program costs likely to result from this action ($500 million in 2002) is justified in light of the access improvements and cost-sharing simplifications that would be achieved. Skeptics ques- tion whether a more generous outpatient mental health benefit would significantly improve appropriate access and worry that inappropriate utilization could increase. Ambiguous billing requirements — The willingness of mental health providers to serve Medicare beneficiaries may be further compromised by confusion regarding reimbursement policies and practices that result in protracted adjudication processes and a high rate of claims denials. Up National Health Policy Forum | www.nhpf.org 14 Background Paper November 27, 2006 to 20 percent of claims for medication management services and up to 50 percent of claims for group therapy services are denied. Denials are made for a variety of reasons, and some proportion of these denials are likely valid. The DHHS OIG concluded that one-third of outpatient mental health services provided to Medicare beneficiaries were medically unnecessary, billed incorrectly, rendered by unqualified providers, or improperly docu- mented. Group therapy services and psychological testing were found to be particularly problematic.32 Many providers believe that local policies Up to 20 percent of claims for medication set by Medicare carriers for outpatient mental health services are too vague, too management services and up to 50 percent of narrowly defined, or too often misapplied, claims for group therapy services are denied. resulting in inappropriate denial of claims or costly resubmissions in order to supply necessary documentation. Provid- ers complain that mental health services are subject to variable and poorly articulated payment rules. Medicare limits coverage to services that are medically “reasonable and necessary for the diagnosis or treatment of illness or injury or to improve the functioning of a malformed body member.”33 Relatively few criteria for making specific coverage decisions are set at the national level. Rather, Medicare carriers are given a fair degree of latitude in establishing rules to assess medical necessity and set service-specific requirement to ensure the appropriateness of services. Inconsistency in coverage policies across regions is not unique to mental health services, but mental health providers believe that the diversity of practices related to specialty mental health services adds to the confusion. Providers also believe that carriers implement established policies improp- erly. For example, some carriers have policies stating that psychotherapy services for patients with cognitive impairments will not be covered if the severity of those cognitive limitations precludes the patient from deriving meaningful benefit from the therapy. Approximately 30 percent to 40 per- cent of people with dementia also have co-occurring mental disorders such as depression or psychosis34 that may be helped through psychotherapy, but these comorbid disorders often go undiagnosed or undertreated.35 Providers report that Medicare carriers routinely deny claims for mental health services delivered to any individual with dementia, regardless of severity level, despite evidence that patients with mild to moderate de- mentia benefit from psychological interventions that improve functioning and coping. Documentation requirements to validate severity level are not clearly delineated. Screening services — Mental health screenings to identify conditions that may not be readily apparent to providers are not covered for most exist- ing Medicare beneficiaries, but a new “Welcome to Medicare,” one-time physical examination that includes depression screening is covered for any new beneficiary joining the program after January 1, 2005. In general, Medicare does not cover preventive health screenings, except for certain, statutorily defined services. This initial preventive physical examination National Health Policy Forum | www.nhpf.org 15 Background Paper November 27, 2006 was added as a benefit by the Medicare Prescription Drug, Improvement and Modernization Act of 2003 (MMA) and includes a broad range of preventive services. Although depression screening is explicitly identified as a component of the physical, Medicare does not require or recommend any specific screening tool. Primary care physicians believe that many of the standardized, validated depression screening tools are too cumbersome to administer during a short office visit. The American Academy of Family Physicians has advised an approach recommended by the U.S. Preventive Services Task Force.36 This approach, the PHQ-2, involves asking two questions: “Over the past two weeks, have you felt down, depressed or hopeless?” and “Over the past two weeks, have you felt little interest or pleasure in doing things?” Answers can either be yes/no or scaled from zero (not at all) to three (nearly every day). An affirmative answer to either question should prompt seven additional questions that provide more detailed evaluation for depression. The extent to which the “Welcome” physi- cal in general and the depression screen in particular will be used is as yet unclear. Patient advocates have raised concerns that even if these screenings are helpful for new beneficiaries, coverage is not available for either screening beneficiaries enrolled before January 2005 or conducting periodic reassessments. psychiatric practice incentives — Psychiatrists have increasingly discon- tinued offering psychotherapy as a component of their practice, prefer- ring to focus on more lucrative medication management services. The financial incentives established by Medicare and other payers encourage psychiatrists to provide psychopharmacologic treatments in contrast to psychotherapy. Reimbursement rates show a clear economic advantage to providing medication with brief follow-up visits and a clear financial disincentive to provide psychotherapy37. Psychiatrists earn about $100 less an hour for providing one 45- to 50-minute session of psychotherapy than for providing three medication management visits in the same time. Some private managed care organizations expressly prohibit psychiatrists from providing psychotherapy. Inadequate psychotherapy training in some psychiatric residency programs appears to reinforce the trend away from psychotherapeutic approaches. These forces have made the delegation of psychotherapy services—if they are offered at all—to nonphysician providers common practice, with little coordination between therapists and prescribing physicians. Many worry that this practice of “splitting” treatment has a negative impact on the quality of patient care and patient outcomes. Opinions dif- fer as to the best way to resolve this problem. Some would advocate for psychiatrists to increase their involvement in providing psychotherapy. Others believe that the delegation of psychotherapy to less highly com- pensated, yet qualified, professionals is economically efficient and desir- able but worry that these providers are not properly trained to monitor prescription drug use. Still others advocate giving limited prescription National Health Policy Forum | www.nhpf.org 16 Background Paper November 27, 2006 privileges to psychologists. Some would argue that more should be done to build a team approach to service delivery. Reimbursement of nonphysician providers — The central role of non- physician providers in delivering psychotherapy suggests that these are the providers most likely to be influenced by Medicare’s outpatient mental health limitation. For many of them, the financial disincentives of the higher copayment level are compounded by other reimbursement policies that decrease their payment levels relative to those of physicians. Nonphysician providers recognized by Medicare receive different rates for their services. For example, psychologists are paid 100 percent of the physician fee sched- ule, whereas social workers receive 75 percent of the physician rate. In addition to psychologists and social workers, Medicare also makes direct Although pharmacological interventions are most payment to nurse practitioners with effective when delivered in combination with psy- the equivalent of a master’s degree in psychotherapy and to clinical chotherapy, nearly half of all patients taking psy- psychiatric nurse specialists. Other chotropic drugs receive no other form of treatment. types of mental health service pro- viders, such as licensed professional clinical counselors and marriage and family therapists, can only be paid indirectly by Medicare (for example, through a physician who employs, supervises, and pays them) and only in limited circumstances. It has been argued that expanding the number of professional disciplines approved for direct payment from Medicare and raising the payment rates for professions paid below the physician fee schedule could improve access to mental health services, particularly in rural areas with health provider shortages. However, MedPAC recently recommended against such expansions, finding that, although utilization would probably rise, the increase in program costs likely to result from this action would not represent a prudent use of resources. Prescription Drugs Psychotropic drugs, broadly defined, include any chemical substance that alters brain function. Psychotropic medications commonly used to treat mental disorders include antidepressants, antipsychotics, antianxiety agents, and stimulants. Increasingly a central component in the treatment of mental illness, such drugs were used in 75 percent of mental health treatment cases in 2001.38 Pharmacological interventions are most effective when delivered in combination with psychotherapy.39 However, for nearly half of all patients taking psychotropic drugs, medication is the only form of treatment received.40 Psychotropic medications can be particularly complicated to monitor, and establishing effective treatment regimens for vulnerable patients with co- occurring neurological disorders is complex. Individuals with severe men- tal impairments may resist changing medications because treatment often National Health Policy Forum | www.nhpf.org 17 Background Paper November 27, 2006 requires considerable trial and figure 4 error before an effective medica- Average Annual Drug Spending Among Medicare Beneficiaries tion or medication combination can be identified. Disruptions in DISABLED Drug Spending g p AGED medications can be dangerous Not in Medicaid Dually Enrolled Not in Medicaid Dually Enrolled for these individuals, resulting in rapid deterioration, impaired Mental Disorder functioning, increased health care 3,000 utilization, and overuse of more Mental Disorder expensive urgent care services 2,500 and inpatient hospitalizations. 2,000 Medicare Part D began cover- ing outpatient prescription psy- 1,500 chotropic and other drugs in January 2006. Coverage is pro- 1,000 vided through either private, stand-alone prescription drug 500 plans or drug plans offered by 0 Medicare Advantage managed care plans (both types referred No Mental Disorder No Mental Disorder to here as “drug plans.”) In Psychotropic Drugs general, individual plans have Non-Psychotropic Drugs a great deal of latitude in de- termining which drugs are Source: Adapted from exhibit 5 in Julie Donohue, “Mental Health In The Medicare Part D Drug Benefit: included on a plan’s formulary A New Regulatory Model?” Health Affairs, 25, no. 3 (May/June 2006), p. 711. and cost-sharing requirements, but the formulary and benefit design must be approved by CMS. Benefi- ciaries can select from the drug plans available in their geographic region. Dually eligible enrollees were automatically enrolled in drug plans in an attempt to avoid gaps in coverage as these beneficiaries transitioned from Medicaid to Part D coverage. Protections — CMS recognized that private health plans would have some incentive to discourage enrollment by high-risk, costly populations, includ- ing those with mental illness, and implemented protections to minimize risk selection. Beneficiaries with a mental disorder have high per capita drug expenditures,41 spending, on average, 61 percent more on drugs than ben- eficiaries without mental illness.42 (See Figure 4 for average drug spending among Medicare beneficiaries.) In order to minimize the development of benefit designs and enrollment strategies biased against such high-cost patients, CMS required drug plans to conform to a variety of safeguards. nProtected drug classes. Plans are required to cover “all or substantially all” drugs available in six special classes: antidepressants, antipsy- chotics, anticonvulsants, antineoplastics, antiretrovirals, and immune suppressants.43 These drug classes are essential to vulnerable groups known to have high prescription drug costs, such as people with mental illness, HIV/AIDS, and cancer. National Health Policy Forum | www.nhpf.org 18 Background Paper November 27, 2006 nUtilization management. Drug plans are precluded from imposing utilization management (such as prior authorization, quantity limits, and step therapy) on individuals who were stabilized on drugs in protected drug classes before being enrolled in the drug plan.44 Plans must also continue coverage for protected medications during the annual benefit period, even if the medication is removed from the formulary. However, plans can apply utilization management tools on any prescription initiated after enrollment in the plan (such as new or different medications or dosage changes). It is not clear how long these protections will remain in place, and plans are not required to provide all doses and forms of a medication (for example, continuous release formulations), which can be particularly problematic for psychotropics because many are prescribed “off-label” at dosage levels that have not been evaluated by the Food and Drug Administration.45 nCost sharing. Plans have flexibility in determining placement of drugs on cost-sharing tiers within the formulary. However, plans are explicitly prohibited from using their formularies to discriminate against any class of beneficiaries, and CMS has reviewed drug plans’ cost-sharing requirements to ensure that protected drug classes are not systemati- cally assigned to the highest cost-sharing tiers. “Benzos” exclusion — Although Part D mandates coverage for many of the psychotropic drugs used to treat mental illness, some types of psychotropic drugs (such as antidementia medications) are not in a protected class, and benzodiazepines are statutorily excluded from coverage. Benzodiazepines, which are tranquilizing agents, table 1 have been the subject of contro- Percentage of Medicare Beneficiaries Using Psychotropic Drugs versy for several decades. These by Therapeutic Category and Eligibility, 2002 medications can lead to depen- dency, and drug-related adverse events (such as oversedation, Disabled Disabled Aged aged falls, and hip fractures) may be Dually Eligible Other Dually Eligible Other more common in older patients Any Psychotropic Drug Use 53.1% 47.8% 31.0% 22.6% taking this drug. Nonetheless, benzodiazepines continue to Antidepressants 39.8% 38.1% 20.6% 15.9% be widely used today, and a Antipsychotics 22.6 8.6 5.0 1.4 growing clinical consensus has Use by developed around their appro- Therapeutic Antimanics 16.3 12.9 4.5 2.3 priate use. (See Table 1 for an Category* Dementia Drugs 0.1 0.4 3.4 2.2 overview of psychotropic drug Benzodiazepines 10.7 11.7 9.0 6.6 use by therapeutic category and Barbiturates 0.5 0.4 0.2 0.1 eligibility class.) Abrupt termi- nation of benzodiazepines can *Percentages of use in therapeutic categories will exceed drug use total because some beneficiaries use lead to severe withdrawal reac- multiple categories of drug. tions.46 In large part because of Source: Adapted from exhibit 2 in Julie Donohue, “Mental Health In The Medicare Part D Drug Benefit: A New Regulatory Model?” Health Affairs, 25, no. 3 (May/June 2006), p. 710. these clinical concerns, all state National Health Policy Forum | www.nhpf.org 19 Background Paper November 27, 2006 Medicaid programs have historically provided some level of coverage for benzodiazepines, despite the fact that the Medicaid statute allows states to exclude coverage of these drugs at their discretion. Transitions from Medicaid to Medicare — Advocates have raised con- cerns about how coverage changes could affect the more than 6 million beneficiaries who are dually eligible for Medicaid and Medicare, 39 percent of whom have mental disorders. Access to psychotropic drugs is likely to improve for many Medicare beneficiaries who did not have prescription drug coverage before the implementation of Part D but, for dually eligible people, the transition could be problematic. Enrollment in Part D was mandatory for the dual eligibles who had received drug coverage through state Medicaid programs before 2006. Duals in some states may encounter new formulary rules and higher cost-sharing requirements. Some Medicaid programs exempt psychotropics from utilization management restrictions, and pharmacists are legally required to fill prescriptions, regardless of the individual’s ability to pay the copayment amount. Sources of drug spend- ing for Medicare beneficiaries are shown in Figure 5. Under Medicare Part D, dual eligibles will continue to be exempt from paying premiums or deductibles (as they were under Medicaid) if they are enrolled in below-average premium plans. However, they are responsible for low copayments ($1 for generic drugs and $3 for brand name drugs, depending upon their income) and pharmacists are now able to require co- payments before filling prescriptions. Studies have shown that even small copays can significantly lower utilization figure 5 of prescription drugs among low-income populations.47 Sources of Prescription Drug Spending These copayments, though seemingly small, can be a sig- for Medicare Beneficiaries, 2002 nificant burden for duals with multiple chronic conditions who struggle to cover costs of several medications with Psychotropic All Other SSDI or Social Security subsidies that are often inadequate Drugs Drugs to cover housing and food costs. 48 27.4% 15.0% Risk adjustment — Federal payments to drug plans are Medicaid intended to minimize the disincentives associated with 34.2% enrolling high-cost patients, such as those with mental 22.6% Out of Pocket disorders, by (i) tying payment level to the risk associated with individual enrollees and (ii) limiting the collective risk incurred by plans through the use of risk corridors 17.6% Employer-Sponsored 25.0% and reinsurance payments for larger-than-expected cata- Coverage strophic expenses. A new risk adjustment model designed 10.8% for Part D adjusts payments based on age and specific Uncollected ] 2.5% 21.7% Liability 23.4% comorbidities. The risk adjuster accounts for 23 percent of the variation in drug spending. Although the Part D Other risk adjuster appears to offer an improvement over older risk adjustment models and is considered to offer fairly Source: Adapted from exhibits 3 and 4 in Julie Donohue, “Mental Health generous protection, the extent to which it obviates risk In The Medicare Part D Drug Benefit: A New Regulatory Model?” selection incentives is not yet clear. Patient advocates worry Health Affairs, 25, no. 3 (May/June 2006), p. 712. National Health Policy Forum | www.nhpf.org 20 Background Paper November 27, 2006 that plans may still have financial incentives to use aggressive utiliza- tion management tools and pricing practices to discourage enrollment by persons with mental illness.49 Care Management Efforts to improve the management and integration of the multiple mental health services covered under Medicare are limited. Medicare’s struggle to coordinate care and control costs under a largely fee-for-service program lacking robust utilization management mechanisms has led Congress to explore the utility of private market disease management programs, which focus on coaching patients, rather than attempting to influence provider practices directly. The recent Medicare Health Support program autho- rized by the MMA pays participating disease management vendors and insurance plans a monthly per beneficiary fee for managing a population of chronically ill beneficiaries. To date, no disease management effort has focused on a mental disorder, and few have incorporated mental health screening and care coordination to address the impact of mental comorbidities on physical conditions like dia- betes. Given the prevalence of mental disorders in the Medicare population and the association of mental disorders with higher health care spending, a strengthened emphasis on mental health could prove useful. Medicare managed care plans do not appear to have made great strides beyond the fee-for-service program. Mental health benefits, care man- agement techniques, and outcomes have not been rigorously studied in Medicare managed care plans. Available evidence suggests that Medicare +Choice (now Medicare Advantage) plans provide poorer quality mental health services than employer-sponsored managed care plans.50 In only one area—mental health—do Medicare managed care plans score lower than their employer-sponsored counterparts on quality of care measures. Fewer Medicare beneficiaries receive appropriate follow-up after hospi- talization for a mental illness and fewer receive appropriate management of their antidepressant medications. These findings suggest that Medicare managed care plans have not taken steps to optimize use of appropriate mental health services and have likely modeled their benefits on the ex- isting fee-for-service structure. This is not surprising, as more generous benefits relative to the traditional program could lead to adverse risk selec- tion for the plans, which have no incentive to attract high-need patients with mental disorders. Managing care for beneficiaries with severe mental illnesses is especially challenging because these patients need psychosocial, rehabilitative, and supportive services that are not typically covered through traditional insur- ance programs outside of Medicaid. Such services recognize that the nature of mental illness interferes with treatment compliance and assists people at various stages of the recovery process to reduce distress and prevent relapse or rehospitalization. Studies have shown psychosocial models to be National Health Policy Forum | www.nhpf.org 21 Background Paper November 27, 2006 more effective at identifying and delivering mental treatment, improving activities of daily living, and reducing utilization of inpatient and other, nonpsychiatric health care services.51 The addition of these services, while cost-effective in terms of the outcome improvements achieved, may not be cost-saving. Therefore, it will be difficult for managed care organizations to implement evidence-based practices if the costs of psychosocial support services are not factored into their capitation rates. While dual eligibles have some access to rehabilitative and psychosocial services through Medicaid and local state agencies, care management for these beneficiaries is complicated by the complexity of coordinating cover- age and payment policies across Medicare and Medicaid. A recent MedPAC report identified a number of conflicts, inconsistencies, or unclear policies that have the potential to hinder care delivery for dual eligibles.52 MedPAC recommended that many of these coverage and payment issues could be alleviated if dual eligibles were enrolled in the same plan for both Medicare- and Medicaid-covered services and the plan took steps to integrate benefits. This integration has occurred in demonstration projects, such as A recent MedPAC report identified a number of con- the Program of All-Inclusive Care for the Elderly (PACE) flicts, inconsistencies, or unclear policies that have and the Wisconsin Partnership the potential to hinder care delivery for dual eligibles. Program. However, the cost- effectiveness of these plans has not been well established and they serve only a small fraction of dual eligibles. Medicare Advantage Special Needs Plans (SNPs), authorized by the MMA, offer another potential avenue to better manage care for Medicare benefi- ciaries with mental disorders, but a SNP focused on this population has not yet been developed. SNPs provide an opportunity to integrate acute and long-term care services and increase coordination of Medicare and Medicaid financing and benefit structures. These plans are subject to the same rules and requirements as other Medicare Advantage plans but they are able to provide products focused exclusively on specialized popula- tions, such as the dual eligibles, the institutionalized, and other chronically ill populations. SNPs may limit enrollment to one of the special needs populations, tailoring benefits and provider networks to best meet the needs of these vulnerable groups. Concerns have been raised that SNPs are not required to coordinate with state Medicaid programs and may do little to integrate care for vulnerable populations. Medicare Spending for Mental Health Services Medicare expenditures for mental health services reflect the program’s benefit design. Before the introduction of Part D, Medicare spending on mental health accounted for a relatively small proportion of total mental health expenditures and an even smaller proportion of total Medicare National Health Policy Forum | www.nhpf.org 22 Background Paper November 27, 2006 spending. Medicare spending on all mental health figure 6 services in 2001 was $7.2 billion, representing 3 Mental Health Spending by Payer Class, 2001 percent of all Medicare spending ($242 billion). In contrast, mental health as a proportion of all Total Mental Health Spending health care expenditures accounted for 12 percent $103.7 billion of Medicaid funds and 4 percent of spending by private insurance in 2001.53 Mental health spend- State ing across payers is presented in Figure 6. (excluding Medicaid) Other 26% Medicaid The limited nature of Medicare’s mental health Federal 6% 26% benefits has shielded the program from the Private mental health costs of its beneficiaries. Al- 23% Out of pocket though roughly 20 percent of all persons with 12% 7% mental disorders are covered by Medicare,54 the program contributes only 7 percent of total spending on mental health services. This dis- Total Medicare Spending proportionately low funding rate is attributed $242 billion Medicare Spending on to the absence of a prescription drug benefit Mental Health prior to 2006, limited coverage for outpatient $7.2 billion (3% of total Medicare spending) services, and other benefit design features that may decrease covered spending and limit ser- vice utilization for mental health services. Source: Tami L. Mark et al.,”U.S. Spending For Mental Health And Substance Abuse Treatment, 1991–2001,” Health Affairs Web Exclusive (March 29 2005), Total mental health spending by Medicare ben- p. W5-137. eficiaries55 has historically been heavily skewed toward inpatient services, with 56 percent of the total going to inpatient care and only 30 percent toward outpatient services. This relationship is in contrast to national trends showing a re- versal in inpatient and outpatient spending over the past decade, during which inpatient spending declined from 40 percent to 24 percent while outpatient spending increased from 36 percent to 50 percent of all mental health spending.56 Medicare’s mental health costs, while small relative to those of other pay- ers, have been rising rapidly in recent years. The growth rate for Medicare spending on mental health services increased from 3 percent to 7 percent over the past decade, making it second only to Medicaid’s. This growth is attributed to the rising number of Medicare beneficiaries who qualify for the program as the result of a mental health disability.57 With the advent of Part D, the growth of mental health spending under Medicare will accelerate significantly. Medicaid spending on psychotropic medications for dually eligible beneficiaries totaled more than $6.3 billion in 2002—an amount almost equal to total mental health expenditures by Medicare in 2001. This suggests that Medicare spending for mental health services will at least double as a result of Part D. Spending increases will likely be even higher in light of psychotropic drug expenditures associ- ated with non–dual eligibles and apparent increases in psychotropic drug prices relative to those paid by state Medicaid programs. Because these National Health Policy Forum | www.nhpf.org 23 Background Paper November 27, 2006 drugs were in a protected class, drug plans contend that they had little figure 7 leverage to negotiate price on these products with manufacturers. Plans Psychotropic Drug were mandated to cover a broad range of products in protected classes Expenses as a Percentage and, therefore, had little latitude to offer volume guarantees in exchange of Overall Mental Health for price discounts. Spending, 1987–2001 Although prescription drug costs have not been a significant cost factor for Medicare until recently, it is important to note that psychotropic medications account for almost all growth in mental health spending for other public 21.0% and private insurance plans over the last 15 years. In 1987, 7.7 percent of all mental health care spending in the United States was for psychotropic medications. By 2001, spending on psychotropic drugs accounted for 21 percent of total mental health spending.58 Between 1992 and 1997, total U.S. spending on psychotropic drugs grew at twice the rate of total drug 12.8% spending and, since 1997, spending growth for psychotropic medications has outpaced both total health care spending and total drug spending.59 7.7% In 2003, more than $18 billion was spent on antidepressant and antipsy- 7.2% chotic drugs.60 Figure 7 displays growth in psychotropic drug spending 1987 1992 1997 2001 as percentage of total mental health spending. Source: Richard G. Frank, Rena M. Conti, Conclusion and Howard H. Goldman, “Mental Health Policy and Psychotropic Drugs,” Mil- A dramatic increase in spending for mental health services under Medicare, bank Quarterly, 83, no. 2 (2005), pp. 271–296; available at www.macfound. resulting largely from the introduction of the prescription drug benefit, org/atf/cf/%7bb0386ce3-8b29-4162-8098- has the potential to trigger a much broader assessment of the program’s e466fb856794%7d/mental%20health% mental health policies. As Medicare expenditures for mental health begin 20ftf%203%2006.pdf. to more accurately reflect the burden of disease borne by beneficiaries, policymakers may take a much closer look at the value of the services being purchased. The conclusions they reach could well hinge on an assessment of the effectiveness and efficiency of the services delivered, along with an exploration of the extent to which financing incentives are contributing to these clinical practices: Are the “right” mental health services being delivered to the “right” Medicare beneficiaries at the “right” time at the “right” price? While the same questions could be posed for any service covered by Medicare, mental health services have often been singled out for special consideration, and the influence of these “exceptional” policies is unclear. Medicare and other insurers have often treated mental health services differently from other services because mental health care can differ sig- nificantly from the traditional medical model. Medicare’s mental health policies are often debated in a piecemeal, disjointed fashion, frequently as footnotes to more global policy changes. There may be some benefit to considering Medicare’s mental health policies more holistically in order to better consider how these policies interact and reinforce (or contradict) each other. In monitoring protections developed under Part D to prevent discrimination against persons with mental illness and other costly conditions, policymakers National Health Policy Forum | www.nhpf.org 24 Background Paper November 27, 2006 may find it necessary to consider how policies related to other services are influencing psychotropic drug utilization. Assessing the impact of restric- tions on outpatient mental health services, tracking the influence of prospec- tive payment on inpatient psychiatric capacity, exploring mechanisms to better leverage Special Needs Plans and other care management strategies, and examining eligibility policies for disabled persons are among the issues that may come under further scrutiny as Medicare’s spending on those with mental disorders continues to rise. ENDNOTES 1. Marilyn Moon, “Growth in Medicare Spending: What Will Beneficiaries Pay?” Urban Institute, January 7, 1999; available at www.urban.org/url.cfm?ID=407788. 2. Estimates of mental health disorders in this paper include all diagnosable mental and substance abuse disorders listed in American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association: Arlington, VA, 4th ed., 2000). Statistics that specifically refer to adults with serious mental illness , serious and persistent mental illness, or substance abuse disorders are noted. Estimates of mental disorders in the general population are based on rigorous epidemiological studies that rely on clinically guided surveys using standardized diagnostic criteria to identify mental disorders. Estimates of mental disorders among Medicare beneficiaries are from Kaiser Family Foundation analysis of 2005 Medicare Current Beneficiary Survey (MCBS) Cost and Use File. See Henry J. Kaiser Family Foundation (KFF), Medicare Chartbook ( Menlo Park, CA: Henry J. Kaiser Family Foundation, 3rd ed., Summer 2005); available at www.kff.org/ medicare/upload/Medicare-Chart-Book-3rd-Edition-Summer-2005-Report.pdf. 3. KFF, Medicare Chartbook. 4. Ronald C. Kessler et al., “Prevalence, Severity, and Comorbidity of 12-Month DSM-IV Disorders in the National Comorbidity Survey Replication” Archives of General Psychiatry, 62, no. 6 (June 2005), pp. 617–662. 5. Prevalence rate for Medicare beneficiaries calculated by author based on a weighted average of prevalence rates in elderly and disabled population, with relative weights derived from Centers for Medicare & Medicaid Services (CMS), “Medicare Enrollment: National Trends 1966–2005,” available at www.cms.hhs.gov/MedicareEnRpts/Downloads/HISMI05. pdf. Prevalence rate for general population from Ronald C. Kessler et al., “Prevalence and Treatment of Mental Disorders, 1990 to 2003,” New England Journal of Medicine, 352, no. 24 (June 15, 2005), pp. 2515–2523. 6. Sandra M. Foote and Christopher Hogan, “Disability Profile And Health Care Costs Of Medicare Beneficiaries Under Age Sixty-Five,” Health Affairs, 20, no.6 (November/December, 2001), pp. 242–253; available at http://content.healthaffairs.org/cgi/reprint/20/6/242.pdf. 7. Ronald C. Kessler et al., “Estimating the Prevalence and Correlates of SMI in Com- munity Epidemiological Studies,” chap. 12 in Mental Health, United States, 2002, Ronald W. Manderscheid and Marilyn J. Henderson, Eds., DHHS Publication No. SMA04-3938 (Rockville, MD: U.S. Department of Health and Human Services, 2002); available at http:// mentalhealth.samhsa.gov/publications/allpubs/sma04-3938/Chapter12.asp. 8. Social Security Administration, “Annual Statistical Report on the Social Security Disabil- ity Insurance Program, 2004,” March 2006; available at www.ssa.gov/policy/docs/statcomps/ di_asr/2004/index.html#editions. 9. Authors’ calculations based on data in KFF, “Medicare Chartbook.” Endnotes / continued  National Health Policy Forum | www.nhpf.org 25 Background Paper November 27, 2006 Endnotes / continued 10. Medicare spending per aged beneficiary is $6,367; for each disabled beneficiary under the age of 65 without end-stage renal disease, Medicare spends $5,419. See Medicare Payment Advisory Committee (MedPAC), A Data Book: Healthcare spending and the Medicare program, June 2006, p.19. Once disabled beneficiaries reach age 65, they are classified as aged, regard- less of disability status. 11. Foote and Hogan, “Disability Profile And Health Care Costs,” p. 245. 12. Authors’ calculations based on disease prevalence rates cited in this paper and popula- tion statistics obtained from CMS, “Medicare Enrollment: National Trends. Medicare Aged and Disabled Enrollees by Type of Coverage,” 2005; available at www.cms.hhs.gov/Medica- reEnRpts/Downloads/HISMI05.pdf. See also Craig W. Colton and Ronald W. Manderscheid, “Congruencies in increased mortality rates, years of potential life lost, and causes of death among public mental health clients in eight states,” Preventing Chronic Disease, 3, no. 2 (April 2006); available at www.cdc.gov/pcd/issues/2006/apr/05_0180.htm. 13. Alzheimer’s Disease International, “Factsheet 3: The prevalence of dementia,” April 1999; available at www.alz.co.uk/adi/pdf/3preval.pdf. 14. Medicare provides an additional 60 hospital days (lifetime reserve days) for coverage beyond the 90-day limit; these 60 days can only be used once in the beneficiary’s lifetime. 15. Section 1812(b)(3) of the Social Security Act. 16. Before 2005, Medicare paid for inpatient services in two ways: (i) care in scatter beds in general hospitals were paid under the inpatient prospective payment system (PPS) and (ii) care in general and freestanding IPFs were paid under TEFRA (the Tax Equity and Fiscal Responsibility Act of 1982) rules, with payment based upon the hospital’s base year costs (first full fiscal year of operation) and then updated to the current fiscal year by the annual allowable rate of increase. If costs were below the target cost per case, the providers received a bonus payment of 50 percent of the difference between the target and the actual cost per case, up to 5 percent of the target amount. If costs exceeded the target rate, Medicare provided additional payments equal to half of the amount by which the hospital’s costs exceed the target amount, up to 10 percent of the target amount. Hospitals could also apply for additional Medicare exceptions payments. Many facilities established in the early 1980s consistently exceeded their TEFRA limits, and higher payments under TEFRA rules led to a steady growth in the number of PPS-exempt facilities. 17. For more detailed information about payment methodology, see MedPAC, “Payment Basics: Psychiatric Hospital Services Payment System,” September 2006; available at www.medpac.gov/publications/other_reports/Sept06_MedPAC_briefs_psych.pdf; and CMS, “Final Inpatient Psychiatric Facility (IPF) Prospective Payment System (PPS) Fact Sheet,” December 22, 2004; available at www.cms.hhs.gov/InpatientPsychFacilPPS/Down loads/IPFPPSFactsheet.pdf. 18. Assertive Community Treatment Association, “ACT Home”; available at www.actassociation.org/. 19. CMS, Medicare Benefit Policy Manual, “Hospital Services Covered Under Part B,” chapter 6, 2005; available at www.cms.hhs.gov/manuals/Downloads/bp102c06.pdf. 20. Office of Inspector General (OIG), Review of Partial Hospitalization Services Provided Through Community Mental Health Centers, Report no. OEI-A-04-98-02146). Rockville, MD: Office of Inspector General, U.S. Department of Health and Human Services, 2001; available at http://oig.hhs.gov/oas/reports/region4/49802146.pdf. 21. For billing instructions for partial hospitalization services, see CMS, “Hospital Outpatient Services,” §100, chap. 4 in Medicare Claims Processing Manual, revised September 18, 2006; available at www.cms.hhs.gov/manuals/downloads/clm104c04.pdf. 22. The Office of the Inspector General has raised concerns that state licensure laws may not provide sufficient detail regarding providers’ scope of practice to allow rigorous claims adjudication and quality monitoring. Endnotes / continued  National Health Policy Forum | www.nhpf.org 26 Background Paper November 27, 2006 Endnotes / continued 23. Ronald C. Kessler et al., “The epidemiology of major depressive disorders,” Journal of the American Medical Association, 289, no. 23 (June 18, 2003), pp. 3095–3105. 24. Ronald A. Finch and Kathryn Phillips, An Employer’s Guide to Behavioral Health Services: A Roadmap and Recommendations for Evaluating, Designing, and Implementing Behavioral Health Service. (Washington, DC: National Business Group on Health; 2005); available at www. businessgrouphealth.org/pdfs/fullreport_behavioralhealthservices.pdf. 25. Section 1833(c) of the Social Security Act: The formula for determining actual payment amounts for outpatient therapy can be confusing. Medicare pays 80 percent of the cost for outpatient services. However, the outpatient mental health limitation requires Medicare to reduce its payment for outpatient therapy services by 62.5 percent of the approved amount. Essentially, Medicare pays 62.5 percent of 80 percent of the approved amount. This reduction does not reduce the fee entitled to the provider; thus, the outpatient mental health treatment limitation in effect shifts payment responsibility from Medicare to the beneficiary. The end result is that Medicare pays 50 percent of the approved amount (62.5 percent of 80 percent) and the beneficiary must pay the remaining 50 percent of the original approved amount. 26. In 2002, Medicare authorized payment for six current procedural terminology (CPT) codes for health and behavior assessment and intervention (H&B) services to prevent, treat, or manage behavioral health issues related to physical (rather than mental) health condi- tions. Interventions provided under these codes might include helping patients cope with medical illness and medical procedures, improving management of chronic conditions, or increasing adherence to medication and treatment protocols. Significantly, reimbursements for the H&B services are billed under medical services and are not subject to the 50 percent copay requirement for outpatient mental health services. 27. Outpatient mental health benefits were once subject to both high cost sharing and low spending limits until the late 1980s, when Congress raised and ultimately eliminated the original limits on total payments for outpatient psychiatric services. Before 1987, Medicare covered an annual maximum of only $250 (50 percent of the first $500) of outpatient mental health expenses. The $250 annual payment cap was increased to $1,100 with the enactment of the Omnibus Budget Reconciliation Act (OBRA) of 1987. Two years later, OBRA 1989 eliminated the annual limit and also allowed licensed clinical psychologists and certified social workers to bill Medicare for mental health services. These factors led to a substantial increase in Medicare payments for mental health services in nursing homes during the early and mid-1990s. Many providers do not believe the repeal of the annual limits had a major impact on access, because beneficiaries are still responsible for a 50 percent copayment for outpatient therapy services, which serves as a significant deterrent to utilization. 28. OIG, “Carrier Determination of Copayments for Medicare Mental Health Services,” OEI- 09-04-00221, October 2006; available at http://oig.hhs.gov/oei/reports/oei-09-04-00221.pdf. 29. Judith R. Lave and Howard H. Goldman, “Medicare Financing for Mental Health Care,” Health Affairs, 9, no. 1 (Spring 1990), pp. 19–30. 30. Janet B. Mitchell and Susan G. Haber, “State Payment Limitations on Medicare Cost- Sharing: Impacts on Dually Eligible Beneficiaries and Their Providers,” RTI International, July 31, 2003; available at www.rti.org/pubs/StatePaymentLimits.pdf. 31. MedPAC, Report to the Congress: Assessing Medicare Benefits, June 2002, p.47; available at www.medpac.gov/publications/congressional_reports/Jun02_Entire%20report.pdf. 32. OIG, “Medicare Part B Payments for Outpatient Mental Health Services,” OEI-03-99- 00130, May 2001; available at http://oig.hhs.gov/oei/reports/oei-03-99-00130.pdf. 33. Social Security Act, section 1862(a)(1)(A). 34. American Geriatrics Society (AGS) and American Association for Geriatric Psychiatry (AAGP), “The American Geriatrics Society and American Association for Geriatric Psychia- try Recommendations for Policies in Support of Quality Mental Health Care in U.S. Nursing Homes,” Journal of the American Geriatrics Society, 51, no.9 (September 2003), pp. 1299–1304. Endnotes / continued  National Health Policy Forum | www.nhpf.org 27 Background Paper November 27, 2006 Endnotes / continued 35. AGS and AAGP, “Recommendations for Policies in Support of Quality Mental Health Care”; and M. A. Smyer, D. G. Shea, and A. Streit, “The provision and use of mental health services in nursing homes: Results from the National Medical Expenditure Survey,” American Journal of Public Health, 84, no. 2 (1994), pp. 284–287. 36. Randall O. Card, “Getting Paid: How to Conduct a ‘Welcome to Medicare’ Visit,” Family Practice Management, 12, no. 4 (April 2005), p. 27–34; available at www.aafp.org/ fpm/20050400/27howt.html. 37. Steven Sharfstein, “How Will Psychologists Practicing Medicine Affect Psychiatry?” Psychiatric News, 41, no. 9 (May 5, 2006), p. 3. 38. Samuel H. Zuvekas, “TRENDS: Prescription Drugs And The Changing Patterns Of Treatment For Mental Disorders, 1996–2001,” Health Affairs, 24, no. 1 (January/February 2005), pp. 195–205 39. World Health Organization, The World Health Report 2001: Mental Health—New Under- standing, New Hope (Geneva, Switzerland: World Health Organization, 2001); available at www.who.int/whr/2001/en/whr01_en.pdf. 40. Julie Donohue, “Mental Health In The Medicare Part D Drug Benefit: A New Regulatory Model?” Health Affairs, 25, no. 3 (May/June 2006), pp. 707–719. 41. Donohue, “Mental Health In The Medicare Part D Drug Benefit.” 42. Bruce Stuart, Linda Simoni-Wastila, and Danielle Chauncey, “Assessing the Impact of Coverage Gaps in the Medicare Part D Drug Benefit,” Health Affairs Web Exclusive, April 19, 2005, pp. w167–w179. 43. See Guidance from CMS regarding this requirement; available at www.cms.hhs.gov/ PrescriptionDrugCovContra/10_RxContracting_SpecialGuidance.asp. 44. See CMS guidance on “all or substantially all,” available at http://new.cms.hhs.gov/ PrescriptionDrugCovContra/Downloads/FormularyGuidance AllorSubAll.pdf. 45. Additional information about the impact of Part D for people with mental disorders, see American Association of Community Psychiatrists, American Association for Geriatric Psychiatry, American Psychiatric Association, National Alliance on Mental Illness, National Association of State Mental Health Program Directors, National Council for Community Behavioral Healthcare, National Mental Health Association, and Treatment Effectiveness Now, “Mental Health Part D”; available at www.mentalhealthpartd.org/. 46. Stephen B. Soumerai et al., “Effects of Limiting Medicaid Drug-Reimbursement Ben- efits on the Use of Psychotropic Agents and Acute Mental Health Services by Patients with Schizophrenia,” New England Journal of Medicine, 331, no. 10 (September 8, 1994), pp. 650–655. 47. Bruce Stuart et al., “Who Bears the Burden of Medicaid Drug Co-payment Policies?” Health Affairs, 18, no. 2 (March/April 1999), pp. 201–212; and R. Tamblyn, “Adverse Events Associated with Prescription Drug Cost-Sharing among Poor and Elderly Persons,” Journal of the American Medical Association, 285, no. 4 (January 2001), pp. 421–429. 48. National Alliance on Mental Illness, “Fact Sheet: Five Reasons Cost Sharing Is Bad Public Policy for Individuals with Serious Mental Illnesses,” June 2006; available at www. nami.org/Template.cfm?Section=Medicaid&Template=/ContentManagement/ContentDisplay. cfm&ContentID=36213. 49. Donohue, “Mental Health In The Medicare Part D Drug Benefit.” 50. MedPAC, Report to the Congress: Medicare Payment Policy, March 2004, p. 45; available at www.medpac.gov/publications/congressional_reports/Mar04_Ch2.pdf. Endnotes / continued  National Health Policy Forum | www.nhpf.org 28 Background Paper November 27, 2006 Endnotes / continued 51. For review, see Mark Olfson, Merrile Sing, and Herbert J. Schlesinger, “Mental Health/ Medical Care Cost Offsets: Opportunities For Managed Care,” Health Affairs, 18, no. 2 (March/April 1999), available at http://content.healthaffairs.org/cgi/reprint/18/2/79.pdf; Lewin Group, “Costs of Serving Homeless Individuals in Nine Cities” chart book, November 19, 2004, available at www.rwjf.org/files/newsroom/cshLewinPdf.pdf; Bazelon Center for Mental Health Law, Avoiding Cruel Choices, (Washington, DC: Bazelon Center for Mental Health Law, 2002), available at www.bazelon.org/issues/children/publications/TEFRA/ avoidingcruelchoices.pdf. 52. MedPAC, “Dual Eligible Beneficiaries,” chap. 3 in Report to the Congress: New Approaches in Medicare, June 2004; available at www.medpac.gov/publications/congressional_reports/ June04_Entire_Report.pdf. 53. Tami L. Mark et al., “U.S. Spending For Mental Health And Substance Abuse Treatment, 1991–2001,” Health Affairs Web Exclusive (March 29 2005), W5-133. 54. Authors’ calculations based on disease prevalence rates cited in this paper and population statistics obtained from CMS, “Medicare Enrollment”; and U.S. Census Bureau, “2005 American Community Survey”; available at http://factfinder.census.gov/servlet/ ADPTable?_bm=y&-geo_id=01000US&-qr_name=ACS_2005_EST_G00_DP1&-ds_name= ACS_2005_EST_G00_&-_lang=en&-_caller=geoselect&-state=adp&-format= . 55. Inclusive of out-of-pocket payments and supplemental insurance. 56. Does not include spending on residential treatment (primarily services related to child abuse and substance abuse) or spending on insurance administration. See Mark, “U.S. Spending for Mental Health and Substance Abuse Treatment.” 57. Social Security Administration, “Annual Statistical Report on the Social Security Disabil- ity Insurance Program, 2004,” March 2006; available at www.ssa.gov/policy/docs/statcomps/ di_asr/2004/index.html#editions. 58. Richard G. Frank, Rena M. Conti, and Howard H. Goldman, “Mental Health Policy and Psychotropic Drugs,” Milbank Quarterly, 83, no. 2 (2005), pp. 272; available at www. macfound.org/atf/cf/%7bb0386ce3-8b29-4162-8098-e466fb856794%7d/mental%20health%20 ftf%203%2006.pdf. 59. Tami L. Mark et al., National Estimates of Expenditures for Mental Health Services and Substance Abuse Treatment, 1991–2001 (Rockville, MD: U.S. Department of Health and Hu- man Services, 2005); available at www.samhsa.gov/spendingestimates/SEPGenRpt013105 v2BLX.pdf. 60. IMS, “IMS Reports 11.5% Dollar Growth in ’03 US Prescription Sales,” press release, February 17, 2004; available at www.imshealth.com/ims/portal/front/articleC/0,2777, 6599_3665_44771558,00.html. The National Health Policy Forum is a nonpartisan research and public policy organization at The George Washington University. All of its pub- lications since 1998 are available online at www.nhpf.org. National Health Policy Forum | www.nhpf.org 29