|
Mental Health In The Medicare Part D Drug Benefit: A New Regulatory Model?
The Medicare Prescription Drug, Improvement, and Modernization Act (MMA) of 2003 represents the most significant expansion of Medicare benefits since the program’s inception and has important implications for mental health. Medicare will become a major payer for psychotropic medications through the new prescription drug benefit. The structure of the drug benefit’s delivery system creates incentives for plans to underprovide medications, like psychotropic drugs, that are used persistently and are associated with high expected costs. Regulators have put policies in place to counteract these incentives. This paper examines these strategies’ likely success and suggests additional approaches to be considered.
THE MEDICARE PRESCRIPTION DRUG, Improvement, and Modernization Act (MMA) of 2003 represents the most dramatic expansion of Medicare benefits since the program’s inception forty years ago. MMA authorizes the creation of a new prescription drug benefit (Medicare Part D) that expands drug coverage to the 18 percent of beneficiaries who lack drug coverage year round and to an additional 27 percent who lack drug coverage for part of the year.1 Medicare Part D also expands Medicare’s role in financing mental health services. Together with Medicaid, Medicare will finance a sizable share of psychotropic medication use in the United States at a time when pharmacotherapy has become the dominant form of mental health treatment. Historically, Medicare’s coverage of mental health services, like that of third-party payers, was more limited than coverage for medical services for other conditions. With the implementation of MMA, a new model of Medicare mental health policy is emerging, one that integrates mental health and sets up special provisions to protect people with mental disorders. A key challenge to integrating mental health services and general medical care in insurance markets is the risk of adverse selection and the incentives this risk creates for insurers to stint on mental health treatment. Medicare is attempting to counteract these incentives by putting into place special provisions to ensure access to pharmacotherapy by people with mental illnesses. In this paper I provide some background on the role of psychotropic drugs in modern mental health care, followed by a brief history of mental health in Medicare and discussion of the prevalence of mental illnesses and use of psychotropic medications among beneficiaries. I discuss some of the resulting concerns about ensuring access to psychotropic medications for the aged and disabled. I argue that relative to beneficiaries with other health needs, incentives may be heightened for Medicare drug plans to discourage the enrollment of beneficiaries with mental health needs or to place limits on their care. I then discuss strategies for combating these incentives, examine these strategies’ likely success, and suggest additional approaches that could protect the vulnerable populations affected.
Psychotropic drugs have become the predominant treatment for mental illnesses. In 2001, 75 percent of people who were treated for mental health or substance abuse disorders received pharmacotherapy either alone (34 percent of those treated) or in combination with treatment in ambulatory care settings (41 percent).2 Several new therapeutic agents were introduced in the 1990s to treat depressive and anxiety disorders, attention deficit hyperactivity disorder (ADHD), psychotic disorders, and other mental health conditions. The availability of these new medications, along with a host of other factors, led to an increase in the number of people receiving psychotropic drug treatment as well as the substitution of newer, more expensive agents for older drugs. The amount spent on psychotropic drugs grew from $2.7 billion in 1987 to $17.8 billion in 2001.3 Since 1997, spending on psychotropic medications has grown at a higher rate than spending on health care and prescription drugs overall.4
Mental health has had an uneven history in Medicare. Some mental health services are covered at parity with other medical services, while others are more limited. Inpatient psychiatric hospitalization in general hospitals has been covered on the same terms as all medical care under Medicare Part A. Yet Medicare placed a 190-day lifetime limit on inpatient services in public or private psychiatric hospitals. Outpatient mental health benefits were subject to both high cost sharing and low spending limits until Medicare reforms were made in the late 1980s.5 Medicare continues to require 50 percent coinsurance for psychotherapy and other outpatient mental health services that do not involve medication management, in contrast to most private insurance plans, which impose the same cost sharing for outpatient mental health services as for other medical services.6 High cost sharing for outpatient mental health benefits in Medicare likely limited access to treatment for some beneficiaries with mental illnesses.7 In part because benefits are more limited, Medicare pays for only 7 percent of mental health and substance abuse spending, compared with 17 percent of overall health care spending.8 Only 3 percent of total Medicare spending is on mental health and substance abuse services, compared with 4 percent of private insurance and 12 percent of Medicaid spending, in spite of the high prevalence of mental disorders among Medicare beneficiaries. Approximately 15 percent (6.3 million) of Medicare beneficiaries are under age sixty-five with permanent disabilities.9 The Social Security Administration estimates that roughly one-third of Social Security Disability Insurance (SSDI) recipients are disabled as a result of a mental disorder.10 Aged and disabled beneficiaries are very different populations: They differ in terms of the prevalence of mental illnesses as well as the level of prescription drug use, spending on prescription drugs, and sources of drug coverage. Here I present population-based estimates of the diagnoses, prescription drug use, spending, and sources of payment for psychiatric drugs based on analyses of the 2002 Medicare Current Beneficiary Survey (MCBS) and associated Medicare claims.11 I provide data for aged or disabled beneficiaries and data by Medicaid eligibility because of the important role that Medicaid plays in financing psychotropic drugs and the changes that Part D will bring about for dual eligibles.12
Prevalence of mental illnesses.
The diagnosed prevalence of mental illnesses among Medicare beneficiaries varies by eligibility category from 2.5 percent to 17.1 percent and is highest among the dually eligible disabled population (Exhibit 1
Psychotropic drug use and spending. Psychotropic drug use is high among all Medicare beneficiaries. Approximately one-quarter of the aged and one-half of disabled beneficiaries used one or more psychotropic drugs in 2002 (Exhibit 2  ). Antidepressants are the most commonly used therapeutic category among all eligibility categories, followed by antipsychotics for the disabled and benzodiazepines for aged beneficiaries. The high use of benzodiazepines among Medicare beneficiaries is notable, given that the standard Medicare drug benefit will not cover that therapeutic category. Abrupt termination of benzodiazepine use because of non-coverage under Part D might lead to withdrawal reactions, seizures, and emergency department (ED) visits.14 As a result, most state Medicaid programs have elected to cover benzodiazepines, which treat anxiety disorders, insomnia, and other neurological and rheumatologic disorders, for dually eligible enrollees.
Among beneficiaries with a mental disorder, per capita psychotropic drug spending tends to be highest in the dually eligible disabled population, at a mean of $1,408 in 2002 (Exhibit 3 ).
Total drug spending. Furthermore, beneficiaries with mental illnesses are likely to have high overall prescription drug spending both because of the high cost of psychotropic medications per se and because these beneficiaries frequently suffer from other chronic medical conditions that require pharmacotherapy. Medicare beneficiaries with depression have high rates of medical service use for other conditions such as hypertension, congestive heart failure, diabetes, and osteoarthritis.15 A majority of people with schizophrenia have at least one serious medical condition that requires treatment with medication.16 Regardless of eligibility category, per capita drug spending for beneficiaries with a mental disorder is higher than that for beneficiaries who do not have a mental disorder (Exhibit 4 ).
Source of drug coverage. Before implementation of Part D, the aged and disabled had somewhat different sources of drug coverage. In 2002, Medicaid was the most likely source of drug coverage for disabled beneficiaries (33 percent), whereas employer-sponsored coverage was the most likely source for the aged (35 percent).17 Dually eligible disabled beneficiaries constitute only 8 percent of Medicaid enrollees yet account for 25 percent of Medicaid drug spending.18 In 2002, Medicaid paid for 27.4 percent of all psychotropic medications for Medicare beneficiaries, compared with 15 percent of all other prescription drugs (Exhibit 5 ).
Under Part D, dually eligible beneficiaries will move from Medicaid drug coverage to Medicare prescription drug plans (PDPs). Because the dually eligible and other low-income Medicare beneficiaries will receive generous subsidies under Part D (low or no premiums or deductibles and copayments of $1–$5), they are unlikely to incur much higher out-of-pocket drug costs under the new system. However, the dually eligible could face more restrictions on access to psychotropic medications, because of the use of prior authorization, step therapy, or quantity limits, than under Medicaid. Thus far, few states have excluded psychotropic drugs from their preferred drug lists.19 State Medicaid programs vary with respect to how they apply management tools to psychotropic drug classes, so the effect of the transition to Medicare Part D will vary by state of residence.
Medicare drug plans. Beneficiaries enrolling in Part D may either enroll in a Medicare Advantage prescription drug (MA-PD) plan to obtain coverage for all Medicare-covered benefits, including prescription drugs, or stay in traditional Medicare and enroll in a stand-alone PDP. In 2006, depending on the region, beneficiaries will face a choice of thirteen to twenty-two plans, some offering multiple insurance products. Some of these are national plans; others are available only in one of the thirty-four drug plan regions. These plans vary in terms of premiums, benefit design (deductibles, copayments), and use of formularies and other pharmacy benefit management tools. Some plans will adopt the standard drug benefit package as outlined in the MMA statute, but most will alter some dimension of the benefit.20 For example, in Pennsylvania beneficiaries can choose from twenty-three PDPs offering fifty-two products. Monthly premiums vary among these PDPs from $10.14 to $68.61.21 Two-thirds of the PDP products have either a zero or reduced deductible; the remainder have the standard $250 deductible. The number of drugs listed on PDP formularies varies from plan to plan. Adverse selection in mental health and drug insurance markets. An important lesson of mental health economics is that the problems with adverse selection that lead to health insurance market failure are exacerbated for mental health.22 People with mental illnesses have high expected health care spending, and mental health spending is highly predictive of health plan choice.23 Prescription drug usage exhibits some of the same features. Information asymmetry between consumers and their health plans is greater for prescription drugs because spending on prescription drugs is more persistent and therefore more predictable to consumers than other forms of health spending.24 As a result, consumers are likely to self-select into health plans on the basis of prescription drug use and spending.25 Both the lack of a private market for stand-alone prescription drug coverage and the minimal drug coverage available in Medigap policies have been attributed to adverse selection.26
To avoid adverse selection, plans have an incentive to underprovide services, such as mental health services, that are predictable to the consumer and correlated with high total spending, because these services are likely to affect consumers’ choice of plans.27 Because PDPs are competing for enrollees and will face some financial risk under Part D, they will have an incentive to structure their formularies, cost sharing, pharmacy networks, and pharmacy management tools so that they do not enroll a disproportionately high number of people with high costs. Because psychotropic drug use is chronic and associated with high expected total drug costs (Exhibit 4 Strategies for attenuating selection incentives. MMA legislation and the Centers for Medicare and Medicaid Services (CMS) put in place a number of strategies for reducing the risk of adverse selection and attenuating plan incentives to stint on provision of certain drugs. Risk adjustment. Adverse selection is a problem only when prices paid to plans do not adequately reflect heterogeneity in expected costs. PDPs are new entities created in response to MMA. Unlike pharmacy benefit management (PBM) organizations, PDPs will face some financial risk for delivering the drug benefit. Monthly payments will be risk-adjusted to account for differences in enrollees’ drug spending. The risk adjustment model used for Medicare Part D is a modified version of the CMS-HCC (Hierarchical Condition Category) risk adjustment model developed for MA plans; it adjusts rates for age, sex, disability status, and the presence of multiple health conditions. The proposed risk adjustment model for Part D explains roughly 23 percent of the variation in drug spending—an improvement over the 12 percent of variation in spending predicted by overall health care risk adjustment models.28 However, the potential for risk adjustment models to reduce selection should be evaluated based on a comparison of actual predictive power with the amount of variation that could be explained.29 The amount of explainable variation in drug spending is much higher than that for outpatient or inpatient medical spending. Marian Wrobel and colleagues explained 55 percent of the variation in Medicare drug spending by adding the prior year’s drug spending to a risk adjustment model similar to that proposed by the CMS for Part D.30 Given that the proposed risk adjustment system explains less than half of the explainable variation (23 percent versus 55 percent), it will not fully attenuate selection incentives. Risk sharing. In addition to the risk-adjusted payment system, Medicare will share risk with drug plans through reinsurance and risk corridors. Medicare will reinsure PDPs for 80 percent of drug costs after a beneficiary’s drug costs have exceeded the out-of-pocket maximum in the standard benefit ($5,100). Through risk corridors, Medicare sets a drug spending target equal to total premiums minus administrative costs. Medicare then specifies corridors of risk around that target and holds plans responsible for a portion of the costs that fall within each corridor. If costs fall below the target amount, Medicare will share savings with the plan. In 2006 and 2007, plans will pay all of the drug costs between the target amount and 2.5 percent above the target. Plans will be responsible for 20 percent of the costs between 2.5 percent and 5 percent above the target amount, and for 25 percent of the costs above 5 percent of the target. Reinsurance and risk corridors do not target losses that are predictable to the plan, and they still leave plans incurring some losses on high-cost cases. Therefore, these forms of risk sharing reduce plans’ financial risk without reducing incentives to encourage enrollment of good risks or discourage the enrollment of bad risks.31 Tightening up the formulary rules. MMA requires that plans offer drugs in every therapeutic category used by beneficiaries and that at least two drugs be offered per category or pharmacologic class. Applying this rule alone would have given plans leeway to exclude the most commonly used and most costly psychotropic medications from their formularies, to discourage beneficiaries with mental illnesses from enrolling in those plans. For example, plans could have offered two tricyclic antidepressants and no selective serotonin reuptake inhibitors (SSRIs), even though the latter are the most widely used for treatment of depression.32 CMS formulary guidelines, however, imposed additional requirements on plans. They stipulate that for certain therapeutic categories—antidepressants, antipsychotics, anticonvulsants, anticancer drugs, immunosuppressants, and HIV/AIDS drugs—plans are required to list "all or substantially all" of the drugs in the category. Both a clinical and an economic rationale were offered in support of this decision. The CMS stated that including all drugs in these classes is consistent with best practices in the public and private sectors, is consistent with prescribing patterns for these drugs for current beneficiaries, and is attributable to concerns about "selection and/or discrimination" against people who use medications in these classes.33 The formulary guidelines do allow plans to place some antidepressants and antipsychotics on higher cost-sharing tiers. Moreover, the CMS has not prohibited plans from using other drug cost containment tools to manage the use of psychotropic medications (such as prior authorization, step therapy, therapeutic substitution, or quantity limits). If plan premiums do not adequately reflect the heterogeneity in beneficiaries’ expected drug spending, plans could have an incentive to use these tools to restrict use of psychotropics in a way that discourages enrollment by people with mental illnesses. The advantage of tools such as prior authorization is that they might be more responsive to individual patients’ needs than more blunt instruments such as beneficiary cost sharing or closed formularies, which are imposed uniformly on all beneficiaries, regardless of clinical need. The disadvantage is that they will be more difficult for regulators to monitor. The flexibility given to plans to use pharmacy utilization management tools, while necessary for cost containment, leaves open a primary means with which plans may stint on psychotropic medications to avoid adverse selection. The CMS might consider additional strategies to discourage stinting, including alternative forms of risk sharing and careful monitoring of plans’ use of pharmacy management tools. Additional strategies. Risk sharing for "high risks." The system of risk sharing now in place for Part D involves retrospective risk sharing for plans that attract high risks. This system reduces plans’ financial risk without reducing selection incentives. An alternative system is for Medicare to share risk with plans for a subset of beneficiaries identified prospectively as being high risks.34 Instead of receiving a per capita monthly premium for all enrollees, plans could identify people ex ante for whom Medicare would reimburse all or some acceptable spending. For example, plans could designate beneficiaries with schizophrenia and other serious mental disorders as being high risks and could be reimbursed by Medicare for part of these beneficiaries’ drug costs. Because the population and the amount of the payment would be determined prospectively and not be based on actual costs, such a system would dampen the incentive for plans to discourage enrollment of high risks without reducing incentives to manage the drug benefit efficiently.35 Monitoring plans’ use of drug utilization management tools. The CMS’s requirement that PDPs offer "all or substantially all" antidepressants and antipsychotics is supported by the scientific literature. Evidence suggests that mental illnesses are biologically heterogeneous; that medications differ in side effects, contraindications, and other important characteristics; and that people with mental illnesses must often try several medications before finding the right treatment match.36 For instance, studies of antidepressants show that these medications have similar rates of effectiveness at the population level but not at the individual level.37 We know very little, however, about patient-level predictors of success or failure with a particular medication.38 Given the enormous price differences in some drug classes, PDPs may put prior authorization or step therapy systems into place that create incentives for physicians to prescribe lower-cost therapeutic agents. Let us compare two antidepressants: fluoxetine and sertraline, both SSRIs. Sertraline, which maintains patent protection and is available only by brand name, has a retail price three times that of fluoxetine, which is available in generic form.39 A PDP may place sertraline on the highest tier of the formulary, requiring a substantial copayment of $30–$50, or may require that consumers try fluoxetine or some other generic antidepressant before they can fill a prescription for sertraline. Such a policy would steer physicians and consumers toward use of lower-cost drugs and give PDPs the ability to negotiate lower prices with drug companies. In the classes where the CMS is requiring that PDPs include all or substantially all drugs, these management tools will be important mechanisms that plans can use to negotiate discounts or rebates with drug companies. But the impact of these management tools on treatment patterns and health outcomes will depend on how they are implemented in practice. For instance, how cumbersome or time-consuming is a prior authorization process for physicians? Some prior authorization programs require physicians to document the dates and outcomes associated with all previous medication trials before they will pay for the restricted medication. PDPs may require that a certain form be used and that the information be transmitted by fax, and physicians and their patients may wait a few days for a response. These features create barriers to medication access. Depending on how these systems are implemented in practice, they could either promote cost-efficient, high-quality care or limit access to drugs, disrupt treatment regimens, or encourage disenrollment of a vulnerable population. Some patients will respond to the medication on the lowest tier; others will require several trials and flexibility in moving from one therapy to another. Little is known about how well tools such as prior authorization affect use patterns and health care costs for people with mental illnesses because they have not been widely adopted by systems that serve large numbers of these patients (such as the Department of Veterans Affairs or Medicaid). Use of tools such as step therapy in the antipsychotic class may complicate problems with treatment nonadherence and discontinuation among people with schizophrenia.40 The CMS could monitor the use of these tools and their impact on treatment patterns and quality of care in several ways. First, it could conduct surveys of physicians or audit utilization management systems such as prior authorization or step therapy to determine how great a burden they place on providers and consumers. Second, it could monitor beneficiaries’ enrollment, disenrollment, and plan-switching behavior. Plans with disproportionately high levels of disenrollment by beneficiaries with mental illnesses may be using pharmacy management tools in such a way that imposes heavy restrictions on psychotropic medications on which beneficiaries depend. Third, the CMS could examine the impact of formularies, cost sharing, and management tools on utilization patterns of psychotropic drugs and other medications commonly used by people with mental illnesses. Fourth, and most importantly, the CMS could assess the impact of various plan features on service use and health outcomes. Some outcomes, such as ED visits and inpatient hospital stays, can be observed from claims data. A rigorous quality assessment effort will necessitate that PDPs submit all utilization data to the CMS for analysis. Assessing the impact of Part D on other important outcomes such as consumer satisfaction will require data collection. The sheer number of plans operating in the Part D market represents a major challenge to regulators monitoring quality of service delivery and ensuring adequate provision of medications. In the future, once the CMS and the plans have more experience managing the drug benefit, Medicare could require that plans meet certain performance standards and create financial incentives to encourage them to meet these targets. MMA contains a number of important demonstration projects related to quality improvement. Under the Premier Hospital Quality Incentive Demonstration, Medicare is providing financial incentives to nearly 300 hospitals to improve their performance on thirty-four quality measures for five clinical conditions. In spite of the prevalence of mental illnesses among Medicare beneficiaries, no mental health conditions or measures are included at this time.
The extent to which Medicare beneficiaries benefit from the new drug benefit will depend on the previous sources of coverage, level of drug spending, and specific characteristics of the plan they choose. Beneficiaries who lacked prescription drug coverage prior to 2006 will in general have lower out-of-pocket spending under Part D.41 Depending on the features of a plan chosen, such as the placement of drugs on tiers in the formulary and use of pharmacy management tools, a beneficiary who is stabilized on a psychotropic medication enrolling in a plan that places that drug on the highest tier in terms of cost sharing may be forced to either switch drugs or take on a high out-of-pocket burden. An elderly depressed Medicare beneficiary with a lower income who does not qualify for the low-income subsidy might be better off enrolling in a plan with a high monthly premium if that plan has placed the antidepressant they are taking on a lower tier. A disabled dually eligible beneficiary with schizophrenia who is stabilized on an antipsychotic medication will be better off in a plan that allows him or her to maintain use of that drug. This beneficiary will not be exposed to high out-of-pocket costs in terms of premiums, deductibles, or copays, but the flexible application of pharmacy management tools will be very important to avoiding disruptions in treatment and allowing for medication changes. Because mental illnesses frequently co-occur with other chronic medical conditions, it will be important for beneficiaries with mental illnesses to enroll in plans that offer adequate coverage of other classes as well. The more information available to beneficiaries on plan formularies, the better able they will be find a plan that meets their needs. Yet along with this information comes a greater risk of adverse selection, and the incentive for plans to restrict use of psychotropic medications to encourage the disenrollment of beneficiaries with mental illnesses. Policymakers should consider additional forms of risk sharing and close monitoring of plans’ use of management tools to ensure adequate access to psychotropic medications for this vulnerable population.
Julie Donohue (jdonohue{at}pitt.edu) is an assistant professor in the Department of Health Policy and Management, Graduate School of Public Health, University of Pittsburgh (Pennsylvania). The author thanks Richard Frank, Howard Goldman, Marcela Horvitz-Lennon, Judith Lave, Harold Pincus, the editors, and two anonymous reviewers for helpful comments on an earlier draft of this paper. Funding from the Treatment Effectiveness Now Project for a related paper is gratefully acknowledged.
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Role Of Psychotropic Drugs...
Mental Health In Medicare
