Medicaid programs have struggled for some time to balance their charge to provide high-quality health care with their need to contain cost increases, particularly in the area of pharmaceutical spending. Medicaid prescription drug spending, which currently accounts for about 13 percent of total costs, has increased by some 14 percent annually in recent years, making it one of the fastest-growing components of program spending (1). States have attempted to control this growth through a number of strategies, including the use of preferred drug lists (PDLs).
As of March 2005, at least 38 Medicaid programs had implemented or enacted PDLs (2). Reports of cost savings in the first states to adopt PDLs (3) undoubtedly fueled this trend. However, as PDLs have been rapidly implemented across the country, the collection and use of evidence related to how they affect beneficiaries have not kept pace, making PDLs major policy experiments in balancing cost control with access to vital care for some of the sickest and poorest Americans.
Medicaid law limits a state's ability to restrict access to care, which is a critical protection given the vulnerable population the program serves. This protection may be particularly important in the case of prescription drugs, because the elderly and disabled enrollees who constitute roughly a quarter of Medicaid beneficiaries currently account for more than 85 percent of the program's drug spending. There is considerable evidence that diminished access to drugs for these populations can result in adverse effects on health status and increases in the use of more expensive types of care (4).
To protect patients' access to medically necessary drugs, Medicaid statute does not permit hard formularies. With few exceptions, section 1902(a)(54) of the Social Security Act requires states to cover all approved drugs manufactured by companies that have signed a federal drug rebate agreement with the Secretary of Health and Human Services. Generally, a Medicaid PDL is a type of formulary that comprises drugs that the program agrees to reimburse without prior authorization. A prescriber or pharmacist must obtain permission from the state before a drug that is not on the list is dispensed. By using prior authorization as the tool by which it enforces the PDL, the state creates a procedural barrier to access to nonpreferred medications. PDLs are also used as leverage in price negotiations, given that states often request supplemental rebates from manufacturers to lift prior authorization requirements for nonpreferred drugs. Therefore, in practice, the PDL must be evaluated in at least two ways: how the list is developed (the strength of clinical evidence and financial considerations and how they are balanced) and how the list is implemented, including the responsiveness of the system to authorization requests and appeals when authorization is denied.
Medicaid PDLs must be established by an authorized committee appointed by the governor or state drug utilization review board. In recognition of the greater need of the population served by Medicaid, Medicaid law permits only limited drug exclusions, requires states to respond to requests for authorization within 24 hours, requires the provision of a 72-hour supply of the drug in question in an emergency, and, to handle disputes, requires that an appropriate appeals process be in place. States recognize that PDLs and prior authorization may pose undue burdens on some beneficiaries and many choose to exempt medications in certain classes—for example, antiretroviral or psychotropic drugs—or for individuals with particular diagnoses.
Some states work hard to ensure that they are using the best possible data to construct their PDLs. For example, Oregon's evidence-based approach to drug selection is widely praised, and several states are using Oregon's data reports to develop their own PDLs (5,6). However, PDL construction is only one part of the PDL process, and case study research in several states documents various ways that PDL implementation can hinder Medicaid beneficiaries' ability to get needed medications (7,8). There are many opportunities for Medicaid beneficiaries to slip through the cracks of the PDL process, whether because of relatively strict lists, long wait or response times for authorization, or physicians and pharmacists who may feel overburdened and less able or inclined to pursue authorization. In addition, some patients, for a variety of reasons, including cognitive impairments or transportation difficulties, have trouble navigating the system, and states are often unaware of how elements of the process, such as prior authorization, affect beneficiaries (9).
The future of drug utilization control under Medicaid is complicated by the implementation of Medicare Part D. The seven million people who are dually eligible for Medicaid and Medicare, a group that accounts for half of Medicaid outpatient drug spending, will switch to Medicare coverage on January 1, 2006. Although this change would seem to provide states with fiscal relief, the experiences of individual states are likely to vary. States still face a number of drug-related costs after implementation of Part D, including payments to the federal government to help fund Part D (the "clawback") and potentially diminished ability to obtain supplemental rebates. In addition, if states choose to wrap around the Part D benefit for the dually eligible, they will still have drug outlays for this population, but without federal matching payments.
Federal funds for Medicaid are likely to decrease in other ways: 29 states face reductions in their federal match rate in October 2005, and the federal budget has targeted a reduction in Medicaid spending totaling $10 billion over four years. The resulting increased financial pressure is likely to stimulate state responses to control utilization and spending, such as increasing copayments and decreasing reimbursements to pharmacies.
States face hard decisions in time frames and under circumstances that do not allow for the rigorous testing of every policy decision. However, as the Medicaid drug benefit changes and cost and utilization controls evolve, what is known should be taken into consideration. Many Medicaid beneficiaries are vulnerable and highly dependent on access to drug therapies. Assuming that the goals of any Medicaid drug benefit are to deliver high-quality care and to control drug utilization and cost increases without generating increases in other expenses, the best controls will direct expensive or risky medications toward the patients who can truly benefit from them rather than imposing blunt procedural barriers that hinder both necessary and unnecessary access. A heavy reliance on strategies to curb drug spending growth that unduly burden beneficiaries potentially harms those who are most in need and will not produce a long-term solution to a problem that plagues not just Medicaid but the health care system in general.
Many state governments now face an ongoing budgetary crisis, with Medicaid expenses one of the driving cost centers. Given the intensity of the political forces associated with these circumstances, it is hardly surprising that program administrators are aggressively seeking means of reducing costs. One of the more popular strategies now in play, the use of PDLs, is seemingly a source of significant savings. However, this strategy not only may result in an increase in the overall use of health care resources but also is associated with a significant social impact as a form of health care rationing to the elderly and increased costs to the working poor. Finally, the use of PDLs further usurps the prescriber's professional autonomy while placing psychiatric patients at special risk in order to achieve what may be an illusory budgetary benefit.
Briefly, a PDL is a list of drugs, created by a Medicaid program, for which prescribing is acceptable. Products are excluded from the list subject to a requirement for prior authorization. This process circumvents federal restrictions on the use of formularies in Medicaid while still modifying prescribing patterns. In theory, the savings attributable to PDLs are generated by promoting the use of less expensive alternatives. In fact, any savings come from collection of "extra-federal" rebates from pharmaceutical firms, with rebates for drugs on the PDL exceeding what is required by the Omnibus Budget Reconciliation Act of 1990 (OBRA '90).
There is reason to question whether PDL systems generate savings at all. January's issue of the American Journal of Managed Care is devoted entirely to this topic. In that special issue, Tamer Abdelgawad and I reported evidence of unintended consequences of PDL implementation in the form of increases in hospitalizations and office visit expenses for patients subject to PDL restrictions (10).
It should be recognized that such a system is a de facto rationing of health care that particularly affects older patients, who constitute a substantial fraction of Medicaid recipients. Although rationing may be defensible social policy, the PDL mechanism is flawed. Because the "extra-federal" rebates are voluntary, rational economics suggests that the firms who are the most willing to pay the rebate are those whose products command a smaller share of the market, presumably because they are either less efficacious or have a more objectionable side-effect profile than more successful alternatives. If that is true, the vulnerable populations served by Medicaid may be receiving effective medications with less efficacy and more side effects.
The impact is not limited to Medicaid recipients. Consider the physician whose practice includes a large fraction of Medicaid recipients. It seems reasonable to assume that in many cases, the rest of the practice is likely to consist of patients who are economically disadvantaged. As shown by Wang and colleagues (11), the imposition of Medicaid restrictions results in "spillover" effects to non-Medicaid patients in these circumstances. Physicians learn the Medicaid "ropes" out of necessity but export them to their other patients as well. If you accept the possibility of additional cost and health care burden associated with these restrictions as discussed above, PDLs impose these additional burdens, these economic externalities on that segment of the population that is least able to bear the cost—the working poor.
Furthermore, PDLs, although not formularies per se, do effectively restrict physician alternatives. In practice, a small handful of selective serotonin reuptake inhibitors or second-generation antipsychotics may be approved. These drugs then become the acceptable standard for that therapeutic classification. This is the commoditzation of therapy; you end up not prescribing a particular product but a class of therapeutic agents. It eliminates a great deal of the art in medicine, the sometimes intuitive match between product and patient. This is a just another case of the incremental usurping of the physician's professional autonomy, justified in the service of cost containment—even while that cost-containment may prove illusory.
Not least in this discussion and with this readership is the nature of treatment for individuals with psychiatric issues. As is well known, this population is famous for idiosyncratic responses to therapeutic alternatives—what works wonderfully for one patient may often prove disappointing for another, or even provoke an exacerbation of symptoms. In circumstances like these, ready access to the full pharmacopeia is clearly of great importance. Although the exact magnitude of savings to the Medicaid pharmacy budget from PDL implementation remains unclear, the preliminary evidence suggests that the savings are limited at best. It doesn't take too many hospitalizations for acute schizophrenia to eradicate those savings, to say nothing of the patient suffering involved.
Finally, let's not forget the net intent impact on the pharmaceutical industry. There is a straightforward link between profits and research initiatives; companies seek to innovate in the hopes of realizing rewards for their risk taking. The psychiatric population is dependent to a large extent on Medicaid—reduced profits in this area logically will result in fewer investments in this therapeutic area, fewer new tools in the toolbox in the future. Given the incredible complexity of the biochemistry involved, the progress made in psychiatric pharmacotherapy in the past few decades is nothing sort of miraculous. Hobbling that progress for an economic policy of dubious benefit seems imprudent, at best.
Should state Medicaid programs implement preferred drug lists? The answer is, Absolutely. But, it is important how it's done. Our Washington State physicians and pharmacists supported legislation in 2003 (SB 6088, available at www.rx.wa.gov) that mandated evidence-based drug evaluations to determine preferred drug selection in state agencies that purchase or reimburse for prescription drugs for Washington residents. The physicians said that as long as the selection process was based on a credible review of the literature they would support the state preferred drug list (PDL). Washington contracted with Oregon to participate in the Drug Effectiveness Review Project (DERP, www.ohsu.edu/drugeffectiveness) to obtain credible, robust systematic reviews of the evidence for each drug class that is on the Washington PDL. Now, two years later, Washington Medicaid has a success story to show that the physicians were true to their word.
Washington Medicaid's success with the evidence-based PDL is described in the 2005 Report to the Legislature (www.rx.wa.gov). Within four months of implementing 11 drug classes, Washington Medicaid had an overall compliance rate to the preferred drug or drugs of 90 percent. This took into consideration the provision within SB6088 that allows prescribers to "opt out" by writing "Dispense as written" to get the nonpreferred drug that is medically necessary for their patients. Another unique provision of SB6088 is the Therapeutic Interchange Program, whereby retail pharmacists are mandated to substitute the preferred drug for a nonpreferred drug on the Washington PDL if the endorsing prescriber indicates that substitution is allowed. Fifty-three percent of the prescriptions for drugs in PDL drug classes for Medicaid clients are written by endorsing prescribers, a measurement of success in recruiting physicians to participate in and support the Washington PDL. It also underscores prescribers' confidence in the drug class reviews conducted by the Oregon Health and Science University's Evidence-based Practice Center (OHSU-EPC).
In July this year, Medicaid implemented its first mental health drug class, the second-generation antidepressants. The OHSU-EPC has published the report on its Web site (www.ohsu.edu/drugeffectiveness/reports/final); and this evidence-based report has been reviewed by Washington's pharmacy and therapeutics committee. SB 6088 provides for grandfathering clients who are already taking antidepressants. On the basis of the evidence, the committee recommended that for new starts, generic drugs should be tried first. It named four antidepressants that must be included on the PDL but stated that the rest were essentially equal and could also be considered on the basis of their net cost to the state.
The atypical antipsychotic drug class has also been evaluated by the OHSU-EPC (the report is on DERP's Web site). The process to select the preferred drug in this drug class will be the same as mentioned above for the second-generation antidepressants.
The DERP reports also help Washington Medicaid establish drug initiatives to promote appropriate prescribing of drugs that are frequently used for off-label indications. One such drug class review is for the second-line antiepileptics that include Neurontin (gabapentin), Topamax (topiramate), and Keppra (levetiracetam). Washington Medicaid has used this report as the topic of discussion with a stakeholder work group of mental health experts, patient advocacy groups, and drug companies to examine the quality of evidence in the literature to support off-label use of these drugs. With the assistance of this stakeholder work group, a business plan has been developed to limit the use of these drugs to only those indications for which there is good-quality evidence of safety, efficacy, and effectiveness to support their use beyond FDA labeling.
The evidence-based Washington PDL has had a positive influence on the state's Medicaid program in the following ways. First, there has been excellent provider and community support for rational drug therapy that is safe, efficacious, and cost-effective. Second, reliable, nonbiased drug information is provided to prescribers to help them select appropriate drug therapy. Third, drug manufacturers are using national standards for reporting drug information so that critical elements of the drug's safety, efficacy, and effectiveness can be more easily determined. Fourth, drug manufacturers are competing for the lowest net cost by submitting supplemental rebate offers to state Medicaid programs in drug classes for which the evidence clearly shows that the drugs are essentially the same. Fifth, grading the quality of evidence is catching on, bringing attention to good-quality, well-designed research studies and uncovering the flaws of poorly designed ones. Sixth, limited state resources are now being spent wisely on rational drug therapy instead of on the drugs most highly promoted. Finally, early evidence of safety issues in a report on nonsteroidal anti-inflammatory drugs (NSAIDs) provided an early warning for Washington Medicaid to require prior authorization for Vioxx (rofecoxib) before it was pulled from the market.
In summary, if a state Medicaid program is going to establish a PDL, it is important to incorporate evidence-based drug class reviews by a credible research center into the PDL selection process. In doing so, the state is serving its providers, clients, and taxpayers well. States are using taxpayers' money to evaluate what is the best drug to use for the client's condition, and when the evidence shows that there is little or no difference between drugs in terms of safety, efficacy, and effectiveness, it is important that states use the lowest-cost drug to protect the state's precious resources.
With Medicaid funds in short supply, there has been an increasing tendency to attempt to conserve these funds by adopting managed care plans, which typically include restricting access to medications by creating preferred drug lists (PDLs). The National Alliance for the Mentally Ill (NAMI) has very serious concerns about restrictions on access to psychotropic medications that result from such PDLs.
In many cases, such restrictions may be reasonable. For example, recently my Medigap insurer decided that that it would no longer support the use of Vancenase (beclomethasone) for allergy treatment but would support Flonase (fluticasone). Upon making the change in this allergy medication, I realized that there is no discernible difference in the effectiveness of these two products, at least in this instance.
However, the same is not true for medications that are used to manage severe and persistent mental illness. It is clear that there is a biological—and therefore genetic—basis for severe and persistent mental illness, but it is also evident that the hereditary aspect of these illnesses is complex (12). This is quite different from the inheritance of diseases such as phenyketonuria or sickle-cell anemia, each of which is determined by single-gene mutations. Because the heredity of mental illness involves multiple genes, and because various combinations of these genes are present in different individuals who manifest symptoms of mental illness, it follows that particular medications designed to treat these illnesses will not be equally effective in all cases (13). Medications that successfully manage schizophrenia for person A may or may not be effective in managing person B's schizophrenia. One size definitely does not fit all. It therefore is crucial that psychiatrists who treat these illnesses be allowed the discretion to have access to the full armamentarium of available medications to maximize the chances for the successful treatment of the patients under their care. To do less will seriously jeopardize persons with severe mental illness, their families, and society at large. Such restrictions would be penny-wise and pound-foolish.
Furthermore, there is evidence (14) that there can be unintended expenses incurred by restricting access to medications and that these expenses can reduce or eliminate the "savings" realized by restricting access to medications. Having to make do with medications that are suboptimal in their effectiveness can result in costly increases in the number of visits to physicians and in hospitalizations.
It seems clear that the assumptions being made to support the claim that adoption of PDLs reduces the cost of treatment and does not compromise the effectiveness of treatment must be reevaluated, especially given that such policies affect the lives of persons who have severe and persistent mental illness.