Best Practices: Using Best Practices to Manage Psychiatric Medications Under Medicaid

Although pharmacy benefit costs remain a relatively small portion of total U.S. health care costs (approximately 10 percent), they have increasingly attracted the attention of benefit managers because their annual rate of increase has been higher than other areas of health care. The costs of psychiatric medications, in particular, have increased more rapidly than those of other therapeutic drug classes; in state Medicaid fee-for-service plans, these medications usually represent a third of the total pharmacy benefit costs (1). To control costs, pharmacy benefit managers have relied primarily on mandatory, coercive interventions such as step therapy, preferred drug lists, and prior authorization algorithms based solely and primarily on cost (1).

Initially, pharmacy benefit managers avoided placing these restrictions on psychiatric medications because of the complexity of treating mental illnesses and because of strong advocacy from consumer groups. However, recent public budget crises are increasingly forcing the inclusion of psychiatric medications in utilization and cost control interventions. But there is increasing evidence that such cost-based mandatory approaches often result in unanticipated clinical problems and actually create the opposite effect on health care costs—they increase the rate of polypharmacy, increase the use of inpatient and emergency department services as a result of inadequate access to the most effective medications, and increase the rate of patients' nonadherence to treatment (2,3). In this column we describe a noncoercive education-based physician intervention program focusing primarily on best practices to inform prescribing behaviors.

Background

Missouri's fee-for-service Medicaid pharmacy benefit covers more than half a million Medicaid-eligible individuals. The benefit includes all eligibility categories except for women and children in urban areas who are not permanently and totally disabled or in state custody. Psychiatric medications are prescribed for approximately one-third of the covered eligible individuals by approximately 8,000 prescribers in an open panel. Missouri Medicaid has used step therapy and cost-based prior-authorization algorithms for nonpsychiatric medications but has decided not to use these interventions or preferred drug lists with psychiatric medications. Instead, a program called the Mental Health Medicaid Pharmacy Partnership, directed by the Missouri Department of Mental Health (in cooperation with the state Medicaid authority) and assisted by Comprehensive NeuroScience Inc. (CNS), was implemented. The concept was originated by the second author and is funded by a direct grant to CNS from Eli Lilly and Company.

The partnership program differs from traditional pharmacy utilization management approaches on several key principles. First, it focuses primarily on the quality of clinical care, anticipating cost savings as only a secondary effect. Second, it strives not to punish the many for the sins of the few—the intervention focuses on a subgroup of physicians, a substantial proportion of whose practices differ from the usual best practices. Finally, the program preserves the autonomy of both physicians and patients by relying on feedback, education, and persuasion rather than coercive restrictions. This approach upholds the premise that each individual clinical situation is unique and therefore that pharmacy management interventions should aid physicians in making their own informed medication decisions rather than mandating them.

Program and implementation

The Mental Health Medicaid Pharmacy Partnership program was implemented in Missouri in the spring of 2003. Pharmacy claims data are analyzed monthly and compared against a series of nine clinical quality indicators. These indicators are in the form of "questionable practice" prescription patterns: prescription of three or more antipsychotics, multiple prescribers of antipsychotics, failure to refill an antipsychotic prescription, two or more anxiolytics or sedative-hypnotics, multiple prescribers of anxiolytics or sedative-hypnotics, a child receiving three or more psychotropics, polypharmacy in several therapeutic classes, an unusually high or low dosage of antipsychotics, and rapid switching of antipsychotics.

These indicators were chosen on the basis of best practice guidelines and expert consensus, including practice guidelines of the American Psychiatric Association, the Expert Consensus Guidelines Series, and the Texas Medication Algorithm Project. Every month selected prescribers receive mailings that include a quality control informational letter alerting them to possible deviations from best practice, a 90-day pharmacy claims history of any patient in their practice to whom any of the selected indicators applied, a benchmark report comparing physicians' prescribing practice with that of their peers, and a medication best practice briefing related to the particular clinical issue. All prescribers have been offered expert consultation in psychopharmacology issues, which can be provided by respected local Missouri psychiatrists under contract with CNS. Any selected prescriber with a particularly high volume of quality edits receives calls initiated by the expert consultants.

The primary goal in implementing this initiative has been to engage prescribers in a dialogue about the current standard of practice in prescribing psychiatric medications. The general duty of all physicians is to prescribe within the usual accepted standard of practice. However, with the many new advances in psychopharmacology in recent years and the availability of new medications, the accepted standard of practice for prescribing psychiatric medications is not nearly as clear now as it was ten years ago.

As part of the implementation process, an advisory council was recruited, consisting of representatives from the three Missouri psychiatric societies, the psychiatry departments of the four Missouri medical schools, the Missouri medical society, private practice psychiatry groups, the mental health advocacy community, and the community mental-health centers. The role of the advisory council was to assist in prioritizing the clinical quality indicators most pertinent for prescribing practices in Missouri so that these can be used in the quality improvement approach and data analysis. In addition, the advisory council's role was to review the templates for cover letters and educational materials to ensure that they reflected the current standard of care in Missouri, that they were appropriate in tone and content, and ultimately that they would assist in educating their colleagues about the intent and purpose of this initiative.

All Missouri Medicaid prescribers were mailed a brief description of the initiative and the reasons behind it. Multiple presentations were made throughout the state at psychiatric society meetings, departmental grand rounds, and pharmaceutical industry-sponsored dinner meetings to describe the initiative and its intent and to present the initial analysis of the overall practice patterns with the rates at which potential quality-of-practice issues had been occurring. All communications included repeated requests for feedback and recommendations about the appropriateness and usefulness of the materials. They communicated and solicited additional ideas for improvements to this particular project and any other ideas for improving clinical quality and pharmacy utilization management in the area of psychiatric medications.

As a result of the numerous recommendations and feedback received, the mailings have been continually modified, and no two monthly mailings have been identical. The main cover letter for each new monthly mailing identifies the changes that have been suggested by the community of prescribing physicians. Some pharmaceutical manufacturer representatives were given separate briefings and background information in an attempt to reduce any anxiety and misinformation related to this new intervention.

Program results

The initial analysis of prescribing patterns confirmed some of our assumptions. For example, the most common drug class prescribed for persons receiving psychiatric medications was antidepressants (63 percent). On the other hand, the analysis contained many surprises. For example, the use of antipsychotic medication was surprisingly high, occurring among 29 percent of persons receiving any psychiatric medication. Antipsychotic polypharmacy was lower than anticipated, present among fewer than 9 percent of persons receiving antipsychotics. Rates of antipsychotic use above the usual recommended dosage and rates of rapid switching of antipsychotics were both lower than anticipated. Nonantipsychotic mood stabilizers were found to be the fastest growing class by percentage in terms of utilization and cost. About 2 percent of all prescribers of psychiatric medications accounted for more than half of prescribing practices that were outside of the recommended best practice guidelines.

Initially prescribers had a low response rate to the intervention in terms of both change to their prescribing patterns and their feedback. However, the response rate increased remarkably around the fourth mailing. In addition to contacting prescribers with the results of the analysis, we notify them about failure to refill and multiple prescribing. In total, approximately 3,000 contacts are made monthly, with about a 5 percent response rate via mail, fax, e-mail, or telephone.

Because the pharmacy claims analysis is updated monthly, it has been possible to track changes in individual prescriber practices over time. The change in response to the mailings has been surprisingly robust. Depending on the quality indicator, there has been considerable variation in the percentage of prescribers who stop the identified practice over a six-month period (from a low of 31 percent to a high of 98 percent). For example, of 4,400 prescribers who received letters addressing the concurrent prescription of at least two atypical antipsychotic medications within a 60-day period, 2,546 (60 percent) ceased this practice after the first six-month period, resulting in an average response time of 44 days. Prescriber acceptance has been higher than anticipated, with positive comments outnumbering negative comments by four or five to one. Compared with psychiatrists, nonpsychiatrists are more likely to return positive comments and less likely to return negative ones.

It is our impression that secondary fiscal savings have substantially exceeded the cost of the project. We are preparing a detailed report on the analysis of the current patterns of practice, physicians' response to the intervention, and its cost savings.

Discussion and conclusions

It is our hope that the current crisis presented by rising pharmaceutical benefit costs and shrinking health care budgets will be successfully addressed by means other than mandatory coercive formulary restrictions on the part of pharmacy benefit managers. On the basis of our experience in behavioral health care, a multidisciplinary approach will be necessary for achieving success. The program described here has demonstrated that pharmacy benefit managers can focus more on helping prescribers to improve the quality of clinical care. Prescribers in Missouri focused and dedicated their attention to providing the most cost-effective psychopharmacologic interventions, including factoring cost into the quality portion of their equation.

In the coming months, the partnership mailings will include information about the cost effects of the questionable-practice prescribing patterns or clinical quality indicators that the program addresses. We believe that pharmaceutical manufacturers should focus their educational and sales efforts on reducing inappropriate use of their products. Advocacy groups should identify utilization management strategies that they can support, not just ones that they oppose. State departments of mental health must assist their Medicaid agencies in appropriately controlling psychopharmacy benefit costs by focusing on improving quality and facilitating better relationships among the Medicaid agency, mental health providers, and mental health advocates. The results of the partnership so far show that this approach can change patterns of practice to conform to consensus standards of quality without mandating nonindividualized medication changes that can infringe on the autonomy and judgment of patients and physicians.

Acknowledgment

The authors thank Eli Lilly and Company for its support of the program described in this column.

Dr. Parks is medical director of the Missouri Department of Mental Health in Jefferson City and clinical assistant professor of psychiatry at the University of Missouri, Columbia, Missouri 65211 (e-mail, [email protected]). Dr. Surles is senior vice president of Comprehensive NeuroScience in Pennington, New Jersey. William M. Glazer, M.D., is editor of this column.