Effects of prior authorization on medication discontinuation among Medicaid beneficiaries with bipolar disorder.

OBJECTIVE

Few data exist on the cost and quality effects of increased use of prior-authorization policies to control psychoactive drug spending among persons with serious mental illness. This study examined the impact of a prior-authorization policy in Maine on second-generation antipsychotic and anticonvulsant utilization, discontinuations in therapy, and pharmacy costs among Medicaid beneficiaries with bipolar disorder.

METHODS

Using Medicaid and Medicare utilization data for 2001-2004, the authors identified 5,336 patients with bipolar disorder in Maine (study state) and 1,376 in New Hampshire (comparison state). With an interrupted time-series and comparison group design, longitudinal changes were measured in second-generation antipsychotic and anticonvulsant use; survival analysis was used to examine treatment discontinuations and rates of switching medications.

RESULTS

The prior-authorization policy resulted in an 8-percentage point reduction in the prevalence of use of nonpreferred second-generation antipsychotic and anticonvulsant medications (those requiring prior authorization) but did not increase use of preferred agents (no prior authorization) or rates of switching. The prior-authorization policy reduced total pharmacy reimbursements for bipolar disorder by $27 per patient during the eight-month policy period. However, the hazard rate of treatment discontinuation (all bipolar drugs) while the policy was in effect was 2.28 (95% confidence interval=1.36-4.33) higher than during the prepolicy period, with adjustment for trends in the comparison state.

CONCLUSIONS

The small reduction in pharmacy spending for bipolar treatment after the policy was implemented may have resulted from higher rates of medication discontinuation rather than switching. The findings indicate that the prior-authorization policy in Maine may have increased patient risk without appreciable cost savings to the state.