The rise and fall of gabapentin for bipolar disorder: a case study on off-label pharmaceutical diffusion

Abstract

Context: Rising drug costs have increased focus on how new pharmaceuticals diffuse into the marketplace. The case of gabapentin use in bipolar disorder provides an opportunity to study the roles of marketing, clinical evidence, and prior authorization (PA) policy on off-label medication use.

Design: Observational study using Medicaid administrative and Verispan marketing data. We examined the association between marketing, clinical trials, and prior authorization on gabapentin use.

Setting and patients: Florida Medicaid, bipolar disorder-diagnosed enrollees ages 18 to 64 for fiscal years 1994 to 2004.

Results: Gabapentin prescriptions increased from 8/1000 enrollees per quarter in 1994 to a peak of 387/1000 enrollees in 2002. Its uptake tracked marketing efforts towards psychiatrists. The publication of 2 negative clinical trials in 2000 and the discontinuation of marketing expenditures towards psychiatrists were associated with an end to the steep rise in gabapentin prescriptions. After these events gabapentin use remained between 319/1000 and 387/1000 enrollees per quarter until the PA policy, which was associated with a 45% decrease in prescriptions filled. After 1 year, scientific evidence and marketing discontinuation were associated with a 5.4 percentage point decrease in the predicted probability of filling a gabapentin prescription and the PA policy, a 7.1 percentage point decrease.

Conclusions: Pharmaceutical marketing can influence off-label medication prescribing, particularly when pharmacologic options are limited. Evidence of inefficacy and/or the cessation of pharmaceutical marketing, and a restrictive formulary policy can alter prescriber behavior away from targeted pharmacologic treatments. These results suggest that both information and policy are important means in altering physician prescribing behavior.

Figure 1

Interrupted time series of number of prescriptions by quarter and marketing by quarter. Four events have been noted in the time series: (1) the quarter of FDA approval for valproic acid for BPD; (2) the quarters of the publication of the 1^st and 2^nd randomized placebo-controlled trials (RCT) that showed gabapentin’s lack of efficacy in BPD; (3) the quarter in which Florida’s prior authorization (PA) policy was implemented.