VA Cooperative Studies on Anti-Epileptic Drugs: Do They Impact Practice?

Session number: 1091

Abstract title: VA Cooperative Studies on Anti-Epileptic Drugs: Do They Impact Practice?

Author(s):
MJ Pugh - Center for Health Quality, Outcomes, and Economic Research
J Knoefel - Neurology Department, VAMC Albuquerque
A Charbonneau - Center for Health Quality, Outcomes, and Economic Research Boston University Medical Center
L Kazis - Center for Health Quality, Outcomes, and Economic Research Boston University School of Public Health
DR Berlowitz - Center for Health Quality, Outcomes, and Economic Research Boston University School of Public Health Boston University Medical School

Objectives: VA cooperative studies convincingly demonstrated that, despite equal efficacy, phenytoin and phenobarbital are more likely to cause adverse side-effects and drug interactions than newer medications. These negative effects were considered so problematic, particularly for the elderly, that the current VA cooperative study did not even consider them. We used pharmacy data to determine the extent to which these potentially problematic medications were used in the VHA in FY1999 and FY2000.

Methods: We linked administrative (FY1998-FY1999) and pharmacy databases (FY1999-FY2000) to identify veterans with epilepsy who were on anti-epileptic drugs (N=71,818). We examined the extent to which veterans received phenobarbital or phenytoin in FY1999 and FY2000. Veterans seen in a neurology clinic visit were coded as such for current and subsequent years. Chi-square and logit analyses explored changes over time, differences by age (elderly > 60 years), and differences between those seen in Neurology at some point and those seen only in primary care (PCC).

Results: Over half of all patients were on potentially problematic medications (phenytoin: 44%, phenobarbital: 8%). Little change was noted for phenobarbital, but use of phenytoin decreased slightly from 46% to 43% from FY1999-FY2000 (p<.001). Those seen in Neurology were less likely to receive phenobarbital (8% vs. 9%) and phenytoin (39% vs. 51%), and the elderly were more likely to receive phenobarbital (12% vs. 8%) and phenytoin (52% vs. 39%) (p’s<.001). Logit models indicated a clinic by age interaction for phenytoin. Elderly veterans seen in PCC were more likely, by a factor of 1.84, to receive phenytoin than younger veterans. However, elderly veterans seen in Neurology were more likely, by a factor of 1.60, to receive phenytoin compared to younger veterans.

Conclusions: Some change was evident, but many veterans remain on potentially problematic medications. The more vulnerable elderly population was more likely to be on these medications, especially when seen only in PCC. These patterns suggest delays in dissemination/ adoption, especially for generalists.

Impact statement: Physicians may be unwilling to change an efficient medication despite side-effects, but research increasingly emphasizes the need to factor quality of life issues into this equation. Dissemination of this information is crucial to quality of care.