A Medicare database review found that physician preferences increasingly outweighed patient characteristics as determinants of first-time prescriptions for COX-2 inhibitors

J Clin Epidemiol. 2005 Jan;58(1):98-102. doi: 10.1016/j.jclinepi.2004.06.002.


Objective: Although innovative drugs may be underprescribed by some physicians, it is possible that rapid adoption after market introduction may lead to prescribing such drugs to patients without a clear indication. We sought to quantify the relative contributions of patient vs. physician factors to the decision to prescribe selective cyclooxygenase-2 (COX-2) inhibitors during the first 2 years of their availability.

Methods: A cohort of 37,957 Medicare beneficiaries who were enrolled in the Pharmaceutical Assistance Contract for the Elderly in Pennsylvania was identified. All patients had started using nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) or selective COX-2 inhibitors between January 1, 1999, and December 31, 2000, and had no prior NSAID use. All had full prescription drug coverage, including NSAIDs and selective COX-2 inhibitors. Subsequent prescriptions were not considered. We quantified the amount of variation in first-time COX-2 prescribing that could be explained by predictors of gastrointestinal (GI) toxicity, other patient characteristics, or physician preferences. Explained variation was calculated as the R(2) (standardized to range from 0 to 1) from unconditional logistic regression and random intercept mixed effects logistic regression, fitted separately in each of eight consecutive 3-month periods.

Results: COX-2 inhibitors were adopted as the preferred NSAID by 55% of physicians within 180 days after they were marketed. In new NSAID users, COX-2 prescribing was twice as dependent on physician prescribing preferences (R(2)=60%) as on the combined predictors of GI toxicity (R(2)=3%) and other patient factors (R(2)=30%). The ratio of COX-2 prescribing explained by physician preferences over the contribution of patient factors increased from 2 to more than 10 over a 24-month period.

Conclusions: First-time COX-2 inhibitor prescribing was somewhat dependent on patient factors in the first quarter of marketing, but the proportional influence of physician preferences increased substantially over the following 2 years, raising the question of why physician factors and not patient risk factors influence COX-2 inhibitor prescribing.

Publication types

  • Research Support, Non-U.S. Gov't
  • Research Support, U.S. Gov't, P.H.S.

MeSH terms

  • Aged
  • Anti-Inflammatory Agents, Non-Steroidal / adverse effects
  • Cohort Studies
  • Cyclooxygenase Inhibitors* / therapeutic use
  • Drug Utilization Review / methods*
  • Female
  • Health Services Research
  • Humans
  • Male
  • Medicare / statistics & numerical data
  • Patient Selection
  • Pennsylvania
  • Practice Patterns, Physicians' / statistics & numerical data*
  • Risk Factors


  • Anti-Inflammatory Agents, Non-Steroidal
  • Cyclooxygenase Inhibitors