Effectiveness of national provider prescription of PPI gastroprotection among elderly NSAID users

OBJECTIVES: Our aim was to quantify the effect of provider adherence on the risk of NSAID-related upper gastrointestinal events (UGIE). METHODS: We identified from national pharmacy records veterans >65 yr prescribed an NSAID, a coxib, or salicylate (>325 mg/day) at any Veterans Affairs (VA) facility (January 1, 2000 to December 31, 2002). Prescription fill data were linked in longitudinal fashion to VA inpatient, outpatient, and death files and merged with demographic, inpatient, outpatient, and provider data from Medicare. Each person-day of follow-up was assessed for exposure to NSAID alone, NSAID+proton pump inhibitor (PPI), coxib, or coxib+PPI. UGIE was defined using our published, validated algorithm. Unadjusted incidence density ratios were calculated for the 365 days following exposure. We assessed risk of UGIE using Cox proportional hazards models, while adjusting for demographics, UGIE risk factors, comorbidity, prescription channeling (i.e., propensity score), geographic location, and multiple time-dependent pharmacological covariates, including aspirin, steroids, anticoagulants, antiplatelets, statins, and selective serotonin reuptake inhibitors. RESULTS: In our cohort of 481,980 (97.8% male, 85.3% white, mean age 73.9, standard deviation 5.6), a safer strategy was prescribed for 19.8%, and 2,753 UGIE occurred in 220,662 person-years of follow-up. When adjusted for prescription channeling, confounders, and effect modification-associated PPI, risk of UGIE was 1.8 (95% confidence interval [CI] 1.6-2.0) on NSAID alone, 1.8 (95% CI 1.5-2.0) on coxib alone, 1.1 (95% CI 0.7-4.6) on NSAID+PPI, and 1.1 (0.6-5.2) on coxib+PPI. When the analysis was adjusted for cumulative percent time spent on a PPI, risk of UGIE decreased from HR 3.0 (95% CI 2.6-3.7) when a PPI was prescribed 0-20% of the time to 1.1 (95% CI 1.0-1.3) when a PPI was prescribed 80-100% of the time. CONCLUSIONS: Provider adherence to safer NSAID prescribing strategies is associated with fewer UGIE among the elderly. An adherent strategy lowers, but does not eliminate, risk of an NSAID-related UGIE.