Mortality and prescription drug use: an analysis using the CHS limited access data

Ongoing interest in the use of drug profiles derived from pharmacy benefit program databases continues as a tool for life insurance risk assessment. The Cardiovascular Health Study (CHS) limited access data may be another way to explore the association between the number of prescriptions used by an individual, classified by drug class, and the risk of death. The CHS captured medication use of participants at baseline. An exploratory analysis, using Cox proportional hazards regression, was performed to assess a simple score based on the number of drug classes used and the hazard of death over the 8 years of follow up available (n=5795). When adjusted for gender, current smoking status, and age, the hazard ratio (HR) of death associated with this score was 1.19 (95% CI: 1.15-1.22). Use of calcium supplements and hormone therapy were associated with favorable mortality risk. As expected, insulin, oral agents for diabetes, CHF therapy, and oral steroids were associated with the greatest increase of mortality risk (49%-101%). A score was also created by summing only the 10 significant drug classes derived from a stepwise model. The HR of death associated with this "significant sum" was 1.37 (95% CI: 1.32-1.44), adjusted for gender, age, and current smoking status. This exploratory analysis suggests that a drug score based on a simple sum of prescription drug classes used could be a useful predictor of mortality in an elderly (aged 65 and over) population.