Biostatistics Weekly Seminar

Previous work suggests an association between opioids and cardiovascular disease (CVD). We study if this association exists using the Veterans Aging Cohort Study (VACS). We determined the exposure to opioids through the pharmacy fill/refill of prescribed opioids. The CVD outcome was defined as the time to acute myocardial infarction, congestive heart failure, ischemic stroke, and receipt of a cardiac revascularization procedure. We implemented two analytic approaches. In the first approach, the exposure was crudely determined as any or no opioid receipt in the first three years following the baseline period. To balance the propensity of receiving prescribed opioids, we used Cox proportional hazards regression with matching weights, a method that approximates propensity score caliper matching. In the second approach, we determined opioid exposure longitudinally and adjusted for the potential time-variant confounder, the VACS comorbidity score, using the framework of Marginal Structural Models. The results of the first approach show that the hazard of CVD was higher among patients who received prescribed opioids compared to those who did not. The results of the second approach is currently showing no evidence of the association between prescribed opioid receipt and the hazard of CVD. This is an ongoing project.

Zoom (Link to Follow)
13 October 2021
1:30pm