Abstract:
|
The PROVE IT-TIMI 22 randomized controlled trial found that the use of higher potency statins had a protective effect (HR: 0.85) on a combined endpoint of major cardiovascular events or death. We estimate the effect of high vs. low statin potency on the outcome of a major cardiovascular event or death within one year, in a cohort of new users of statins within 20 days of a first myocardial infarction, in the Truven MarketScan database. We compare results from different estimation techniques, including targeted maximum likelihood estimation (TMLE) and augmented inverse probability of treatment weighting (A-IPTW). 21,624 individuals with a new statin prescription within 20 days of an initial MI were identified. Patient characteristics were well balanced in the cohort. Incidence of a major cardiovascular event was 44% among high potency statin users, and 40% among low potency statin users. Adjustment via all modeling strategies resulted in very similar risk ratios just below the null (RR: 0.95-0.98). The relative efficiency for TMLE was slightly improved compared to all other models. A-IPTW and TMLE performed similarly, with the TMLE modeling having a very slight precision advantage.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.