A1297
Title: A Bayesian approach to studying major adverse cardiovascular events: Leveraging information from clinical trials
Authors: Amber Hackstadt - Vanderbilt University Medical Center (United States) [presenting]
Cara Lwin - Vanderbilt University Medical Center (United States)
Robert Greevy - Vanderbilt University (United States)
Kathryn Snyder - Vanderbilt University Medical Center (United States)
Christianne Roumie - Vanderbilt University Medical Center (United States)
Abstract: Multiple meta-analyses of trials estimated that the risk of major adverse cardiovascular event and heart failure hospitalization outcomes (MACE+HF) was lower for patients treated with sodiumglucose cotransporter 2 inhibitors (SGLT2i) versus dipeptidyl peptidase 4 inhibitors (DPP4i). However, the results were more varied in cohorts of patients without a history of cardiovascular disease (primary prevention). A Bayesian approach is applied to further investigate the association of SGLT2i with MACE+HF in a large primary prevention cohort of veterans. The Bayesian approach allows straightforward incorporation of prior information from other studies, including clinical trials. A Bayesian survival analysis model is used where the covariates are directly modeled in the hazard function, and a flexible M-spline function is used to estimate the baseline hazard. Information is incorporated from previous studies via an informative prior on the coefficient for the treatment effect in the hazard function. The Bayesian approach allows the estimation of the probability of a protective effect of SGLT2i for the MACE+HF outcome and its components. Different choices are examined for the priors. The Bayesian analysis suggested a protective effect for SGLT2i versus DPP4i for the MACE+HF outcome but not all the components of the MACE+HF outcome.
