Description of a methodological approach to verify the outcome-optimization of tailored therapeutic choices and test application to PCI vs. CABG.

Abstract

Background: The scientific evidence guiding the choice between Percutaneous Coronary Intervention (PCI) and Bypass Graft Surgery (CABG) is inconclusive. Yet, tailoring the choice to the patients' characteristics is generally considered important to optimize outcomes. Objectives:To verify the supposed outcome benefits of tailoring the choice of the revascularization procedure. Methods:We calculated a propensity score (PS) -i.e. the probability, given the patients' characteristics, of undergoing one of the two alternative procedures -for the 11,750 patients with severe coronary disease who underwent coronary revascularization between 2002 and 2008 in an Italian region. Then we investigated the effect-modification of the PS on the Hazard Ratios (HR) of PCI vs. CABG for death, myocardial infarction, stroke and repeat revascularization with a likelihood ratio test and by visual inspection. Results:Only the least important outcome (repeat revascularization) significantly differed across deciles of PS (p=0.05) and its graphical trend supported a favorable effect of the decision process. Conclusions:In agreement with the current scientific uncertainty, but contrary to common opinion, the medical decision process between PCI and CABG based on demographic and clinical factors is marginally capable of optimizing the post-procedural outcomes. The study relies on the assumption that the variables considered by clinicians were among those included in the PS.