Using observational data to estimate prognosis: an example using a coronary artery disease registry

With the proliferation of clinical data registries and the rising expense o f clinical trials, observational data sources are increasingly providing ev idence for clinical decision making. These data are viewed as complementary to randomized clinical trials (RCT). While not as rigorous a methodologica l design, observational studies yield important information about effective ness of treatment, as compared with the efficacy results of RCTs. In additi on, these studies often have the advantage of providing longer-term follow- up, beyond that of clinical trials. Hence, they are useful for assessing an d comparing patients' long-term prognosis under different treatment strateg ies. For patients with coronary artery disease, many observational comparis ons have focused on medical therapy versus interventional procedures. In ad dition to the well-studied problem of treatment selection bias (which is no t the focus of the present study), three significant methodological problem s must be addressed in the analysis of these data: (i) designation of the t herapeutic arms in the presence of early deaths, withdrawals, and treatment cross-overs; (ii) identification of an equitable starting point for attrib uting survival time; (iii) site to site variability in short-term mortality . This paper discusses these issues and suggests strategies to deal with th em. A proposed methodology is developed, applied and evaluated on a large o bservational database that has long-term follow-up on nearly 10 000 patient s. Copyright (C) 2001 John Wiley & Sons, Ltd.