To examine the association between hospital-level trial participation and in-hospital mortality, we conducted patient-level analyses. For these analyses, to reduce bias, as has been recommended by others,8 we excluded patients who were transferred to another hospital (20 835 patients [12.0%]). We accounted for within-hospital clustering, whereby patients at the same hospital were more likely to have similar responses to each other relative to patients treated at other hospitals (eg, within-center correlation), by using logistic generalized estimating equation models to estimate marginal effects of percentage of hospital-level trial participation.14,15 This method produces estimates similar to those from ordinary logistic regression, but the variances of the estimates are adjusted for within-hospital clustering of responses. Variables entered into our model included sociodemographic factors (age [continuous], sex, race [white vs nonwhite], and insurance status), family history of coronary disease, medical history (hypertension, diabetes mellitus, smoking, hyperlipidemia, myocardial infarction, percutaneous coronary intervention, coronary artery bypass grafting, heart failure, stroke, and renal insufficiency), features of the initial clinical presentation (ST-segment depression, transient ST-segment elevation, signs of heart failure, heart rate [continuous], systolic blood pressure [continuous], body mass index, and specialty of attending physician [cardiology vs noncardiology]), and all hospital-level characteristics. All analyses were performed using SAS statistical software, version 9.1.3 (SAS Institute Inc, Cary, North Carolina).