Percutaneous transluminal coronary angioplasty (PTCA) is currently performed in many patients seeking care because of severe manifestations of multivessel coronary artery disease. Previously, the majority of such patients would have undergone coronary artery bypass grafting (CABG). No definitive evidence is available as to which initial revascularization strategy has the best long-term clinical and economic outcomes. The Bypass Angioplasty Revascularization Investigation (BARI) is the largest of several recent clinical trials that were designed to test the hypothesis that an initial strategy of PICA in selected patients with multivessel coronary artery disease does not compromise long-term clinical outcome compared with an initial strategy of CABG. This report describes how patients were screened, selected, and recruited in BARI and how this process may influence the results and the interpretation of the trial. During the enrollment period, 25,200 patients undergoing diagnostic coronary angiography at the participating institutions or with off-site angiograms referred to BARI investigators were screened for BARI eligibility. Excluded from screening were patients without coronary artery disease, those with single-vessel disease, prior revascularization, primary congenital, valvular, or myocardial disease, and age > 80 years. Slightly more than half of the patients screened (12,670) were not clinically eligible for BARI because of left main disease, insufficient symptoms, emergency revascularization, or other logistic reasons. Thus, 12,530 patients had severe angina and/or ischemia and were clinically eligible for BARI. Nearly 33% of them (4,110) had multivessel disease, which was suitable for both PTCA and CABG. Of the 8,420 patients (67%) who were technically unsuitable, 60% were judged not to be candidates for PTCA, only 3% for CABG, and 3% for both procedures. Nearly half (1,829) of the patients who were clinically and angiographically eligible were randomly assigned to either PTCA (915) or CABG (914). The remaining 2,281 patients did not consent to randomization but 2,013 (88%) agreed to be followed in the eligible, not randomized registry. In addition, a random sample of 422 patients judged to be technically unsuitable for PTCA and/or CABG was enrolled in a registry of angiographic exclusions. Finally, to document ongoing revascularization practice patterns, the BARI investigators have provided semiannual 1-week surveys of all revascularizations at their respective centers as well as a 1-time survey at other institutions in the United States and Canada. Interpretation of the results of the BARI requires understanding the patient population entered into the study. Of the 25,200 patients with multivessel coronary artery disease age < 80 years screened for the study, 12,670 were excluded for a variety of reasons (mostly left main disease or insufficient angina to warrant revascularization), and 8,420 were excluded because of technical unsuitability for both procedures. Of the eligible patients, 1,829 were entered into the study and randomized. Nonrandomized patients and a sample of ineligible patients were enrolled into registries. The data from the BARI registries and surveys will be important to put into perspective the relative long-term efficacy and safety of PTCA and CABG in patients who are suitable for both procedures.