OBJECTIVE: To assess the appropriateness of indications for coronary artery bypass graft (CABG) surgery and percutaneous transluminal coronary angioplasty (PTCA). METHODS: A modified Delphi group judgement process with input from a panel of six interventional cardiologists and six cardiopulmonary surgeons. There was one clinician from each of the 12 tertiary referral heart centres in The Netherlands. MAIN OUTCOME MEASURE: Ratings by panel members, on a 1 to 9 scale, of indications presented as a choice between two treatments (CABG v medical treatment, PTCA v medical treatment, and CABG v PTCA) for 1182 model cases. Each case represented a unique combination of clinical features in terms of symptoms, medical history, and results of tests. Ratings were analysed with respect to degree of agreement among panelists, degree of appropriateness of indications, and panel's preference for invasive or medical treatment. RESULTS: The panel agreed on 58.6% and disagreed on 3.2% of the indications. The panel opted for invasive treatment in 48.2% and medical treatment in 22.8%, and had no clear preference for either method in 29.0% of the cases. When compared with medical treatment, CABG was more often rated appropriate than PTCA: 35.4% v 21.6% (P < 0.001). Panel scores depended on severity of anatomical disease. For instance, for 51.5% of the model cases with one-vessel disease not including the proximal left anterior descending artery, the panel preferred medical treatment to invasive treatment, while the latter was preferred in 18% of the cases. In cases with type C lesions, the panel frequently rated PTCA as inappropriate. Panel scores were also affected by nonclinical factors. Cardiologists and surgeons rated the procedure of their own specialty higher than the alternative invasive intervention. CONCLUSIONS: The panel method yields logically consistent scores of the appropriateness of indications for carrying out medical procedures. It may be an aid in formulating clinical practice guidelines.