Brook and Shekelle raise several important issues related to the validity of the RAND criteria for the appropriateness of coronary angiography. One question involves whether the relationship between appropriateness ratings and the identification of potential life-threatening lesions (left main- and three-vessel disease) is stronger than the relationship between the appropriateness ratings and the presence of any abnormal coronary arteries on angiography. The answer according to our data is no. The percentage of patients who met the appropriateness criteria was as follows: 86% of those with one-vessel disease, 90% of those with two-vessel disease, 86% of those with three-vessel disease, and 92% of those with left mainvessel disease. There was a relationship between having an anatomic disease and appropriateness, since only 48% of those with no disease met the appropriateness criteria. We cannot say how this relationship might be affected by the reliability of the reading of the stress tests and