Teaching medical students to estimate probability of coronary artery disease.

The authors conducted a randomized trial of two methods for teaching medical students how to estimate the probability of coronary artery disease in patients with chest pain. Eighty-two students were given a pre-test consisting of written protocols summarizing the histories of 25 patients who had undergone coronary arteriography. The students estimated the likelihood of coronary artery disease for each case and were then randomized to receive one of two short written clinical lessons: a cardiology textbook chapter on interpreting chest pain, or a lesson based on a prediction rule for estimating probability of coronary artery disease. All students were given a post-test similar to the pre-test. Students who were given the textbook lesson showed no change in the accuracy of their probability estimates. Students who were taught the prediction rule significantly improved their probability estimates, as measured by a statistical index of calibration. The authors conclude that traditional teaching methods do not provide students with guidance in estimating disease probability, and that better teaching methods are needed. In this study, a clinical prediction rule fulfilled the need for instruction in probability estimation.