Referral for coronary angiography after exercise testing: traditional decision-making versus decision analysis

To audit the referral decisions made by a single cardiologist for coronary angiography after exercise testing, we retrospectively reviewed the charts of 303 consecutive patients in a community hospital. The outcomes of these decisions, in terms of angiograms performed and quality-adjusted life expectancy gains as predicted by a decision analysis model, were compared with the theoretical decisions that would have been made using the model. The 97 patients sent for angiography exercised for a shorter time (5.6 +/- 3.1 vs. 6.9 +/- 3.2 min, p < 0.001) had more ST deviation (2.7 +/- 1.4 vs. 1.7 +/- 1.0 mm, p < 0.001), more angina (53.6 vs. 36.9% of patients, p < 0.01) and were more likely to have had a previous myocardial infarction (59.8 vs. 33.5% of patients, p < 0.001) than the 206 not referred. However, of those not referred, 137 were each predicted to gain up to 5.7 quality-adjusted life years (QALYs) from bypass surgery. The overall predicted gain from the cardiologist's decisions was only 0.1 +/- 2.5 QALYs/patient. Had the decisions been made using the model, the mean gain would have been 1.9 +/- 1.3 QALYs/patient, and an extra 128 patients would have been sent for angiography. Decision analysis makes consistent decisions with defined risks and benefits. Such decisions can be reproduced, reviewed and analysed, whereas traditional decision-making may inconsistently reflect the clinician's beliefs and values.