[Cost/efficacy ratio in the diagnosis of coronary disease. Bayes' analysis by computer: respective role of the exercise test, isotopic methods and coronarography].

UNLABELLED

Between January 1983 and May 1984, 104 patients with no known cardiac pathology were referred by their cardiologist for diagnosis of chest pain. They all underwent coronary angiography which was used as the reference investigation and the following sequential Bayes' analysis was performed. The percentage probability of coronary artery disease was estimated from clinical date (age, sex, characteristics of the chest pain subdivided into 3 groups); an exercise ECG was performed in all cases (classified as positive, negative or non diagnostic); if the probability of coronary artery disease was greater than 95% (or less than 5%) after exercise stress testing the patients was diagnosed as having (or not having) coronary artery disease. If the probability was between 6 and 94% the patient underwent Thallium myocardial scintigraphy (Thallium dipyridamole; analysis on a colour television screen); the coronary risk probability before Thallium was that calculated after exercise stress testing. If after myocardial scintigraphy the coronary risk remained between 6 and 94%, an exercise angioscintigraphy was performed and interpreted in the same way. The clinical and complementary date was analysed on a mini-computer, the values of the sensitivity and specificity of the tests used for the calculation of the probability of coronary artery disease were those previously published by our group.

RESULTS

31/88 (35%) of patients were classified in the 5% risk groups after exercise stress testing (24 coronary artery disease; 7 normals: no errors of classification). Fifty six out of the 88 patients (65%) were classified in the 5% risk group after myocardial scintigraphy (42 patients with coronary artery disease with 41 abnormal coronary angiographies and 14 normal patients, all of whom had normal coronary angiographies; this represents a 1.8% divergence of classification compared with coronary angiography). Angioscintigraphy only classified 3 of the remaining patients, one wrongly, and did not seem to be useful diagnostically as a third-line investigation after Thallium scintigraphy or as a second-line investigation instead of Thallium scintigraphy. This strategy is less costly than carrying out coronary angiography systematically in these patients: if diagnostic coronary angiography is performed alone in patients with a risk of 6 to 94% the cost is 4 800 FF vs 10 400 FF per patient; if coronary angiography is performed in all patients in whom coronary artery disease is possible or certain (all patients with a risk of over 5%), the cost is 8 400 FF vs 10 400 FF per patient, a saving of 20%.(ABSTRACT TRUNCATED AT 400 WORDS)