Economics of myocardial perfusion imaging in Europe--the EMPIRE Study.

Abstract

Physicians use myocardial perfusion imaging to a variable extent in patients presenting with possible coronary artery disease. There are few clinical data on the most cost-effective strategy although computer models predict that routine use of myocardial perfusion imaging is cost-effective. To measure the cost-effectiveness of four diagnostic strategies in patients newly presenting with possible coronary artery disease, and to compare cost-effectiveness in centres that routinely use myocardial perfusion imaging with those that do not. We have studied 396 patients presenting to eight hospitals for the diagnosis of coronary artery disease. The hospitals were regular users or non-users of myocardial perfusion imaging with one of each in four countries (France, Germany, Italy, United Kingdom). Information was gathered retrospectively on presentation, investigations, complications, and clinical management, and patients were followed-up for 2 years in order to assess outcome. Pre- and post-test probabilities of coronary artery disease were computed for diagnostic tests and each test was also assigned as diagnostic or part of management. Diagnostic strategies defined were: 1: Exercise electrocardiogram/coronary angiography, 2: exercise electrocardiogram/myocardial perfusion imaging/coronary angiography, 3: myocardial perfusion imaging/coronary angiography, 4: coronary angiography. Primary outcome measures were the cost and accuracy of diagnosis, the cost of subsequent management, and clinical outcome. Secondary measures included prognostic power, normal angiography rate, and rate of angiography not followed by revascularization. Mean diagnostic costs per patient were: strategy 1: 490 Pounds, 2: 409 Pounds, 3: 460 Pounds, 4: 1253 Pounds (P < 0.0001). Myocardial perfusion imaging users: 529 Pounds, non-users 667 Pounds (P = 0.006). Mean probability of the presence of coronary artery disease when the final clinical diagnosis was coronary artery disease present were, strategy 1: 0.85, 2: 0.82, 3: 0.97, 4: 1.0 (P < 0.0001), users 0.93, non-users 0.88 (P = 0.02), and when coronary artery disease was absent, 1: 0.26, 2: 0.22, 3: 0.16, 4: 0.0 (P < 0.0001), users 0.21, non-users 0.20 (P = ns). Total 2-year costs (coronary artery disease present/absent) were: strategy 1: 4453 Pounds/710 Pounds, 2: 3842 Pounds/478 Pounds, 3: 3768 Pounds/574 Pounds, 4: 5599 Pounds/1475 Pounds (P < 0.05/0.0001), users: 5563 Pounds/623 Pounds, non-users: 5428 Pounds/916 Pounds (P = ns/0.001). Prognostic power at diagnosis was higher (P < 0.0001) and normal coronary angiography rate lower (P = 0.07) in the scintigraphic centres and strategies. Numbers of soft and hard cardiac events over 2 years and final symptomatic status did not differ between strategy or centre. Investigative strategies using myocardial perfusion imaging are cheaper and equally effective when compared with strategies that do not use myocardial perfusion imaging, both for cost of diagnosis and for overall 2 year management costs. Two year patient outcome is the same.