Concerns about the increasing medical care costs are causing the medical community to focus its attention on the appropriate of diagnostic tests such as echocardiography. Prerequisite to a better utilization of the limited economic resources assigned to our health care system is an analysis of how, why, and with which results diagnostic tests with a widespread use and relevant cost, like echocardiography, are requested. During the last 2 weeks of September 1994, a transversal, observational study was carried out at 13 hospital echocardiographic laboratories. Ordering physician characteristics, reasons for ordering the test, cardiological diagnostic tests previously performed and their relationship with the test results, were evaluated with a questionnaire completed by the physician who performed the test, in all the out-patients undergoing echocardiogram in that fortnight. Five hundred and sixteen consecutive questionnaires were successfully completed. Fourty-five percent of the echocardiograms were ordered by cardiologists, 35% by general practitioners, 10% by internists, and 10% by other specialists. Hypertension (16.4%) and ischemic heart disease (14.8%) were the most common indications for the test, followed by palpitations or arrhythmias (7.5%), mitral valve prolapse or mitral valve disease (7.3%), chest pain or angina pectoris (6.3%), cardiac murmur (5.5%), dyspnea or heart failure (5.2%), aortic valve disease (5%), prosthetic heart valve evaluation (4.6%), others (27%). Before undergoing the echocardiogram, 433 (84%) patients underwent an electrocardiogram, 242 (47%) a cardiological clinical evaluation, 196 (38%) a chest X-ray, and 191 (37%) had had a previous echocardiogram. The most common echocardiographic diagnosis was normal (29.2%) followed by hypertensive heart disease (16.2%), mitral valve disease (12.3%), aortic valve disease (10.5%), ischemic heart disease (9.3%), cardiomyopathy (4.9%) normal prosthetic heart valve function (4.5%), pericardial effusion (3.8%), others (11.3%). Among the echocardiograms ordered by cardiologists, 21.8% were normal in comparison with 35.4% of those ordered by general practitioners (p < 0.004), 35.3% of those ordered by internists (p = 0.04), 35.3% of those ordered by other specialists (p = 0.04). Among the 284 patients whose echocardiograms were not requested by a cardiologist, only 215 (76%) had undergone an electrocardiogram and only 68 (24%) a clinical evaluation by a cardiologist. In these patients, the frequency of normal echocardiograms was not influenced by having undergone a previous electrocardiogram or a chest X-ray. Conversely, patients in whom the echocardiogram was ordered after a cardiology consult showed a significant lower frequency of normal results compared to patients not evaluated by a cardiologist (23% vs 39%; p < 0.05). More than 50% of the echocardiograms performed in out-patients are ordered by physicians who are not cardiologists. Among these echocardiograms, about 1 out of 3 results normal. This finding suggests an improper use of echocardiogram as a screening tool by non-cardiologists in out-patients. A preceding clinical evaluation by a cardiologist, but not an electrocardiogram or a chest X-ray alone, may determine a more appropriate use of the test being associated with a reduced frequency of normal results.