Diagnoses ascribed to patients during their hospitalization were categorized as cardiac, pulmonary, both, or neither. Cardiacascribed diagnoses included coronary artery disease, acute myocardial infarction, congestive heart failure, congenital heart disease, valvular heart disease, cardiomyopathy (dilated or hypertrophic), pericarditis, tamponade, arrhythmia, or hypertension. Pulmonary-ascribed diagnoses included acute upper airway obstruction, asthma, chronic obstructive lung disease, pulmonary embolism, pneumonia, pneumothorax, pleural effusion, restrictive lung disease, or lung tumor. Diagnoses ascribed to neither cardiac nor pulmonary systems included anemia, morbid obesity, renal failure, and psychogenic dyspnea.
Medications taken at the time of, or after hospitalization were categorized as cardiac related (diuretics, digitalis, antianginals, antiarrhythmics, antihypertensives, or vasodilators), pulmonary related (bronchodilators, steroids, or antibiotics), both, or neither (hypoglycemics, analgesics, or hormonal medications). When a particular medication (such as an anticoagulant) could be used for either a cardiac or pulmonary problem, it was categorized according to the principal indication for its use in the recipient.
After hospital discharge, inpatient records were reviewed for the following predefined study end points: duration of hospitalization (the primary end point), and any change in diagnosis or treatment after results of echocardiography were made known, or performance of other diagnostic cardiopulmonary studies. All data were analyzed based on “intention to test,” whether or not patients received the randomized intervention. If patients returned to the hospital during the study period (29 [ 15 percent] of the 196 patients), only the first randomization was included in the analysis. Statistical analyses were performed using Students t test (two-tailed) for nonpaired data and X’ analysis. Statistical significance was considered present at p<=05.