The objective of the study was to assess factors associated with ward assignment in the emergency room for patients < or = 65 years old with first acute myocardial infarction. We analysed uni- and multivariate predictors for ward assignment (coronary care unit versus internal ward). Eight major centrally located Israeli hospitals provided data during one year. The study population included 1252 patients, of whom 83% were men, 37% were hypertensives, 22% were diabetics, and 14% had previous anginal syndrome. Most patients (83%) were admitted to the coronary care unit. Internal medicine ward assignment was significantly associated with advanced age, history of hypertension or diabetes, a longer time from appearance of symptoms to arrival at the hospital, and myocardial infarction type (non-Q-wave or non-anterior). The likelihood of medical ward referral increased stepwise with the increasing number of a patient's predictive factors: those with > or = 4 factors had a > 30% chance of being assigned to a medical ward compared to a < 10% chance when there were 0-3 risk factors. Exclusion of patients with thrombolysis had no effect on the results. The shortage of cardiac care unit beds apparently leads to emergency room selection acting in detriment of patients with poorest prognoses. Clear guidelines for decision making in the emergency room are needed to resolve this paradoxical situation.