Differential effects of experience on house officers' admission diagnoses for patients with possible myocardial infarction.

The effect of the level of training on house staff's abilities to accurately diagnose acute ischemic heart disease (AIHD) and admit patients presenting in an emergency room was assessed. A validated mathematical instrument, which calculates patients' probabilities of AIHD, was used to retrospectively stratify admitted patients. Sensitivity rates of house staff were similar for patients with high probabilities (greater than or equal to 40%) of AIHD (ns). When probabilities of AIHD were low (less than 40%), junior residents correctly admitted a significantly higher proportion of patients with confirmed AIHD than did interns (86% versus 63%, respectively; p less than 0.06). Specificity was unaffected by the level of house staff training for patients with probabilities of AIHD less than 40% (ns). In the high probability group of patients, the interns' specificity rate (54%) was significantly better than the 20% for junior residents (p less than 0.03); however, when clinical severity was controlled for, no differences by level of training were found.