A preliminary evaluation of emergency ultrasound in the setting of an emergency medicine training program

In this article we seek to evaluate the diagnostic accuracy of emergency physicians performing emergency ultrasonography in the setting of an emergency medicine training program. A prospective observational study was performed at an inner city Level I trauma center with an emergency medicine residency training program. From July 1994 to December 1996 a convenience sample of ultrasound exams was recorded. The diagnostic quality ("acceptable or technically limited") was determined by a board-certified cardiologist or radiologist with fellowship training in ultrasonography. The emergency department interpretations were then compared to those of the blinded cardiologist or radiologist. Four hundred and fifty-six ultrasound examinations were videotaped and entered into the study; 408 (89%) of the studies performed were determined to be "acceptable." The diagnostic accuracy (sensitivity, specificity, positive and negative predictive values) of these studies were as follows: cardiac, to rule out effusion (n = 67; 0.83, 0.98, 0.88, 0.98); transabdominal, to rule out abdominal aortic aneurysms (AAA), cholelithiasis, or free peritoneal fluid (n = 263; 0.91, 0.89, 0.88, 0.92); renal, to rule out hydronephrosis (n = 45; 0.94, 0.96, 0.94, 0.96); pelvic, to rule in intrauterine pregnancy (n = 33; 1.0, 0.90, 0.96, 1.0). The 48 "technically limited studies" included: 39 transabdominal (33 gallbladder, 1 abdominal aortic aneurysm, 5 free peritoneal fluid), 6 cardiac, 2 renal, and 1 pelvic ultrasound. This study suggests that emergency physicians with a minimal amount of training display acceptable technical skill and interpretive acumen in their approach to emergency ultrasonography.