Ultrasound is the imaging study of choice for the detection of gallstones, but ultrasound through medical imaging departments (MI Sono) is not readily available on an immediate basis in many emergency departments (EDs). Several studies have shown that emergency physicians can perform ultrasound themselves (ED Sono) to rule out gallstones with acceptable accuracy after relatively brief training periods, but there have been no studies to date specifically addressing the effect of ED Sono of the gallbladder on quality and cost-effectiveness in the ED. In this study, we investigated measures of quality and cost-effectiveness in evaluating patients with suspected symptomatic cholelithiasis during three different years with distinctly different approaches to ultrasound availability. The study retrospectively identified a total of 418 patients who were admitted for cholecystectomy or for a complication of cholelithiasis within 6 months of an ED visit for possible biliary colic. The percentage of patients who had gallstones documented at the first ED visit improved from 28% in 1993, when there was limited availability of ultrasound through the Medical Imaging Department (MI Sono), to 56% in 1995, when MI Sono was readily available, to 70% in 1997, when both MI Sono and ED Sono were readily available (P <.001). There were also significant differences over the 3 years in the mean number of days from the first ED visit to documentation of gallstones (19.7 in 1993, 10.7 in 1995, 7.4 in 1997, P <.001); the mean number of return visits for possible biliary colic before documentation of gallstones (1.67 in 1993, 1.24 in 1995, and 1.25 in 1997, P <.001); and the incidence of complications of cholelithiasis in the interval between the first ED visit for possible biliary colic and the date of documentation of cholelithiasis (6.8% in 1993, 5.9% in 1995, 1.5% in 1997, P =.049). The number of MI Sonos ordered by emergency physicians per case of symptomatic cholelithiasis identified increased from 1.7 in 1993 to 2.5 in 1995 and dropped back to 1.7 in 1997, when 4.2 ED Sonos per study case were also done. The cost of ED Sonos was more than offset by savings in avoiding calling in ultrasound technicians after regular Medical Imaging Department hours. The indeterminate rate for ED Sonos was 18%. Excluding indeterminates, the sensitivity of ED Sono for detection of gallstones was 88.6% (95% CI 83.1-92.8%), the specificity 98.2% (95% CI 96.0-99.3%), and the accuracy 94.8% (95% CI 92.5-96.5%). We conclude that greater availability of MI Sono in the ED was associated with improved quality in the evaluation of patients with suspected symptomatic cholelithiasis but also with increased ultrasound costs. The availability of ED Sono in addition to readily available MI Sono was associated with further improved quality and decreased costs. The indeterminate rate for ED Sono was relatively high, but excluding indeterminates, the accuracy of ED Sono was comparable with published reports of MI Sono.