OBJECTIVE: To determine whether ultrasound changes emergency physicians' estimated likelihood of acute ureterolithiasis in patients with flank pain. METHODS: This prospective, observational study enrolled patients awaiting computed tomographic (CT) scan for presumed renal colic. Using a visual analogue scale, treating physicians estimated the likelihood of acute ureterolithiasis based first on clinical findings and urinalysis, then after ultrasound, and finally after CT. A 20% change in estimated likelihood was considered clinically significant. Test characteristics of ultrasound for any ureteral stone and for those greater than or equal to 5 mm in size were determined. RESULTS: One hundred seven patients were enrolled. Sensitivity, specificity, and negative predictive value of ultrasound for stones observed on CT were 76.3% (95% confidence interval [CI], 59.4%-88.0%), 78.3% (95% CI, 66.4%-86.9%), and 85.7% (95% CI, 74.1%-92.9%) respectively, and for stones >5 mm 90.0% (95% CI, 54.1%-99.5%), 63.9% (95% CI, 53.4%-73.2%), and 98.4% (95% CI, 90.3%-99.9%), respectively. Ultrasound significantly impacted the estimated likelihood of disease in 33 of 107 cases (30.8%, 95% CI, 22.5%-40.6%). Computed tomography further significantly changed physicians' impression of disease in 55 of 107 cases (51.4%, 95% CI, 41.6%-61.1%). CONCLUSIONS: Bedside renal ultrasound had only a limited impact on the physicians' clinical impression of patients with possible ureterolithiasis. The sensitivity of sonographic hydronephrosis was modest for detecting any ureteral stone, but much better for detecting a large stone. Further study is needed to define the precise role ultrasound should play in evaluating patients with suspected ureterolithiasis.