A prospective study of ultrasonography in the ED by emergency physicians.

The objective of this study was to examine the use of limited, goal-directed, two-dimensional ultrasound studies performed by emergency physicians and to assess the frequency, variety, and accuracy of their readings. A 1-year prospective study was performed by using an emergency department (ED) ultrasound machine with a 3.5-mHz mechanical oscillating sector transducer and a 5.0-mHz vaginal transducer. In a series of proctoring sessions, radiologists trained emergency physicians to do limited, goal-directed ultrasonography. Laser print ultrasonograms were collected from all ED ultrasound examinations performed during a 1-year period and were compared with either formal ultrasonograms performed in the radiology department, the patient's hospital record, or both. Sensitivity, specificity, and positive predictive value (PPV), as well as negative predictive values (NPV), were calculated. The setting was a 104-bed community hospital with an ED volume of 25,000 patients annually, and patients whom the emergency physician believed needed ultrasound studies in the ED were entered. ED ultrasonography was performed in 167 patients by 14 physicians during a 1-year period. For 132 patients who completed formal follow-up, the overall diagnostic accuracy of interpretations of ED ultrasonograms yielded a sensitivity of .95, specificity of .98, PPV of .99, and NPV of .89. Eleven categories of ultrasound use were reported. The three studies most commonly performed were for gallbladder disease (53%), intrauterine pregnancy (28%), and abdominal aortic aneurysms (7%). Accuracy of ED gallbladder ultrasonograms for 65 patients showed a sensitivity of .86, specificity of .97, PPV of .97, and NPV of .85.(ABSTRACT TRUNCATED AT 250 WORDS)