How accurate are emergency clinicians at interpreting noncontrast computed tomography for suspected renal colic?

OBJECTIVE: To assess the accuracy of emergency physicians and trainees in the interpretation of noncontrast helical computed tomography (NCHCT) for suspected renal colic by examining the interrater reliability between emergency department (ED) clinicians and radiologists.
METHODS: Information was collected prospectively on all ED patients undergoing NCHCT for suspected renal colic over a 12-month period. Emergency physicians and trainees were asked to report the absence or presence of specific renal parameters (renal tract abnormality, calculus, hydroureter, hydronephrosis, perinephric stranding, and renal parenchymal edema) and nonrenal parameters (nonrenal abnormality, free gas, free fluid, and aortic diameter >3 cm). These reports were compared with the formal radiology report, which was used as the reference standard. The sensitivity, specificity, accuracy, and kappa coefficient were calculated for each of the parameters.
RESULTS: Over the 12-month period, 212 patients underwent NCHCT for suspected renal colic, of whom 127 had both ED and formal radiological reporting. There was an excellent degree of interrater reliability between the ED clinicians and the radiologists (kappa > 0.75) for the presence of renal tract abnormality and renal tract calculus. There was intermediate interrater reliability (kappa 0.4-0.75) for nonrenal tract abnormalities, hydroureter, hydronephrosis, and perinephric stranding. Four patients had potentially significant nonrenal abnormalities missed by ED clinicians.
CONCLUSIONS: Emergency clinicians are able to identify renal calculi with a high degree of accuracy but may miss important nonrenal abnormalities. Therefore, all patients without evidence of renal tract calculus on NCHCT must have early and appropriate follow-up.