Fractures and liver/kidney lesions top list of residents’ ER misses

A research team from Iowa has found that bone fractures, hypo- or hyperdense lesions in the liver and kidney, pulmonary nodules, and gastrointestinal wall thickening or fat stranding are among the most common misses by on-call residents in the emergency room.

Awareness of missed findings by residents is of widespread interest, said lead author Dr. Dean McNaughton, a radiologist from the University of Iowa, in a scientific poster at RSNA 2007.

"To avoid future misses, residents need to know what they are missing. Attending physicians want to know which areas deserve a second glance, while program directors want to know how to adapt the curricula for future learning," he said.

The researchers used a customized search engine to quickly find discordant resident-attending preliminary reports. They sought to identify the most commonly missed findings from the automated search results.

The study focused on 33,000 CT, MRI, ultrasound, and nuclear medicine examinations carried out in the emergency room over a 12-month period during 2006-2007. The authors analyzed matches generated by the automated search for the keyword "not include" contained in the "report disagree" template. The first 500 matches were included in the study, and 79 duplicates, 134 false matches, and 40 "possible" misses were then removed. This left 247 matches, amounting to 336 misses.

A missed finding was associated with one or more other abnormal findings in 185 of the 336 misses (55%). Fractures - mostly involving facial bones, transverse processes, or ribs - accounted for 59 misses (18%). Other common misses were hypo- or hyperdense lesions (26), usually in the liver or kidney; pulmonary nodules (20); and gastrointestinal wall thickening or fat stranding (15).

"Since the first type of missed finding does not represent a specific pathology, it overlaps with the other groups," McNaughton said.

The authors admit their study has limitations. Misses did not include projection radiography exams, and there may be variability in the "report disagree" template usage among attending physicians. Also, it was assumed that the attendings were always correct, and no attempt was made to assess the impact of the misses on patient management.