Overuse of ER imaging is a problem, but subspecialization is not the answer

Although what Dr. Eric Trefelner says in his "Perturbations" column ("ER practice may save us from neutering," March 2005, page 64) is correct, I think he is off the mark.

At my primary hospital, similar to what Dr. Trefelner discusses, night emergency interpretation for specialty CT, MR, and ultrasound is supplied by rotating subspecialists. As I am sure is the case throughout the country, our private-practice group does not have subspecialty trained "emergency radiologists." He suggests this is suboptimal and is leading to "wide variability in care, increased errors, and additional studies/costs."

Our discrepancy rate-the rate by which final interpretations differ significantly from the preliminary interpretation-is about 1.3%. Even if specialty trained emergency radiologists have a lower error rate, I do not believe that eliminating such a small error rate would significantly improve care and reduce overall imaging costs in our emergency department. With our system of quick case review, most of the errors usually lead to a small delay in diagnosis but rarely lead to additional studies.

What Dr. Trefelner does not mention is the very low rate of positive studies for ER patients. (At our hospital, ER and private physicians order studies on patients in the emergency room.) A recent review of 1600 studies ordered from our emergency room between 5 p.m. and 8 a.m. over a 10-week period revealed pulmonary embolism in just 6.25% of pulmonary angiograms, acute findings in just 7.1% of cervical spine CTs, and acute findings in just 6.7% of pediatric emergency room head CTs. None of 27 patients who had combined chest, abdomen, and pelvis CT scans had acute surgical abnormalities. The conclusion is obvious: There is an extremely low percentage of acute findings on studies ordered from the ER. The problem is not on the interpretation side but on the ordering side.

The literature is replete with similar examples of low positive rates on imaging studies. I have perused the Internet, including the Web sites of the American Society of Emergency Radiology and American College of Emergency Physicians, looking for suggested positive rates on imaging ordered by ER physicians and have found none.

It would be unacceptable for any radiologist in my group to recommend additional imaging that would be negative 90% of the time. Yet, after an initial ER evaluation, ER physicians are sending patients for radiological evaluation where 90% have no significant abnormality identified.

The reason for the high ordering of studies is most likely multifactorial. Dr. Trefelner talks about lack of subspecialization in radiology, when, in fact, a lack of expertise by ER physicians and nurse practitioners accounts for some of the overordering. The subject of liability always arises during discussions of utilization with our ER physicians. Staffing and limited time may also contribute to the ER's reliance on imaging. I do not think radiological interpretation plays a significant role.

I would like to see the ASER, ACER, or American College of Radiology develop imaging algorithms for emergency room patients. These could serve as acceptable guidelines and hence act as reasonable standard of care. These would protect the emergency physician from liability, protect patients from unnecessary imaging, and maximize resource utilization. I would also like to see guidelines for measuring the utilization of the more expensive and resource-hungry examinations. Guidelines that set targets for the number of studies with positive results could serve as goals for ER physicians.

-Joel M. Schwartz, M.D.,

New City, NY