Valid protocols contribute to good ER imaging decisions

Running the numbers on clinical- and cost-effectiveness may appear to drain the excitement out of emergency room imaging. But simple, valid protocols that define how to apply diagnostic imaging to trauma cases are essential to obtaining the biggest clinical benefit from expensive CT technology, Dr. C. Craige Blackmore and Dr. M.G. Myriam Hunink said in their opening-session lectures on Sunday.

Well-defined protocols help the radiologist act intelligently in emergency situations when there is no time to methodically weigh clinical options, said Blackmore, an associate professor of radiology at the University of Washington in Seattle.

Given the rising economic costs of imaging, the conditions under which imaging should be prescribed are bound to be analyzed and calculated. Blackmore advised radiologists to get into act themselves to determine how, why, and by whom trauma imaging should be performed.

"Radiologists need to be in control of emergency imaging, using evidence-based methods based on research performed by radiologists," he said.

The temptation to overuse imaging in the ER arises from the basic orientation of physicians who practice trauma medicine, said Hunink, a professor of clinical epidemiology at University Medical Center in Rotterdam, the Netherlands.

"The ER is all about looking for injury, so the physician's first inclination is to order imaging to search for it," she said.

Clinical protocols require a balancing act between social costs and clinical benefits. The societal costs are rising quickly, Hunink said. From 1985 to 2002, the percentage of U.S. gross domestic product allotted to healthcare rose from 10% to 14%. About 41 million people are uninsured, and enormous regional variations in healthcare utilization suggest endemic inefficiencies in the system, she said.

Clinical protocols balance the value of diagnostic imaging against the cost of gathering radiological data. Decision modeling, a technique used in the mid-1990s to determine whether CT is a legitimate alternative to x-ray angiography for diagnosing aortic ruptures, requires the researcher to make methodology assumptions that sometimes jeopardize the legitimacy of this type of study. Exhaustive decision modeling can fail to account for new technologies, such a multidetector CT for emergency exams of the aorta, that are introduced before the results of evidence-based research can be placed into practice.

Developing new protocols can be a relatively straightforward three-step process involving the determination of who should receive a given ER imaging exam, how appropriate the exam is in a specific situation, and why radiologists should use the best available evidence to govern their use of imaging.

These rules have been applied in the development of various ER imaging use protocols. The Ottawa ankle rule, for example, recommends the use of CT for ER patient with ankle injuries with positive physical findings of lateral malleolar tenderness. This rule, the Canadian head CT rule, the NEXUS C-spine rules, and the escalating pelvic hemorrhage rule are examples of protocols that can lead to fast, accurate decision-making about the use of ER imaging.

An imaging recommendation under the Ottawa knee rule depends on whether the patient can flex the knee and bear weight on it, and whether there is tenderness over the head of the fibula or the inferior margin of the patella. If those criteria are absent, imaging is not indicated.

The Canadian head CT rule considers age, mechanism of injury, signs of amnesia, signs of open or basilar skull fracture, and other symptoms. Patients who lack these criteria do not require a CT of the head.

According to the NEXUS prediction rule for cervical spine imaging, radiologists should look for midline tenderness and neurological deficits. If patients lack these two criteria, can be examined, are not intoxicated, are awake, and do not have a head injury, they don't need a cervical spine exam.

Blackmore's own University of Washington-Harborview cervical spine CT criteria follow the same principles.

The protocol calls for a cervical CT study in the presence of the following symptoms:

• focal neurological deficit
• severe head injury
• high-velocity mechanism of injury, such as a high-speed motor vehicle crash.

The basis for this protocol is the fact that people involved in high-impact motor vehicle accidents have a four-fold greater risk of cervical spine injury than other accident patients. Blackmore tested the underlying assumptions of the protocol on 4285 ER patients and found that 12.8% of those who fulfilled these criteria and underwent CT were found to have C-spine fractures.

Through additional calculations, Blackmore concluded that CT is cost-effective in this situation when the risk of fracture is greater than 4%.

"The sensitivity of these rules is 100%, and that is why we can use them to replace imaging. The specificity of these rules tells us how useful they may be, because the specificity tells us how many radiographs we may avoid by using such a tool," he said.