Emery DJ, Shojania KG, Forster AJ, Mojaverian N, Feasby TE. Overuse of magnetic resonance imaging. JAMA Intern Med. Published online March 25, 2013. doi:10.1001/jamainternmed.2013.3804.
eTable. Indications for magnetic resonance imaging and specialty of referring physicians
Customize your JAMA Network experience by selecting one or more topics from the list below.
Emery DJ, Shojania KG, Forster AJ, Mojaverian N, Feasby TE. Overuse of Magnetic Resonance Imaging. JAMA Intern Med. 2013;173(9):823–825. doi:10.1001/jamainternmed.2013.3804
Author Affiliations: Department of Radiology and Diagnostic Imaging, University of Alberta, Edmonton, Alberta, Canada (Dr Emery); Department of Medicine, University of Toronto, Toronto, Ontario, Canada (Dr Shojania); Department of Medicine, University of Ottawa, Ottawa, Ontario (Dr Forster); Ottawa Hospital Research Institute, Ottawa (Drs Forster and Mojaverian); and Departments of Clinical Neurosciences and Community Health Sciences, Faculty of Medicine, University of Calgary, Calgary, Alberta (Dr Feasby).
Overuse of health care services such as magnetic resonance imaging (MRI) has become an increasingly recognized problem.1,2 We studied the appropriateness of requests for outpatient MRI of the lumbar spine and of the head for headache, as these are common indications and might be frequently inappropriate.
We used the RAND–University of California, Los Angeles, appropriateness method to define appropriate care, combining best evidence and expert opinion,3,4 to prospectively determine the appropriateness of requests for MRI studies of the lumbar spine and of the head for headache at the University of Alberta Hospital (UAH) in Edmonton, Alberta, Canada, and The Ottawa Hospital (TOH) in Ottawa, Ontario, Canada. As part of the process 2 expert panels, 1 each for headache and lumbar spine, were created, comprising specialists nominated by various Canadian specialty societies. The literature review and complete listing of scenarios were reviewed by the panels and then revised. The panelists individually rated each scenario's before and after group discussion on a 9-point scale, where 1 to 3 indicates an inappropriate indication for MRI; 4 to 6, uncertain indication; and 7 to 9, appropriate indication. For the final scenarios, we used the median rating for each scenario to determine the appropriateness of the intervention.
We prospectively identified outpatient requisitions for MRI scans at the UAH from May 2008 to September 2009 and TOH from September 2008 to March 2010. We collected data from 500 lumbar spine requisitions and 500 head for headache requisitions at each site. Each case was matched to a clinical scenario for which the appropriateness rating had been determined.
The specialty of the referring physicians and indications for the studies can be found in the eTable.
Only 443 of 1000 requests were considered appropriate. The remainder were split between inappropriate (285 of 1000 [28.5%]) or of uncertain value (272 of 1000 [27.2%]) (Table).
Only 1 indication, investigation of postoperative leg or back pain, was associated with an almost certain rating of appropriate (160 of 167 [95.8%]), but it accounted for only 16.7% of all MRI orders for the low back. For all other indications, accounting for 83.3% of tests, the probability of an uncertain or inappropriate rating was 3 times greater than the probability of an appropriate rating.
Family physicians had a lower rate of appropriate MRI ordering for the low back than other specialties. Only 33.9% of their MRI scans were considered appropriate vs 58.1% of those ordered by other specialties. Within the subspecialties, MRI scans ordered by neurologists and orthopedic surgeons were appropriate in less than half of cases, while those ordered by neurosurgeons were appropriate 75.7% of the time.
Most MRI scans requested for headache (82.8%) were appropriate (Table). The rest were evenly split between tests requested for an inappropriate indication (9.0%) or tests of uncertain value (8.2%). All physician specialties had similar rates of inappropriate ordering.
We found evidence of substantial overuse of lumbar spine MRI scans. Over half the requests (55.7%) were either inappropriate (28.5%) or of uncertain value (27.2%). In contrast, most (82.8%) requests for head MRI scans for headache were appropriate. The proportion of appropriate requests varied between specialties for lumbar spine MRIs but not for headache MRI scans. Family physicians were more likely to order inappropriate lumbar spine MRI than other physicians.
Since rates of appropriateness vary with the indication and part of the body being imaged, we cannot assume that there are universally high rates of inappropriateness. Each indication must be examined individually.
The high rate of inappropriate and uncertain MRI ordering for the lumbar spine is not surprising. Lumbar spine MRI numbers have increased dramatically in recent years,5,6 but the correlation between lumbar spine MRI findings and clinical signs and symptoms is poor.7
Our study has implications for management of MRI usage for the lumbar spine, which makes up approximately a third of all MRI in some regions. Eliminating inappropriate scans and some of uncertain value could reduce the harm that accrues from unneeded investigations and result in significant cost savings.6
We were unable to measure underuse of MRI, as inappropriateness measures only overuse. Underuse can be an important health quality problem. Furthermore, this was not a population-based study, and only 2 large hospitals were involved, so extrapolation to other health systems should be done cautiously. Magnetic resonance imaging is used in conjunction with other imaging modalities, such as computed tomography (CT). Most of the patients in our study referred for MRI of the head for headache had already been prescreened with a CT scan, likely explaining the high rate of appropriate head MRIs. Many CT scans performed for headache may be inappropriate; however, this was not assessed. For many of the lumbar spine MRI requests (63.7%), we were unable to determine appropriateness on the basis of the information originally provided on the requisition. This has implications to study validity insofar as we had to seek out information from other sources.
Overuse of medical interventions, such as MRI, is a considerable problem, leading to excess costs and adverse outcomes. Overuse is driven by many factors, including patient expectations, physician concerns about litigation, and lack of physician accountability for cost. Solutions will require strict adherence to appropriate guidelines and better education of patients. The efforts of the Choosing Wisely2 consortium to mobilize medical societies to show leadership in reducing overuse is a positive step to this end.
Correspondence: Dr Emery, Department of Radiology and Diagnostic Imaging, University of Alberta, 8440–112 St, Edmonton, AB T6G 2B7, Canada (firstname.lastname@example.org).
Published Online: March 25, 2013. doi:10.1001/jamainternmed.2013.3804
Author Contributions: Drs Emery and Forster had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Emery, Shojania, Forster, and Feasby. Acquisition of data: Emery and Mojaverian. Analysis and interpretation of data: Emery, Shojania, Forster, and Feasby. Drafting of the manuscript: Emery, Shojania, Mojaverian, and Feasby. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: Forster. Obtained funding: Emery, Shojania, Forster, and Feasby. Administrative, technical, and material support: Emery and Mojaverian. Study supervision: Emery, Forster, and Feasby.
Conflict of Interest Disclosures: None reported.
Funding/Support: This study was funded by a grant from the Canadian Institutes of Health Research.
Additional Contributions: The panel members were Robert S. Burnham, MD, Werner J. Becker, MD, Michael Rasminsky, MD, Chaim Bell, MD, John You, MD, Maziar Badii, MD, Ernest Schuster, MD, Donald Lee, MD, Neil Berrington, MD, William Maloney, MD, Raja Rampersaud, MD, and Teresa Cavett, MD. Jaime Laliberte, BSc, Michelle Tubman, MD, and Kelly Wiens, MSc, assisted with the project. Eugene Wai, MD, assisted with the design of the study and took part in the panel discussion. Santanu Chakraborty, MD, and Marlise Peruzzo dos Santos, MD, MSc, provided feedback on the ethics application, data collection forms, and scripts for interviewing patients at 1 of the 2 study sites.
This article was corrected for errors on July. 16, 2013.