Decision support nips needless back-pain imaging in the bud

Point-of-care clinical decision support (CDS) has showcased its ability to safely head off unnecessary imaging of patients presenting with low back pain—the poster-child condition of overutilization and defensive medicine—in the busy emergency department of an urban academic medical center.

Adam Min and Bruce Forster, MD, of the University of British Columbia and colleagues describe their success in study findings posted online May 5 in the Journal of the American College of Radiology.

The researchers prospectively looked at lumbar imaging referrals made by 43 emergency physicians at a Vancouver teaching hospital who each saw at least 10 low back pain cases in both pre- and post-intervention periods.

The team compared imaging rates before and implementation of their homegrown CDS—which equipped the department’s electronic order-entry system with a checklist of red flags for potentially serious underlying causes of pain—and measured the potential harms of reduced imaging.

“After introducing CDS,” the authors report, “we observed no significant increase in the percentage of patients who returned to the emergency department or in the percentage of patients discharged without medical imaging who went on to receive imaging at the hospital outpatient imaging clinic for the same presenting complaint.”

Specifically, the CDS led to a 22 percent relative decrease in the median rate of imaging for low back pain.

Meanwhile, one minor thoracic spine compression fracture was missed, but clinical management was not impacted—and not a single serious diagnosis was missed.

The lumbar CDS tool thus “resulted in greater compliance with established professional guidelines, a safe reduction in the use of medical imaging, and improved cost-effectiveness of care,” the authors conclude.  

Min et al. acknowledge several limitations to their study design, including its use of historical controls whereby emergency physicians served as their own controls. Studies employing historical controls are well known to be susceptible to co-intervention bias, they point out.

As for the design of the CDS tool itself, it was created by a working group of emergency physicians, radiologists and family practitioners.

The tool required doctors ordering imaging for low back pain to select from one or more options, such as “suspected compression fracture,” “suspected epidural abscess or hematoma,” “suspected cancer,” “suspected cauda equina syndrome,” “suspected infection” and “severe progressive neurologic deficit.”

If the referrer checked off “other,” he or she had to explain their alternative reason for imaging in their own words.

“This constituted a ‘medium stop’ intervention because we neither denied imaging for orders that did not meet appropriateness criteria (a ‘hard stop’) nor allowed imaging without an explanation (a ‘soft stop’),” the authors explain.