How EHR 'choice architecture' for imaging could be wasting time and money
When choosing and implementing an electronic health record system, it is important to consider how the system’s architecture might affect providers’ decision-making.
Those decisions—which labs and imaging exams to order, etc.—can have a trickle-down effect on healthcare waste, particularly within the realm of medical imaging. An example of this was shared recently in the Journal of Informatics in Health and Biomedicine, where experts provided two examples of how changing an EHR’s “choice architecture” can reduce medical waste.
Corresponding author of the new paper Raman R Khanna, of the Department of Medicine at the University of California San Francisco, and colleagues explained that when it comes to ordering an imaging exam, the ordering provider is not necessarily at fault for filing a requisition incorrectly, but that the design of the EHR itself could be to blame. The authors described this design as the ordering provider’s “choice architecture,” which “nudges” users toward a specific selection—sometimes an inappropriate exam.
The researchers evaluated how three different choice architectures could be nudging providers to make incorrect selections, but only two were relevant to imaging. In both imaging scenarios, the changes in architecture produced positive results, however, one was more costly.
First, the team addressed the inadvertent ordering of incorrect CT exams. Specifically, they noted that providers will sometimes order a CT abdomen when they are actually trying to request a CT of the abdomen and pelvis—a completely different requisition.
In the EHR, they changed the order entry for “CT abdomen/pelvis” to “CT abdomen /pelvis.” By adding an extra space before the slash, “CT abdomen /pelvis” was bumped higher on the list of exams providers could choose from, which “nudged” them to select it. This saves time (and radiation exposure) by having the entire abdomen and pelvis viewable in a single exam, rather than having just the abdomen available for interpretation, which in most cases would result in the patient needing to return to the CT department for additional imaging.
The team also addressed the issue of ordering prescription of benzodiazepines for patients’ anxiety prior to MRI scans. Prior to the adjustment, the quantity of benzodiazepines providers were presented with was the same quantity given to patients who routinely take the medication for chronic conditions. This resulted in providers being “nudged” to prescribe patients much more medication than necessary when they only needed one or two pills.
To address this, the team added a new order labeled as “Lorazepam (Ativan) tablet 0.5 mg for imaging/procedure” that would appear under an ambulatory orders section when providers searched for lorazepam. The quantity of the order defaults to two tablets with zero refills, with directions stating the medication was to be taken as indicated “for anxiety (prior to imaging study or procedure).”
In both cases, the changes in choice architecture were successful in steering providers in the correct direction. However, the authors noted that the costs of implementing such changes versus the time saved need to be considered. For the CT changes, the total institutional implementation time was estimated at three hours, but for MRI the changes took 16 hours.
To learn more, click here.