Perhaps, like me, you're old enough to remember when doctors kept paper records to document your health and medical care (variously called a medical record, file or chart). I certainly remember those days... along with floppy disks, tape decks and rotary phones. And, while I always assumed that there might be idiosyncrasies in how each doctor or medical practice documented my health care, I sort of figured that the patient charts from different medical practices were more similar than they were different. Once paper records became less common and electronic health record (EHR) systems became the norm that assumption just transferred over to how I thought about EHR systems.

This is surprising to me for a few reasons. I see different doctors (who are part of different medical practices altogether), but they use an EHR from the same company. I know this because the patient portal that I access is the same for both, but I have to make sure that I'm logging into the right version of that EHR's patient portal or I'm communicating with the wrong office about medications and appointments.

But that's not all. As a research assistant, I conducted chart audits at clinics from different Community Health Centers (CHCs) in North Carolina. Some of these clinics, associated with different CHC systems, used the same EHR but implemented them very differently. To do my job efficiently, I had to remember those differences — that even though the EHR was familiar I had to navigate each a little differently and look in different places to find the information that I needed for the chart audit. So I really have no excuse for not thinking about the situation previously.

At this point you've either stopped reading altogether or you're really wondering, 'what difference do those differences make anyway?' Well, in a study recently published in the Journal of the American Medical Informatics Association titled "A usability and safety analysis of electronic health records: a multi-center study," the authors characterized the variability of EHRs from two different companies as implemented in four health systems. And through that process, they've highlighted why those differences are very important indeed.

This research was conducted at four different health systems, one of which was UNC Health Care. Saif Khairat, PhD, MPH, an Assistant Professor in the School of Nursing and a member of the North Carolina Translational and Clinical Sciences (NC TraCS) Institute at the University of North Carolina at Chapel Hill led the team conducting research here at Carolina.

For this study, 12-15 emergency department physicians from the four health systems completed common tasks in the EHR that modeled real patient care. The health systems used EHRs from the two largest vendors — two systems use EPIC and two use Cerner. The physicians ordered lab tests, imaging studies and medications in these clinical scenarios. The researchers captured both how long it took the physicians to complete a task in the EHR, as well as how many mouse clicks were required to complete that task. Finally, they assessed accuracy of the completed tasks.

I'm pretty confident that we all want our health care providers to complete tasks in the EHR with a high-degree of accuracy because that has a direct effect on the quality of the care we receive. This is the safety analysis referenced in the title of the publication.

What the researchers found was wide variability in the time it took to complete a task, the number of mouse clicks required to complete a task, and in the accuracy of the completed task... even in health systems using an EHR from the same company.

Just to give you an idea of the findings, the researchers noted that the time it took to complete an MRI order (a type of imaging test) varied from 32 seconds at one site to well over a minute at the other sites. Placing that MRI order took an average of about 14 mouse clicks at one site to more than 30 clicks at a different site.

When ordering Tylenol (a medication order) the accuracy at the sites varied from no errors at to one site to 30% error rate at another site. And that's just for what I'd consider a fairly straight-forward medication order.

"We saw error rates reach 50% while observing providers as they ordered Prednisone taper [a medication that is reduced by a set amount over 12 days] in the EHR, which demonstrates a serious risk to patient safety," stated Khairat in an email exchange about the publication.

These results certainly bring home to me that customization and implementation of the products introduces variability in usability, and most importantly, in patient safety. This happens even with products that have been certified through the Office of the National Coordinator of Health Information Technology, which is a part of the U.S. Department of Health and Human Services. Khairat emphasized that "vendors also have the responsibility of optimizing the interface design and functionality of their products to help users successfully complete tasks in the EHR."

Bottom line — The way that EHRs are customized and implemented introduces differences in those EHRs from the version that was certified by the government. These differences can result in meaningful variations in both usability and patient safety. Current vendor certification processes do not address these issues as the certification occurs prior to implementation of an EHR in a particular medical practice or health system. For the promise of EHRs to be fully realized, physicians and other users need to be able to do what they need to do in the EHR quickly, easily, and with low error rates.

Learn More

Get all the details on what they did and what they found in the paper - https://www.ncbi.nlm.nih.gov/pubmed/29982549.
AMA press release - https://www.ama-assn.org/study-finds-need-performance-standards-ehr-usability-safety