By Michael Burger, Senior Consultant
Last week, a headline caught my eye: “Report Finds EHR Contributed to Dallas Hospital's 2014 Ebola Issues.” The report is written by an independent panel convened by Texas Health Resources, which operates Texas Health Presbyterian Hospital Dallas. This is the hospital where patient Thomas Duncan received care, was not diagnosed with Ebola and subsequently passed away at his residence.
The report illustrates the issues involving electronic health record (EHR) usability, including accusations that some EHR vendors have attested in the Meaningful Use certification process to use “user-centered design” or UCD. The events in Texas illuminate the challenge faced by EHR vendors regarding usability.
The report draws several interesting conclusions about what happened. It finds that the care team failed to monitor the patient’s clinical information as documented in the EHR and that the hospital seemed more concerned with patient satisfaction and less about outcomes. Moreover, the report concludes that the hospital was ill prepared to address Ebola by failing to get Ebola treatment information into the right hands quickly.
The comments related to electronic health records (EHRs) are particularly fascinating. The report concludes that there is “over reliance on the EHR to convey critical information,” that the “EHR configuration did not provide for automatic alerts related to travel history” and that “[the EHR had] no systems in place that would trigger a review or re-asking of critical travel information.”
The report further notes that the sharing of travel history was not easily accessible to the physician, as viewing it “required extra and non-intuitive steps to be taken by the physician to access information highly relevant to clinical decision-making.” In the end, the hospital “[did not] optimally utilize the full capability of the EHR.”
It’s not surprising that Mr. Duncan’s death and the specter of Ebola elicited a post-mortem evaluation. But is it right to blame EHRs for what happened in Texas?
We all agree that, used to their potential, EHRs can be a tremendous diagnostic tool, capable of detailed analysis of vast amounts of data in seconds. The challenge is how to make the analytical results useful in a life-or-death situation. EHR designers observe clinical workflows and build software to accommodate them. Smooth and efficient clinical workflow does not allow for “automatic alerts related to travel history” and certainly not “a review or re-asking of critical travel information.” Having observed clinical workflow more times than I can count, I can say for sure that nothing irks a doctor more than being interrupted by alerts. Alert fatigue is a well-known phenomenon and barrier to EHR user satisfaction. I can only imagine a physician’s reaction if the EHR began prompting to re-ask questions!
And what is the definition of smooth workflow? We have an expression in EHR product management: ask a group of 5 doctors for their opinion, and you’ll get 6 opinions. Consensus is difficult among physicians, particularly when it comes to EHR workflow. It’s not a surprise then that the report concludes that the EHR “required extra and non-intuitive steps” to access relevant clinical information. What is relevant to one physician is irrelevant and non-intuitive to another. Plus, it varies across specialties and work environments, such as this particular hospital. This means that EHR vendors must seek the middle of the road when it comes to workflow - to provide what is most relevant to most physicians. They also must trust that the physicians are relying on their medical training, not the EHR, to make their clinical decisions. Physicians wouldn’t have it any other way; they want to use EHRs as a resource, not a replacement for their training and medical judgment.
In a nod to our legal team’s advice at my former employer, we had a disclaimer that would display as users logged into the EHR. It contained a paragraph of “legalese,” which I’m sure that everyone clicked through without ever reading. Buried in it were important words: “This software contains logic including warnings and reminders meant to be used as reference material. It is not designed and cannot be a substitute for the clinical decision making of a licensed healthcare professional.” While it’s tempting to blame the EHR for lapses in care quality, in the end it comes down to the caregivers. The best caregivers refer to the EHR as a reference, but rely on their diagnostic skills. Sometimes obscure diagnoses will be missed – to err is human.