While the failure of a Texas hospital to diagnose Ebola was almost certainly caused by human error, it also serves as a wake-up call concerning the usability of EHRs, according to a report published in the online journal Diagnosis.
The report, “Ebola U.S. Patient Zero: Lessons on Misdiagnosis and Effective Use of Electronic Health Records” was written by Dean Sittig, PhD, professor, School of Biomedical Informatics in University of Texas Health Science Center, Houston; Hardeep Singh, MD, chief of the Health Policy, Quality & Informatics program at the Houston Veterans Affairs Center for Innovations in Quality, Effectiveness and Safety; and Divvy Upadhyay, MBBS, MPH, a research associate at the Urban Institute’s Health Policy Center in Washington, D.C.
In the paper, the authors detail the case history of Thomas Michael Duncan, who presented at the emergency room of Texas Health Presbyterian Hospital in Dallas on Sept. 25 with symptoms consistent with Ebola. Despite the fact that the nurse’s notes documented in the hospital’s EMR indicated that he had traveled from Liberia, “it’s clear that this information was not appreciated or acted upon, and he was discharged." Two days later, after his condition worsened he returned to the hospital, was admitted and diagnosed with Ebola. He later died of the disease.
On Oct. 1, the hospital announced that the initial ED nurse failed to communicate the travel history to physicians. The next day the misdiagnosis was blamed on a technical flaw in the EHR--the misalignment of the physicians’ and nurses’ workflows--preventing physicians from seeing the travel history in the nurse’s notes. Within 24 hours the hospital reversed its position and announced that the travel history was indeed documented and available to the full care team in the EHR within the physician’s workflow, and there was no flaw in the EHR.
“Assigning blame to the EHR is not new and often reflects a reluctance to address the complex cognitive and/or performance issues involving front line staff, especially those related to responsibility and accountability,” Sittig and Singh wrote. “It is important that we recognize the reality that EHRs suffer from major usability and inter-operability issues, but also to acknowledge that they are only tools and not a replacement for basic history-taking, examination skills and critical thinking.”
Based on the record in this case, said Singh, “We think there was a failure of basic data gathering and basic data interpretation and synthesis due to many issues.”
For example, wrote the authors, in this case the nurse was using a template they said was designed for the administration of influenza vaccine rather than a template more appropriate for documenting a potential Ebola case. In addition, it appears the emergency room doctor either didn’t read a nurse’s note containing the patient’s travel history, or somehow didn’t know how to access those notes, said Sittig, speaking to CMIO.net.
One lesson learned, said Sittig, is that “the computer is kind of getting in the way of people thinking.”
“In this case it seems they were using the vaccination template and it’s possible that this, combined with everything else like the fact that it was late at night, in a busy ER, was just enough to make someone forget about the key thing that could have given them that once-in-a-lifetime opportunity to make that diagnosis of Ebola, and they missed it,” Sittig said. “That’s one thing we want to be particularly careful about. If you’re spending too much time in front of a computer it’s going to be easy to miss those key findings or symptoms, or clues that are right in front of you.”
Another lesson learned, said Singh, is that “we need to improve usability—we’re basically spending way too much time clicking on screens and picking from lists.”
Consequently, he said, computers are impeding clinicians rather than helping them, and added that they have to be able to provide more cognitive support.
Sittig said he takes a broader, socio-technical view about the issues associated with EHR usability. “A lot of people focus on what I call ‘low-level’ usability, where they complain about how many clicks something takes, or how many screens they have to go through, and while I agree that can be a problem, there is sort of a higher level of usability.”
For example, Sittig pointed out that in the Ebola case in Texas, the medical personnel in the ER were using a “good screen” in their EHR, which happened to be the