“I’m an optimist when it comes to electronic health records,” said John W. McEvoy, MB, BCh, a cardiology fellow at the Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Johns Hopkins School of Medicine in Baltimore. “But I can’t think of one doctor who is actually happy with electronic health records.”
And that concerns him because, as he wrote in a recently published article in the American Journal of Medicine, “[c]linical informatics represents arguably the most significant advance in medicine since the deciphering of the human genome. In particular, as the 'front end' of the clinical informatics revolution, the electronic health record has immense potential to transform modern healthcare.”
But, he wrote in the article titled “The Turing Test and a Call to Action to Improve Electronic Health Record Documentation, ” EHR vendors have “mainly focused on the medico-legal, reimbursement, and regulatory requirements of the documents generated by their systems and have placed relatively little emphasis on improving the clinical quality of electronic health record–facilitated documents.”
“They’ve done a poor job for the end user—the physician,” he observed. Specifically, he has problems with the way in which EHRs present information to those end users. He pointed out in his article that many EHRs produce “electronic health record-facilitated documents” by auto-populating information from a patient’s chart into the note.
“However,” he wrote, “physicians often report that electronic health record–facilitated documents commonly have a poor ‘signal-to-noise’ ratio and fail to succinctly communicate the necessary clinical narrative.”
And, according to McEvoy, “the mind is hardwired to the use of a narrative framework” but EHRs end up distorting the narrative by presenting the physician with a lot of superfluous information he or she must wade through.
“The EHR as it currently exists can’t make that narrative happen,” he said. “So what ends up happening is that physicians have to pick out the pieces of the story from a smorgasbord of information, and it’s not easy for them to do it. There’s a disconnect between the information in the system and how they are processing that information. And that makes things a little more difficult from a workflow situation.”
To make matters even more problematic, McEvoy said, it’s possible that because a physician is scanning down through the record looking for the most relevant information—and probably doing so under some time pressure—he or she is going to miss something.
“So I think there is a big concern that this is hurting the patient/doctor relationship,” he added, noting that if mistakes are made it can erode a patient’s trust in the physician. He is also worried that by focusing so much on a computer screen, rather than the patient, physicians could fail to pick up nonverbal cues from their patients during an examination. “Because we’re looking at a computer screen as we type, rather than the patient, we aren’t seeing a lot of things we may have seen before,” McEvoy observed.
McEvoy emphasized that he is not a clinical informaticist (although he wrote the piece for the American Journal of Medicine as a result of taking a course on clinical informatics) and realizes this isn’t an easy problem to solve, but he would like a system that could pass, in his words, the “Turing test.”
Alan Turing, a pioneer in the field of informatics, was highly influential in the development of computer science, and played a pivotal role in breaking German codes during World War II.
“I believe it behooves the electronic health record vendor, clinical informatics, and physician communities to hold electronic health record–facilitated documentation to the standard of Turing: an inability to tell the difference between the output of a person from that of, or facilitated by, a machine,” wrote McEvoy in his article. According to McEvoy, the Turing test would ask that an observer—an interrogator—examine two kinds of notes, those produced by a person, and those from an electronic health record-facilitated document. If the interrogator can’t tell the difference then the note generated electronically passes the test.
“[A]s the users of this technology, physicians are best positioned to provide suggestions and feedback to clinical informatics and electronic health record vendor groups,” McEvoy wrote. “In addition, physicians need to advocate more for electronic health record vendors and other