Earlier this month, I had the privilege (honor? Thrill?) to attend the in-person annual physician IT leadership meeting organized by AMDIS. If you’re not aware of the group, the Association of Medical Directors of Information Services has traditionally been composed of CMIOs and now CHIOs, with a sprinkle of physicians who are CIOs thrown it. You can tell the organization has been around a while because it predates the existence of the CMIO title. (We were all medical directors back in the day, it seems.)
This is a great meeting because it’s relatively small, informal, and focuses on bread-and-butter healthcare IT issues. Hot topics this year included COVID-19 (not surprisingly), cybersecurity (again, an obvious choice), and telehealth (sure, fine, you saw that one coming as well). Most healthcare systems represented at AMDIS use one of the larger electronic health records (EHRs), but presentations are primarily vendor-agnostic.
One EHR company did send some senior folks, and they presented on the last day of the conference to their clients only (intellectual property and all). There were only a few dozen physicians still able to be there at this point, so it was an intimate presentation. This allowed for a lot of frank conversation regarding how the technology works for clinicians, how the software could be better, and what should be prioritized for future development.
One comment from the physician peanut gallery stopped me in my tracks. The physician was describing how most physicians still struggle to find information in the EHR. We all agreed that the data we’re seeking are in there, but where? And while it is difficult to find information that we know is in the chart, an even bigger conundrum is information of which we’re not aware. There’s no way to know what we don’t know.
Here’s the comment that I thought was great. The physician described the process of trying to navigate the EHR as “spelunking through the chart.” Yes! Finding information in electronic health records is like exploring a cave. It’s dark in there, so often you bump into walls that you can’t see. You don’t know how deep the cave is, and while it may extend miles, it may just as easily end after 20 feet. You may walk and walk toward your goal, not realizing that what you sought was only a few feet away in a different direction. Caves may have bats, and you can get rabies if a rabid bat touches you. OK, maybe the last analogy wasn’t quite on the money; I’m not aware of anyone contracting infectious diseases via EHR use, but I’m sure readers will inform me if I’m wrong about that.
As the good doctor pointed out, we need tech companies to help clinicians find data and knowledge that we know is somewhere in the chart, as well as information that may be relevant but of which we are unaware. Computers need to help us humans elevate data that are likely to be related to the patient and the issues at hand. Finding related information isn’t that difficult given the tools we have now. It’s the second part of the problem that’s thornier: the information must not only be related, but more essentially, it must be relevant. Aye, there’s the rub.
Electronic health record developers and medical students both struggle with the concept of relevance in healthcare. Student doctors are taught to present a patient at rounds by saying things like, “The patient has a past medical history significant for ___” and “Significant lab results include ___.” News flash: EHRs and med students are not good at properly identifying significant history or findings. Why? Because what may be significant for patient A may be irrelevant for patient B. Similarly, what’s significant in the eyes of a cardiologist may be not mentioned by a neurologist seeing the same patient. Continuing along this trend, what a physician finds significant for today’s office visit may not be so for a visit a few months later. This is not easy stuff for humans to learn, so it’s not surprising that computers struggle with it even today.
We have reason to be hopeful, though. With recent advances in artificial intelligence and analysis of big data, computers are getting better at identifying relevant and significant information in the EHR. We’re not in nirvana just yet of course. But the tech is developing quickly, and soon EHRs will be much more user-friendly, and maybe they’ll even be…usable!