EHRs Are Not a Magic Bullet

Share this story:



By Michael Burger, Practice Lead, EHRs and EDI

Electronic health records (EHRs) are taking a lot of heat these days. Lloyd Minor, M.D., the dean of Stanford University’s School of Medicine, recently joined a vocal group of physicians who blame EHRs for physician burnout. He also calls for new systems that integrate technologies, like voice recognition, to make things simpler for the EHR user.

I think that marketplace and technology realities are really to blame for physicians’ frustrations, not EHRs. Consider the following.

  • Unknown user demand for diagnostic tools. EHRs today are built as clinical workflow and data storage tools, not as diagnostic tools. Physicians push back against even basic types of EHR decision support (like alerts and reminders), so there isn’t clear user demand to use EHRs diagnostically. It could be that a search engine really IS a better diagnostic tool. If physicians become interested in using EHRs that way, EHRs will need to be rearchitected.
  • Are apps the answer? It makes sense to yearn for the simplicity of an iPhone when thinking about an ideal EHR. But is that practical? There’s a litany of popular mobile applications (apps), like Twitter, Instagram and Snapchat. Each is a simple, single function, limited-ability application; however, such apps are not mission critical and are too simplistic to address the complexities and multifaceted delivery of healthcare.  Autocorrect in a text message can be amusing, but life threatening in an EHR.
  • Speech recognition. Speech recognition has been around for years.  Virtually every EHR supports it. Accuracy is at an all-time high, but still is not at 100%. Do we want clinical data that’s only 95% accurate in our medical records? Just as important, transposing text (using Natural Language Processing) has limitations in what can be converted into codified data—which is what’s needed to power decision support and diagnostic tools.  This challenge is the result of technology limitations, not EHR design.
  • Freeing physicians from keyboards. Minor especially chafes at physician data entry. Until EHRs took over, physicians focused on clinical decision making and dictated notes that were transcribed (by someone else) to the patient's chart. With an EHR, physicians assume responsibility for inputting information instead of outsourcing the task. Somehow there is an expectation that the clinical data will magically appear in the EHR without someone entering it. If data entry is onerous, the answer is to outsource the data entry, not contend that EHRs are causing burnout.

There is clearly room for improvement in the workflows in today’s EHRs.  Features are click-heavy and require a lot of typing to complete basic tasks.  Those are redesigns that make sense, and usability improvements happen as software of any type evolves.  The iPhone has shown the EHR industry how important usability is, and that speech recognition has a place. Vendor and end-user collaboration, coupled with reasonable expectations, are a smart combination to improve usability. Reasonable expectations are key, because EHRs aren’t magical. What do you think?  Email me: michael.burger@pocp.com.

Other recently published articles on this topic:

EHRs Offer Ready-Made Tools for Population Health Management