Researchers at the University of Illinois are in the early stages of developing a computer-generated physician or nurse that can explain test results to patients and next steps in treatment via patient portals available through an electronic health records system.

For now, the work is in the proof-of-concept stage, with researchers describing the project in the March issue of the Journal of Biomedical Informatics. The work is being done at the University of Illinois’ Beckman Institute for Advanced Science and Technology, and the Carle Foundation Hospital’s Research Institute.

Also See: Patient portals critical in the new MACRA, MIPS era

While many patients today have access to their electronic health records when they log into their portal, they see a bunch of test results, often represented as numbers that can be difficult to understand, says Dan Morrow, lead author and educational psychologist at the University of Illinois. “You see a table of numbers and scores but you don’t get context on ranges and risks, and you don’t know if a low score is normal or a high one is normal,” he explains. For now, the numbers are simply a basis for having a conversation with your doctor.

Dan Morrow

And for now, a computer-generated doctor has to wait as researchers take baby steps and build up to that type of technology. For starters, they are working on trying to enhance the presentation of results whether received in the mail or via a portal, by color-coding scores. A low score for cholesterol, for instance, would be green, a medium score would be yellow, and a high score would be red.

But work is underway to create a more realistic patient-physician dialogue. A retired physician recorded scripts for the patient portal messages and other text needed to develop the avatar clinician’s commentary. Multiple scripts have been developed to emulate how various types of results would be explained to patients.

Researchers also testing whether patients’ “gist memory,” which is fuzzy representations of an event, differs depending on whether the avatar speaks in a natural voice or a computer-generated voice.

As a first step, researchers brought older adults into the lab and had them go through several mock patient scenarios that influence heart disease. Some of the adults, acting as patients and presented with mock test results, would see a list of numbers or the same numbers but color coded. Another group would see a video of a provider presenting results and told how they should think about the numbers, then assess how well the older adults understand and retain those results.

The reality is that most people like information in a video but most physicians won’t use video to explain results so the idea was to create an avatar physician or nurse that appears in the video, according to Morrow. “We want to emulate best practices in a real-life face-to-face environment.”

The avatar clinician is programmed to display appropriate facial expressions, gestures and other cues that promote patient understanding as if findings were being presented by a human.

Having received a grant to start the project, the effort recently received another grant to generate other clinician avatars, and now is seeking grants for pilot studies with patient portals next year.

If everything goes well, Morrow believes a first-generation product could be a reality in two or three years.

Register or login for access to this item and much more

All Health Data Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access