Can computer-assisted decision tools help reduce diagnostic errors? Elizabeth Tracey reports


Just shy of about 6% of people who visit an emergency department will not be correctly diagnosed, a federal agency study led by David Newman-Toker at Johns Hopkins has shown. Can computer-based decision support tools help bring that number down? Newman-Toker comments.

Newman-Toker: I think for a long time the medical community has been anticipating this magical future where computer-based decision support really helps us, and in fact, people have been envisioning that future since the 1950s even. It’s a harder nut to crack than people give it credit for. I think there are going to be things that work very soon relatively speaking. Those are things where we already have digitized datasets on which to train these computer systems to make them experts, so things like radiology.  :30

Newman-Toker says all these tools rely on huge amounts of data being collected, analyzed and incorporated into algorithms that point to probable diagnoses a clinician might consider, and while that seems persuasive, there are also errors inherent in this process. At Johns Hopkins, I’m Elizabeth Tracey.