Physicians best computers in diagnostic showdown

Flesh-and-blood physicians outperformed four symptom-checker apps when diagnosing patient simulations in a new head-to-head study.

The study, published in JAMA Internal Medicine, pitted 234 physicians against 23 commonly used computer algorithms called symptom-checker apps. Both physicians and the symptom-checker apps had to evaluate 45 clinical scenarios consisting of 15 high, 15 medium and 15 low-acuity conditions. Twenty-six of the scenarios involved common conditions and 19 incorporated uncommon conditions. For each instance, physicians and the apps had to list the top three most-likely diagnoses for the condition.

The physicians listed the correct diagnosis first 72 percent of the time, compared with 34 percent of the time for the digital platforms. Additionally, 84 percent of clinicians listed the correct diagnosis in the top three possibilities, compared with 51 percent of the virtual counterparts that did the same.

"While the computer programs were clearly inferior to physicians in terms of diagnostic accuracy, it will be critical to study future generations of computer programs that may be more accurate," said senior investigator Ateev Mehrotra, MD, an associate professor of healthcare policy at Harvard Medical School in Boston.

Despite displaying superior diagnostic skill, physicians still made errors approximately 15 percent of the time. Researchers suggested using computer-based algorithms concurrently with human physicians may help reduce diagnostic errors.

"Clinical diagnosis is currently as much art as it is science, but there is great promise for technology to help augment clinical diagnoses," said Dr. Mehrotra. "That is the true value proposition of these tools."

More articles on quality: 
Severe joint pain increasing among Americans 
Breast cancer treatment costs vary significantly across treatment regimens, study finds 
Which factors most influence physicians' intentions to admit to medical errors? 3 insights

© Copyright ASC COMMUNICATIONS 2020. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 


IC Database-3

Featured Webinars

Featured Whitepapers