Going well beyond Dr. Google, personalized, interactive apps promise consumers fast, convenient and accurate diagnosis – or at least enough information to decide whether it’s urgent to see a doctor. But little is known about the accuracy of those claims. This article examines the evidence.
By Michael L. Millenson
Diagnostic apps with capabilities well beyond static “Dr. Google”-type search engines use patient questionnaires, sensors and crowdsourcing to tell you if a darting pain may dictate a dash to the emergency room or if a darkened mole could denote skin cancer. Some of these direct-to-consumer (DTC) apps promise help in rooting out a rare diagnosis, while others put common diagnoses into context.
App evaluations vary considerably
However, even though some of these apps have been downloaded tens of millions of times, little is known about the evidence supporting their claims. In a recently published article published in Diagnosis, the authors present the results of a review of the peer-reviewed literature, as well as the trade press and other sources of information in what’s called the “gray literature,” for the period January 1, 2014 to June 30, 2017.
The most common types of apps for which there was evidence were those providing general diagnostic and triage advice or those related to dermatology. The evaluations themselves were highly variable; about half simply described app characteristics, while half examined actual performance.
Lack of research on performance quality
Moreover, that performance varied widely, whether in functionality or accuracy. However, the usefulness of the evidence was limited by a frequent failure to name the individual apps evaluated. Moreover, there were no studies of real-world consumer use. Overall, the evidence base for app usefulness was sparse, uneven and inconclusive.
“Glitter with promise”
Differences in the technological underpinnings of apps (e.g. use of sensors vs. questionnaires) were not always recognized by researchers. As DTC diagnostic apps rapidly evolve, clinicians, patients, policymakers and others need much more, and more reliable, information.
“Apps that promise a diagnosis may glitter with promise, but, unfortunately, there’s been little rigorous examination of how well those promises hold up in real life. People deserve better,” explains Michael L. Millenson, a researcher of the study.
Apps are medical devices, and accurate and timely diagnosis is a significant health issue. At a time when the U.S. Food and Drug Administration is reducing its regulatory oversight, this article clearly frames the urgent need for reliable information enabling clinicians and patients alike to distinguish between what’s helpful and what’s hype.
Read the original article here