Are diagnostic health apps accurate? Researchers say there's no way to tell

As more people turn to their phone or laptop for a medical diagnosis, some industry experts are pointing to a growing evidence gap that could leave consumers unable to determine which apps are most effective. 

The number of people who turn to websites and apps to answer their health questions is ever-growing, as is the number of websites and apps that claim to do so. WebMD, the most frequently-accessed health website after, says 75 million unique users access its websites and mobile apps every month. But health apps are only regulated in select instances, and the idea that everything is true on the internet has become laughable.

Still, it isn’t clear whether consumers are accessing false or factual health information, as digital health tools have not been sufficiently evaluated, according to a study published last month in the open-access journal Diagnosis.

The researchers reviewed 36 evaluations of medical apps. Most evaluations examined sites and apps that claimed to diagnose conditions based on symptoms a user selected or images a user uploaded.  

RELATED: Bolstered by FDA's policy shift, ‘digiceuticals’ could soon carve out a bigger role in healthcare

Apps that provided triage advice “varied widely” when it comes to accuracy, according to researchers, who used sensitivity (the ability to rule out a disease) and specificity (the ability to rule one in) as metrics. Apps that tried to diagnose specific diseases “showed few favorable results,” as did specialty symptom checkers (e.g., one for knee pain diagnoses), though apps that linked to sensors performed somewhat better.

“Our team was shocked by the paucity of the evidence,” the study's lead author, Michael L. Millenson, a researcher at Northwestern University’s Feinberg School of Medicine, said in an email to FierceHealthcare.

The authors concluded that far more investment needs to be devoted to the evaluation and research of consumer-facing diagnostic apps, and future evaluations must be tethered by a set of established standards. 

How quickly that happens remains to be seen. Urmimala Sarkar, M.D., an associate professor at the University of California, San Francisco, School of Medicine who studies patient safety, health information technology and the implementation of evidence-based innovations, concurred with the study’s conclusions.

“If there were funding available to do these types of studies, regardless of whether the companies wanted to do them, they would get done,” Sarkar told FierceHealthcare. But currently, she explained, federal funding for these studies is lacking, and app developers have no obligation or incentive to conduct them.

RELATED: FDA maintains its focus on digital health in 2018

Sarkar contrasted the regulatory landscape for health apps with that of another key product that can impact health: prescription drugs.

For a new drug to be approved, its safety and efficacy must be demonstrated through high-quality clinical trials, she said. While clinical trials may not be appropriate for all apps, “there is a floor for evidence for pharmaceuticals, and there’s no floor for digital health.”

That floor may not arise anytime soon. Regulatory oversight in health realm is currently decreasing, not increasing, Millenson noted. While the FDA has some oversight of medical apps, it keeps a long list of examples in which it exercises enforcement discretion, including apps that provide a list of medical conditions based on symptoms. 

"Doctors and patients more than ever need timely and accurate information they can trust," he said.