Three out of the four smartphone apps the study reviewed, incorrectly diagnosed nearly one third of melanomas as "unconcerning", based on evaluation of uploaded images of skin lesions.
Senior investigator Laura Ferris, assistant professor in the Department of Dermatology at University of Pittsburgh School of Medicine, and colleagues, report their findings in the 16 January online issue of JAMA Dermatology.
Health Care AppsToday's smartphone is a hybrid between a mobile phone and a personal computer, with a sophisticated operating system capable of running a wealth of software applications, or "apps".
When they first came out, you could only get apps for communication and entertainment, but nowadays, you can get apps for "everything under the sun, including health care," says Ferris in a statement.
According to sources cited in the study's background information, the largest online app store now markets over 13,000 health care applications to consumers, and in 2011, the mobile app industry was worth over $700m worldwide.
Ferris and colleagues give examples of health care apps that help users learn about adverse effects of medication, track their calorie intake and energy expenditure to manage weight loss, and even a fertility app that monitors a woman's menstrual cycle to pinpoint favorable times for conception.
One area of growth is self-tracking of health, where in combination with wearable sensors, mobile apps collect, process and display a wealth of personal data and daily activity measures to help users keep track of and manage all aspects of their health.
There are also a number of educational apps that give information about melanoma and how to examine your skin to help you track the development of individual skin lesions.
However, the researchers caution that while with the help of such tools, users become more aware of their health, and improve communications with their doctor, it's important they don't let the apps take the place of expert medical advice and diagnosis.
The StudyFerris and colleagues tested four apps from the two most popular smartphone platforms.
The apps are marketed to consumers to help them decide whether their lesions, which they photograph and upload to the apps for analysis, are potential melanomas, or otherwise concerning, or whether they are benign and nothing to worry about.
The apps are offered at low cost or even free, and are unregulated and unvalidated, note the researchers.
Altogether the researchers uploaded 188 photos of skin lesions to each of the four apps. Different apps analyzed them in different ways. One used an automated algorithm, another had them reviewed by an anonymous board-certified dermatologist.
The results show that the ability of the apps to assess melanoma risk "is highly variable", note the authors.
Three out of the four apps incorrectly classified 30% or more of melanomas as unconcerning.
The only app that showed a high degree of sensitivity in diagnosis was the one that had the images analyzed by a dermatologist. The researchers found only one of the 53 melanomas they uploaded photos of was diagnosed as "benign".
This was the most expensive app, with a price tag of $5 per image evaluated.
ImplicationsAlthough the apps carry disclaimers that say the information they provide is for educational use only, the researchers were concerned that people would nevertheless come to rely on them rather than consult a medical professional.
The risk of relying on diagnosis by app could be higher among the uninsured and poorer groups in society, note the researchers, especially as it is patients themselves who tend to be the first to notice a melanoma.
"If they see a concerning lesion but the smartphone app incorrectly judges it to be benign, they may not follow up with a physician," says Ferris.
"Technologies that decrease the mortality rate by improving self- and early-detection of melanomas would be a welcome addition to dermatology. But we have to make sure patients aren't being harmed by tools that deliver inaccurate results," she urges.
The study was funded by the National Institutes of Health (NIH) and the University of Pittsburgh Clinical and Translational Science Institute.