COULD MEDICAL APPS KEEP THEIR PROMISES?

Medical mobile apps are already in wide use, and their use, as well as user base are projected to grow even further. However, it is unknown whether medical apps achieve their claimed behavior effectively and accurately. To determine potential gaps between app claims and app behavior, as well as between app claims and user expectations, we conducted a study on over 2,000 Android apps. We first developed an information retrieval approach that maps an app’s description to medical (ICD) terms, hence delineate the app’s medical scope and stated goals; our analysis has revealed that weight management, heart rate measurement, blood sugar measurement and hearing aids constitute the most common conditions apps claim to address. Next, based on app functionality, we categorize apps into (a) apps that measure or manage a physiological parameter, (b) apps that claim to treat conditions, and (c) apps for self- assessment. Within these three categories, we establish fine- grained subcategories and for each subcategory we compare apps’ claimed behavior with realizable behavior. We found that app widely overstate their behavior and functionality. We also found that apps employ disclaimers and misleading terms to lure users into installing/using the app yet avoid responsibility. Finally, based on our uncovered app behavior and claims, we outline actionable findings w.r.t app claims and actual vs. stated function, meant to make users safer and apps more forthright.