CHARACTERIZING MEDICAL ANDROID APPS

There is a proliferation of medical mobile apps: Google Play alone has thousands of apps in the “Medical” category. Many such apps perform critical tasks (e.g., are used with a medical device or in lieu of a device); handle sensitive patient related information; perform diagnosis; or treat diseases. However, there are wide gaps between an app’s claims and users’ expectations as well as between app implementations and regulatory frameworks’ mandates. We perform the first study, based on analyzing more than 4,000 Android apps, that characterizes medical apps. We begin by introducing an automated classification scheme that integrates textual information extracted from multiple sources to establish the purpose and target audience for an app, based on fine-grained traits and high-level categories; we found that the most common functionalities involved connecting to medical devices (e.g., hearing aids, glucometers), offering tele-health services, or patient management. We then dive deeper into app nature and characterize according to the function and domain of the app. We reveal actionable findings found in various facets of medical applications, regulatory frameworks and user privacy and safety.