BACKGROUND: Epidemiologists, researchers and clinicians conducting health related work often use large datasets and rely on data that may contain protected health information or other information participants do not want to have disclosed. Developing mobile health applications to address these concerns and provide information to impact health and the environment for conducting impactful work involves considering numerous and non-trivial technological and data analysis challenges.
METHODS: We cultivated a diverse team to address these multiple challenges and were funded by CSTE to design a mobile app for the purpose of collecting behavioral health and substance use information from college age students. This effort required the contribution of a multidisciplinary team of experts to address mobile usability, software development, data analytics, social media, communications, psychology, epidemiology and survey research. Our team of developers, designers, engineers, and researchers collaborated on the development effort using a common communication process, and continual knowledge transfer methods. We used a HIPAA compliant platform and applied gamification theory to develop mobile apps for both Android and iOS to survey population-based samples regardless of the population or the disease/condition of interest.We tested the methodology and the branded mobile app on a sample of students at three small state colleges in Vermont. We applied a specific technological approach using knowledge across disciplines to test the user experience.
RESULTS: The technological requirements that were addressed included: designing on both iOS and Android platforms, data security, addressing app registration requirements when designing for HIPPA compliant platforms, consent forms, data management of sensitive data, tracking dashboard, and push notifications. The branded SHAPE app was found to be user friendly, easy to use, and appealing to the target audience. Our prototypes revealed that mobile applications can be leveraged for geo-location analysis, machine learning and text mining in addition to layering of differential privacy to enable analysis by broader groups for more impactful findings.
CONCLUSIONS: Using mobile apps for health surveillance and to impact health holds promise when certain technology approaches and analysis rules are maintained. To use mobile apps for health requires a multi-disciplinary research team, expanding beyond the conventional clinical and survey research expertise to include computer science, data science, and information science.