Digital health in real-world data: challenges and opportunities

This chapter focuses on data collected from digital health technology tools (DHTT) such as smartphones, tablets, and wearable sensors. It discusses the opportunities offered by sensor-derived digital outcome measurements over traditional clinical outcomes. It describes the verification and validation steps needed to ensure that the digital outcome measurements derived from DHTTs are both trustworthy and relevant for comparative effectiveness and personalized medicine research. It then gives an overview of some statistical challenges associated with digital outcome measures, including the impact of low signal-to-noise ratio of frequent digital assessments and the curse of dimensionality with high-dimensional feature spaces. Throughout the chapter, examples of digital outcome measurements in Parkinson's disease, multiple sclerosis, and cardiovascular diseases are used.