|
Abstract:
|
Clinical trials that incorporate mobile health data require special attention to their design and analysis. For example, despite the richness of data generated by mobile health (mHealth) devices, issues with missing data continue to have implications on the interpretation of findings in trials that rely on mHealth data. Points will be illustrated with two clinical trials that generated mHealth data as key endpoints. The Apple Heart Study 1.0 (AHS 1.0) was conducted to evaluate the agreement between Apple’s newly developed irregular pulse watch notification algorithm (IPNA) and a gold standard for identifying signals consistent with atrial fibrillation and atrial flutter. Stanford GOALS was a randomized clinical trial designed to evaluate a multi-faceted intervention to reduce obesity in children, where the role of physical activity was key to understanding changes in body mass index over time. Both faced unique issues related to missingness. Possible solutions for these issues are presented that investigators may consider when designing their own studies.
|
