QuBBD: Statistical & Visualization Approaches for Precision Medicine
Our NIH R01-funded (NIBIB & Office of the Director) research study focuses on leveraging patient-generated health data from patient-reported outcomes and sensor data from wearables and apps to determine precision health recommendations for lifestyle behaviors to improve health outcomes. Our initial use case is focused on data from patients with Inflammatory Bowel Diseases or IBDs through our partnership with IBD Partners’s (formerly CCFA) Patient-Powered Research Network, which has been funded by the Patient-Centered Outcomes Research Institute.
Our work focuses on Bring-Your-Own-Device (BYOD) mHealth data from wearables like Fitbits and smartphone apps that measure physical activity and sleep. There are 3 overall objectives to this work:
- To develop preprocessing methods for mHealth data to deal with sparsity, missingness, and batch effects, which will lead to a high quality data pipeline of mHealth data focused on physical activity and sleep.
- To create new machine learning methods with formal inference to predict health outcomes based on patient-reported outcomes and mHealth data at an individual patient level and cohort level.
- To design and develop an interactive, visual analytics software platform to enable data exploration at the point-of-care.