QuBBD: Statistical & Visualization Methods for PGHD to Enable Precision Medicine

Univ Of North Carolina Chapel Hill, Chapel Hill NC

Investigators

Abstract

The purpose of this proposal is to develop a combination of innovative statistical and data visualization approaches using patient-generated health data, including mobile health (mHealth) data from wearable devices and smartphones, and patient-reported outcomes, to improve outcomes for patients with Inflammatory Bowel Diseases (IBDs). This research will offer new insights into how to process and transform patient-generated health data into precise lifestyle recommendations to help achieve remission of symptoms. The specific aims of this research are: 1) To develop new preprocessing methods for publicly available, heterogeneous, time-varied mHealth data to develop a high quality mHealth dataset; 2) To develop and apply novel machine learning methods to obtain accurate predictions and formal statistical inference for the influence of lifestyle features on disease activity in IBDs; and 3) To design and develop innovative, interactive data visualization tools for knowledge discovery. The methods developed in the areas of preprocessing of mHealth data, calibration for mHealth devices, machine learning, and interactive data visualization will be broadly applicable to other mHealth data, chronic conditions beyond IBDs, and other fields in which the data streams are highly variable, intermittent, and periodic. This work is highly relevant to the mission of the NIH BD2K initiative which supports the development of innovative and transformative approaches and tools to accelerate the integration of Big Data and data science into biomedical research. This project will also enhance training in the development and use of methods for biomedical Big Data science and mentor the next generation of multidisciplinary scientists.