(TA2.34) Toward better collection and analysis of data in healthcare applications

Organiser(s): George Washington University; Renmin University of China; Columbia University; Beijing Institute of Geriatrics National Health Commission

Recent technological revolutions facilitate collection of large volumes of multi-source data in healthcare industry. While analyses of these data open doors to remarkable advancements, challenges impeding the discovery of meaningful information may still exist. For instance, the analysis may be problematic if the data are not appropriately collected. This panel of speakers will discuss ways to improve the collection and the analysis of data with healthcare applications. The panel discussion will begin with an introduction of the speakers and outline of the program. A Q&A session will be held to answer the questions from audience. This panel session will cover: 1.Healthcare data platform. This platform originated from a cohort study for healthy aging. The data was collected from carefully designed questionnaires, wearable devices, and lab results. More than 6000 variables in different categories are collected from 26000 participants ranging from different age groups, occupations, and geo-regions. This well-collected data accomplishes an analysis of aging from health, psychological, and social aspects. The panelist will discuss the methodologies for collecting the useful data based on this data platform. 2. Causal inference and adaptive data collection. The causal discovery for the effect of an intervention becomes increasing important in today’s healthcare application. If the data collection process is confounded with certain unknown factors, the results may not be reliable. Causal inference methods are tools for adjusting analyses from such bias. In addition, adaptive data collecting methods can be used to achieve the compound goals such as reducing risks of confounding, treating more patients with better intervention, obtaining higher efficiency for detecting the effect. The panelist will discuss the application of these methods for improving the data collection process as well as the detection of treatment effects.