CMStatistics 2021: Start Registration
View Submission - CMStatistics
B0842
Title: Functional survey weighted modeling to associate physical activity and non-alcoholic fatty liver disease Authors:  Ekaterina Smirnova - Virginia Commonwealth University (United States) [presenting]
Abstract: Accelerometers present an objective alternative to assessing physical activity in multiple settings and allow continuous monitoring of physical activity in both lab and free-living environments. The National Health Examination Study (NHANES) is the largest US-population study that contains publicly available physical activity data obtained from wearable accelerometers together with extensive health and occupational information. The raw data is typically summarized into a minute-level accelerometry count measure, which leads to functional data collected over 1440 minutes per subject day. The NHANES study participants are recruited from the US population according to a survey design procedure, which has to be accounted for in statistical inference. To accurately associate these data with health outcomes, day-level summaries such as total activity volume, time spent in total sedentary, light and moderate to vigorous activities are often used. However, these summaries may be highly correlated with each other and may not capture the full complexity of the functional data patterns. We illustrate the utility of traditionally summarized features and functional data models that account for the complex survey design in the context of predicting the development and progression of non-alcoholic fatty liver disease (NAFLD). We further discuss the connection of physical activity to racial, occupational and socio-economic disparities in patients with NAFLD.