Comparison of Three Methods to Estimate the Relationship Between Physical Activity and Health Indicators Among Adults with and Without Disability

• Activity Number: 270
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 2:45 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Qing Zhang* and Courtney-Long Elizabeth A. and Michelle Sloan and Stevens Alissa and Dianna D. Carroll
• Companies: CDC and CDC and CDC and CDC and CDC
• Keywords: splines ; graph ; nonlinear ; margin effect ; polynomial logistic

Abstract

According to Department of Health and Human Services physical activity(PA) guidelines, adults who get ? 150 minutes/week of aerobic physical activity may reduce their risk of many adverse health outcomes. Research has shown that there may be a dose-response relationship between PA and health outcomes such that health benefits may accrue as the volume of activity increases. However, there is limited literature addressing this dose-response issue among adults with disability. In this study we use 2010 National Health Interview Survey (NHIS) data to evaluate the relationship between minutes of PA and chronic health indicators (i.e.,obesity, hypertension, and diabetes) among adults ?18 years by disability status. Using 3 estimation methods (locally weighted scatterplot smoothing, restricted cubic splines and quartic polynomial logistic regression) the relationship between PA volume and the 3 health indicators were evaluated and displayed graphically. In the preliminary analysis for each health indicator, all three models suggest that the prevalence decreased as PA volume increased in different ways. The similarities, differences, pros and cons for each estimation method are addressed.