Analyzing the Impact of Modifiable Lifestyle Behaviors on Various Health Outcomes

Abstract

We develop a generalized partially additive model to build a single semiparametric risk scoring system for physical activity across multiple populations. We model each score component as a smooth term, an extension of generalized partially linear single-index models, due to the nonlinear relationship between physical behaviors and various health outcomes. We use penalized splines and propose two inferential methods, one using profile likelihood and a nonparametric bootstrap, the other using a full Bayesian model, to solve computational problems. Both methods exhibit similar and accurate performance in simulations. These models are applied to the National Health and Nutrition Examination Survey (NHANES) and quantify nonlinear and interpretable shapes of score components for all-cause mortality. Modifications are made to the 2018 World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) Score to target recommendations regarding all-cause mortality, cancer mortality, and cancer risk jointly among older adults in the NIH-AARP Diet and Health Study. Weights were incorporated for each Score component; a continuous point scale was developed; and some cutpoint values were changed based on evidence-based recommendations. Exploratory aims examined the impact of separating components with more than one sub-component and whether all components were necessary to retain within this population. Findings suggested incorporating weights improved the score’s predictive performance for all-cause mortality and provided more precise estimates in relation to cancer risk and mortality outcomes. The importance of being a healthy weight, being physically active, and consuming plant-based foods regarding cancer and overall mortality risk were highlighted in this population. A bivariate factor-by-curve spline regression is used to model the joint effect of physical activity measures on fatigue in breast cancer survivors in a randomized controlled physical activity trial. The factor-by-curve component creates a distinct bivariate function for the study groups. This model is applied to the COMPARE study, an intervention focused on increasing physical activity in breast cancer survivors. The resulting functions display a different relationship on fatigue for each group, with a statistically significant group-type interaction. These findings highlight the importance of increasing physical activity intensity and allocating activity in shorter bouts throughout the day to reduce fatigue in this population.