Patterns (e.g. frequency, amount, etc.) of dietary intake and daily physical activity have each been independently linked with an increasing prevalence of obesity. Yet, connecting these patterns to obesity and chronic disease through the integration of time has not been previously considered. Given the strong evidence for independent linkages between each of these patterns and obesity, there is a critical need to determine the potential synergistic correlation of these patterns of behavior within the framework of a relationship with health. In the absence of such insights, opportunities for early detection of behavioral patterns that predispose obesity and chronic disease will be missed, and our long-term research goal to create these early detection strategies will not be met. The central hypothesis of this project is based on the analytical framework and methodology that was previously developed by the investigators: that daily patterns of energy intake, when integrated with physical activity, will be associated with health in a representative sample of U.S. adults 20 to 65 y (NHANES 2003-2006). The objectives in this R21 application include the development of data patterning methodology that can be used to create distinct dietary intake and physical activity patterns and then successfully integrate these patterns to identify population temporal pattern clusters. Next, the investigators will evaluate the cluster relationships with obesity and health outcomes and compare the integrated clusters to the un-integrated dietary and activity pattern clusters. The working hypotheses are firstly, that novel distance measures based on dynamic time warping for the dietary and physical activity data can be integrated to produce meaningful clustering related to health, and secondly, that a population cluster which exhibits a pattern of evenly spaced eating occasions, moderate energy consumption and moderate physical activity patterns in a 24 hour day will be associated with normal weight and without chronic disease, and that relationships with health outcomes will be stronger for the integrated temporal pattern clusters compared with un-integrated temporal dietary clusters and physical activity clusters. The rationale for this research is that its successful completion is expected to create data reduction methods that classify temporal lifestyle patterns linked to disease, further, the expectation is that these analytical techniques will integrate multidimensional temporal dietary and physical activity data. These outcomes are expected to have a significant positive impact, not only in developing/evaluating new analytic methods but in laying the groundwork for data based preventative interventions. This proposed research is potentially significant because results will provide a starting point for understanding the importance of the timing of dietary and physical activity patterns to the prevention of obesity and disease with potentially broad translational
TO NIH The proposed research is relevant to public health because preventative time-based dietary and physical activity patterns will be identified among the US population, which are expected to open new research horizons for evidence-based recommendations for healthy lifestyles. The project is relevant to NIH's mission because individualized early detection of lifestyles supporting health or linked with disease may be attainable through the development of analytical techniques integrating multiple layers of data. Such strategies could be applied to utilize the increasing information from personal electronic monitoring devices and to detect patterns associated with positive and negative health outcomes.
