In the past decade, overweight in children has become a major national health concern. Hundreds of studies, both descriptive and etiologic, have identified factors such as lack of physical activity and poor diet that increase the risk of overweight. This project seeks to enrich the analysis of overweight in children by moving beyond delineating risk factors associated with the condition to specifying those factors that, if appropriately targeted, can lead to the greatest reduction in the number of overweight children in the population. In order to do this, we propose to generate multifactorial population attributable fractions (PAFs) to quantify the potential impact of partially or completely eliminating risk factors for child overweight from the population. While it is numerically straightforward to generate a PAF one factor at a time, it is not straightforward to generate multifactorial PAFs-PAFs that are mutually exclusive and mutually adjusted. While some approaches for estimating these complex PAFs have been proposed in the literature, many methodological issues remain and, as a consequence, the use of PAFs in practice has been limited. This project will use data from the National Survey of Children's Health (NSCH) to produce multifactorial PAFs that reflect the interconnectedness of behavioral, clinical, and social factors related to overweight in children and that can inform strategies for prevention. The methods development will include assessing techniques for separately handling modifiable and unmodifiable risk factors, examining alternatives for averaging and weighting PAFs, and providing computational approaches for estimating the standard errors of the PAFs.
The specific aims of the project are: 1) To extend and refine methods for estimating multifactorial population attributable fractions (PAFs) by assessing various numerical alternatives; 2) To apply the PAF methods to child overweight, illustrating the way in which PAFs can be used to identify a set of risk factors which, if addressed, would have the greatest impact on reducing an outcome in the population; and 3) To disseminate the methods, including dissemination of guidelines, programming code, and other documentation, to practitioners engaged in priority-setting and program development. The team of study investigators at University of Illinois at Chicago School of Public Health (UIC-SPH) includes two maternal and child health epidemiologists and a biostatistician, and has a long history of working with professionals in public health agencies to apply epidemiologic and biostatistical methods to the analysis required for needs assessment, surveillance and monitoring, program planning, and policy development. ? ? ?
