Studying the Child Obesity Epidemic with Natural Experiments

Publication Date:

Publisher: National Bureau of Economic Research

Author(s): R. Sandy; G. Liu; J. Ottensmann; R. Tchernis; J. Wilson; O.T. Ford

Research Area: Health

Type: White Paper

Abstract:

This paper is part of a series on obesity published by the National Bureau of Economic Research. The authors used data from pediatric visits to estimate effects of certain changes to the environment on children’s body mass. Environmental factors examined in this study are fast-food restaurants, supermarkets, parks, trails, violent crimes and 13 types of recreational amenities. The importance of the built environment is increasingly being recognized as child obesity grows in the United States, in part, because policies can affect the built environment more easily than some of the other factors contributing to obesity. Data were drawn from electronic medical records for 60,000 children aged 3–18 who were seen at an academic health care center in Indianapolis, Ind., between 1996–2006. Other data came from aerial photographs, police records, county health department records and the U.S. census.



The study results are complex. The authors used two methods; cross-sectional analyses and fixed-effects (FE) analyses, and the two methods produced conflicting results. For example, in the cross-sectional analysis the presence of fast food increased children’s body mass; in the FE analysis, the effect was opposite. Although the authors view is that the cross-sectional methods produced more reliable results, the FE result “cast doubt on the highly publicized policies to reduce fast food exposure as interventions for preventing obesity.” In general, the study did not elucidate clear-cut methods for reducing children’s body mass. However, the amenities that seemed to produce the most effects were the introduction of playing fields, kick ball fields and volleyball courts.



Limitations of this study are that the time frame was short term (results were assumed to occur within a year of the arrival of any amenity), and that close proximity of an amenity was assumed to be related to its use. No actual data on when and how families used these amenities was collected. Future work examining effects over a longer time period, and incorporating more detailed data about the families using these amenities should yield worthwhile insights into relationships between children’s weight and the built environment.