点击显示 收起
Washington University School of Medicine
Department of Psychiatry
St Louis, MO 63110
E-mail: bishopm{at}psychiatry.wustl.edu
Dear Sir:
I read with interest the recent Journal article by Wang and Zhang (1) on changes in the association between overweight and family income in children and adolescents using data from the National Health and Nutrition Examination Survey (NHANES) between 1971 and 2002. The authors are to be commended on their ambitious undertaking of examining secular trends in the relation between overweight and socioeconomic status (SES); such findings may have important insights for developing effective obesity prevention and management programs as well as public policies. Wang and Zhang point out that each of the commonly used variables of SESincluding education, income, and occupational statushas its own strengths and limitations for studying the relation between SES and health outcomes. In their study, SES groups were determined by the poverty income ratio tertiles at each survey timepoint (NHANES I, 19711975; NHANES II, 19761980; NHANES III, 19881994; and NHANES 19992002), and then low-, middle-, and high-SES groups were compared on the basis of these tertiles. It is important to address some potential limitations to this approach.
The US Census Bureau's poverty threshold is an "absolute" threshold, fixed at a point in time (1963) and updated solely for price changes (2). As such, it does not reflect changes in standard of living, taxes, or other factors affecting SES over time. In addition, the poverty threshold is constant across the United States, and no allowances are made for dissimilar costs of living in different geographic areas. Thus, a child living in the rural Midwest with a family income of $40 000 would be in the same SES category as a child living in a coastal metropolis on the same $40 000 income. Furthermore, the poverty threshold does not account for family expenses, such as childcare costs for working and nonworking families; therefore, a working family spending 10% of their income on childcare would be in the same SES category as a nonworking family making the same income but without the extra expense. These concerns are important not only for determining the SES grouping of children in each individual NHANES cohort, but also for understanding the possible differing SES classifications across timepoints that result from temporal changes in the factors affecting the poverty threshold.
Wang and Zhang found in their analyses that, contrary to widely held perceptions, not all low-SES groups were at increased risk of overweight; they also found that, over the past 3 decades, there has been an overall weakening association between SES and overweight, especially in the later waves of NHANES (19881994 and 19992002). Given the above concerns regarding the definition of SES, their results should be interpreted cautiously. An alternative explanation for the findings could be that the poverty threshold was a more accurate indicator of SES in the earlier cohorts, whereas, for the later groups, nationwide changes such as increases in the numbers of working or single-parent families, unequal geographic changes in the cost of living, or greater consumerism (or all 3 circumstances) make the poverty threshold a less reliable SES measure. These variables may also affect different races to greater or lesser degrees. Thus, the apparently changing associations over time among overweight, race, and SES could in fact reflect the comparison of unequal SES groups.
Despite these concerns over the authors' choice of SES indicator, their study shows the real and significant increase in child and adolescent overweight spanning the last 3 decades and highlights the urgent need for further research into the causes and potential prevention and management efforts required to fight the childhood obesity epidemic.
ACKNOWLEDGMENTS
The author had no personal or financial conflict of interest with the study by Wang and Zhang or with those authors themselves.
REFERENCES