QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Online Supporting Material Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wilde, P. E. Articles by Peterman, J. N. Search for Related Content PubMed PubMed Citation Articles by Wilde, P. E. Articles by Peterman, J. N.

---

Community and International Nutrition

Individual Weight Change Is Associated with Household Food Security Status^1,2

Friedman School of Nutrition Science and Policy, Tufts University, Boston, MA 02111

³ To whom correspondence should be addressed. E-mail: parke.wilde{at}tufts.edu.

	ABSTRACT

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
This study examined the relation between household food security status and current measured weight and change in self-reported weight over 12 mo using data from the 1999–2000 and 2001–2002 National Health and Nutrition Examination Surveys. Current measured BMI categories were as follows: underweight (<18.5 kg/m²), overweight (25 kg/m²), and obese (30 kg/m²). Change in self-reported weight used 2 cut-off points, i.e., a gain/loss of at least 2.27 kg (5 lb) and at least 4.54 kg (10 lb). Household food security categories were as follows: fully secure, marginally secure, insecure without hunger, and insecure with hunger. Multivariate analyses were adjusted for race/ethnicity, household income, education level, and current health status. Compared with women in households that were fully food secure, women in households that were marginally food secure [odds ratio (OR) 1.58] and food insecure without hunger (OR 1.76) were significantly more likely to be obese. Compared with women in households that were fully food secure, those in households that were marginally food secure were significantly more likely to gain at least 4.54 kg (OR 1.68). Compared with men in households that were fully food secure, men in households that were marginally food secure were more likely to be obese and to gain at least 4.54 kg, but these effects were smaller in magnitude than those for women and insignificant in some specifications. This study corroborates previous cross-sectional associations between intermediate levels of food insecurity and obesity for women, and it finds an association between intermediate levels of food insecurity and 12-mo weight gain for women.

KEY WORDS: • food security • weight change • weight status • obesity

An apparently contradictory relation exists between weight and income in the United States. Although hunger has traditionally been associated with limited access to food and low weight (1–3), obesity and overweight are actually higher among some low-income populations (4–7). Recently, researchers suggested that inconsistent access to resources may be partially responsible for the increased prevalence of obesity among individuals in low-income households (1,7–10). One manifestation of inadequate access to resources is food insecurity, or the lack of access to enough food for household members at all times in socially acceptable ways (11). Using various measures of food security/sufficiency and definitions of overweight and obesity, some, but not all researchers reported that some subpopulations of low-income individuals in food-insecure households have increased rates of overweight and obesity.

Those who found positive associations between food insecurity/insufficiency and overweight/obesity reasoned that such a relation may be due to the inconsistent availability of food in food-insecure households (5–8). Researchers suggested conceptual models in which food insecurity leads to a disordered eating pattern of underconsumption, limited consumption, or substitution of foods when resources are constrained, and compensatory overconsumption when resources are adequate, which in turn leads to adiposity (7,8). Studies linking food insecurity/insufficiency and overweight/obesity cited past research documenting overconsumption after periods of deprivation in humans (12–14) as a reason for the associations (7–10).

The link between food insecurity and overweight/obesity has been studied extensively, but with mixed results, depending on the data source and subpopulation considered. Additionally, because the observed relations are based on cross-sectional data sets, causality has been difficult to establish. To account for factors that might influence both weight and food security status, researchers controlled for characteristics such as income, race, and education, but other variables not measured by the survey cannot be controlled in cross-sectional analyses. Until recently, there were no nationally representative surveys that provided longitudinal information on both weight and food security status.

To our knowledge, no studies have focused on the association between food security status and change in weight over time. Fortunately, the National Health and Nutrition Examination Survey (NHANES), which collects both measured and self-reported height and weight information, including self-reported information about past weight, began collecting household food security information in its 1999–2000 data collection period. The 1999–2000 and 2001–2002 NHANES data sets provide information about household food security status over the 12 mo preceding the survey, as well as current weight and weight 12 mo preceding the survey. These data permit analysis of change in weight for the same time frame over which household food security status was measured.

	SUBJECTS AND METHODS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
    Sample and sampling frame. To consider the effect of household food security status on weight and change in weight, we examined data from the 1999–2000 and 2001–2002 NHANES collection periods. NHANES is a nationally representative survey that is conducted on an ongoing basis by the National Center for Health Statistics of the CDC. NHANES is a multistage, stratified survey that samples the noninstitutionalized civilian population (15). Data are collected throughout the year and released in 2-y data sets. For this analysis, we pooled the 1999–2000 and 2001–2002 data sets, and compared outcomes for women and men 18 y old by household food security status.

    Survey measures. Household food security status was based on the NHANES household food security category designation, which was determined using the 18-item U.S. Food Security Survey Module. Only one respondent per household answered the Module, regardless of the number of families living in a household. Screening into the module depended on responses to previous questions, and those households screened out were assumed to have negative responses to all Module questions. Food security questions referred to circumstances over the 12 mo preceding the survey. For example, the survey asks: "In the last 12 mo, did you ever eat less than you thought you should because there wasn't enough money to buy food?" (15). Another question asks about the quality of meals: "I/we couldn't afford to eat balanced meals." The NHANES category "fully food secure" represents no food-related hardship. There are 2 intermediate categories of food-related hardship, i.e., marginally food secure (at least 1 positive answer), and food insecure without hunger (15). The most severe category of food-related hardship is food insecure with hunger. For much of our analysis, we compare the 2 intermediate categories with the fully food secure category.

Current measured height and weight were obtained in NHANES mobile examination centers. Self-reported current and past weight were obtained from the Weight History Questionnaire. Demographic information for gender, race/ethnicity, highest education level, and income were obtained from the Demographic Questionnaire. Current health status was obtained from the Current Health Status Questionnaire.

    Cross-sectional comparisons. For comparisons with the earlier literature, we first conducted bivariate cross-sectional comparisons by household food security status and gender for measured BMI, percentage overweight (BMI 25 kg/m²), percentage obese (BMI 30 kg/m²), and percentage underweight (BMI <18.5 kg/m²), using CDC standards to define our weight categories (16). These analyses used observations with non-missing values for food security status and measured BMI (n = 5080 women, 4618 men).

Cross-sectional multivariate logistic regression analysis controlled for education level, race/ethnicity, household income, and current health status. Some of the control variables had missing values; thus, the multivariate analysis used and reported a somewhat smaller sample size than did the simple comparisons. The nominal race/ethnicity variable categories were as follows: Mexican American, Other Hispanic, Non-Hispanic White, Non-Hispanic Black, and Other Race, including Multi-Racial. For the ordinal education variable, the respondents' highest level of education was measured as follows: high school, high school diploma (including GED), andhigh school. An income:poverty ratio for each family was used to create a variable to control for family income; an income to poverty value 5.00 was topcoded as 5. The variable was broken into 7 categories based on the federal poverty level: 0–50%; 50.01%–100%; 100.01%–130%; 130.01%–200%; 200.01%–300%, 300.01%–500%, and >500%. Current health status was counted as "not good" if the individual responded positively to any of 3 survey questions about head/chest colds, stomach/intestinal illness, or flu/pneumonia. Otherwise, current health status was counted as "good." Results for an alternate specification without the current health status variable were almost exactly the same as the main multivariate estimates, and hence are not reported. All control variables were treated as class variables in the multivariate models. This analysis used observations with non-missing values for measured BMI, food security status, and the control variables (n = 4549 women, 4202 men).

    Changes in weight over time. The second set of analyses compared change in self-reported weight over the past year by household food security status and gender using 2.27 kg (5 lb) and 4.54 kg (10 lb) as cut-off points, due to data heaping at those points. The 4.54-kg cut-off is sufficiently large to indicate physiologically relevant gains and losses. A long-term weight gain of 5 kg is associated with an increased risk of type 2 diabetes (17), stroke (18), and coronary heart disease (19). If maintained, a single 1-y weight gain of 4.54 kg has the potential to increase the risk for these diseases. If the smaller 2.27 kg weight gain was sustained for several years, the cumulative gain would be substantial, leading to increased risk of disease over a lifetime.

The accuracy of self-reported weight is an important consideration in this second group of analyses. Errors in self-reports are systematic instead of random, reflecting both rounding to the nearest point of heaping and a tendency to report weights closer to ideal weight (20). In a review of the literature, no previous analyses of the accuracy of change in self-reported weight were found; thus, it is not known whether errors in changes over time are smaller or larger than errors in weight at a point in time. If respondents tended to understate their weight by a fixed amount, the change in weight would be more accurate than self-reported weight at a point in time. Because the NHANES provides both measured and self-reported current weight, we identified good reporters (the 79.6% of sample whose current weight reporting error <4.54 kg) and very good reporters (the 54.8% of sample whose current weight reporting error <2.27 kg). The main tabulations for weight change restricted the sample to good reporters (n = 4005 women, 3671 men).

Alternate specifications were used to check the robustness of the main results. One supplementary tabulation restricted the sample to very good reporters (n = 2910 women, 2375 men; Supplemental Table 1). Another alternate specification estimated results for each race/ethnicity group separately, to permit a complete set of interactions and allow the effect of food security on weight change to vary by race/ethnicity category. Because the sample size proved insufficient for several of the race/ethnicity categories, supplementary tabulations compared estimates only for Non-Hispanic White and all other categories combined.

Multivariate logistic regression analysis controlled for race/ethnicity, education, family income, current health status, as in the cross-sectional analysis. In addition, the analysis of weight change controlled for self-reported weight 1 y earlier. The analysis used good reporters with non-missing values for the control variables (n = 3569 women, 3337 men).

    Statistical analysis. Statistical analyses were conducted using SAS procedures surveymeans, surveyreg, and surveylogistic (Statistical Analysis System, Release 9.2 for Windows, SAS Institute). The 1999–2000 and 2001–2002 NHANES data sets were merged. All analyses were weighted using the NHANES mobile examination center sampling weights. The Taylor Series expansion method was used to adjust SE and CI for complex survey design.

	RESULTS

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
    Household food security status. The majority of respondents lived in fully food-secure households (84.1% of women, 85.2% of men). Another 6.2% of women and 4.8% of men lived in households that were marginally food secure. The sum of the estimated prevalences for these 2 food-secure categories is 1% higher than the official estimate for the national prevalence of household food security (88.9% in 2002), published by the Economic Research Service of the USDA. In the NHANES sample, 3.6% of women and men lived in households that were food insecure with hunger (Table 1).

View larger version (15K): [in this window] [in a new window]   FIGURE 1  Prevalence of being obese (A) and gaining 4.54 kg over 1 y (B) by food security status for women and men, NHANES 1999–2002. Food security status: (1) fully secure, (2) marginally secure, (3) insecure without hunger, (4) insecure with hunger. Error bars indicate ± 1.96 SE (95% CI) corrected for complex survey design.

In the supplementary analysis using only very good reporters, for women, the results for the prevalence of a 4.54-kg weight gain were very similar to those reported above, as were the results for the prevalence of 2.27 kg weight gain, except for a reduced magnitude and statistical significance for women in households that were food insecure without hunger. For men, the prevalence of a 2.27-kg weight gain and a 4.54-kg weight gain declined for households that were marginally food secure and increased for households that were food insecure without hunger, in comparison with the results above, leaving unchanged the overall pattern of higher prevalences for men in households with intermediate levels of food insecurity (Supplemental Table 1).

In the supplementary analysis with separate estimates for Non-Hispanic White individuals and other individuals, for both men and women, the appearance of the inverted U shape in prevalences of weight gain as a function of increasing food insecurity was most pronounced for the Non-Hispanic White individuals. For men and women in other race/ethnicity categories, the prevalence of weight gain was lowest in households that were fully food secure, significantly higher in households that were marginally food secure, and remained equally high all the way through to households that were food insecure without hunger (Supplemental Tables 2–3).

	DISCUSSION

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 
This analysis confirms that obesity rates appear highest for women in households with intermediate levels of household food insecurity. Similarly, it finds that rates of weight gain in a 12-mo period appear highest for women in households with intermediate levels of household food insecurity during the same period.

For women, the rate of obesity was comparatively low in fully food-secure households, rose as food insecurity worsened, and reached a peak in food-insecure households without hunger (Fig. 1A). This pattern persisted when we controlled for income, education, race/ethnicity, and current health status. This result is qualitatively similar to existing research. Olson reported that food insecurity was associated with increased BMI in women and school-age children in upstate New York (9). Townsend et al. (7) found a higher rate of obesity in women with mild food insecurity compared with food-secure women, as did Adams et al. (10).

Similarly, the prevalence of an increase in self-reported weight during a 12-mo period appeared highest in households that experienced an intermediate level of food insecurity during that period (Fig. 1B). For women, the percentage who gained at least 4.54 kg (10 lb) was 20.7% in households that were fully food secure and 34.6% in households that were marginally food secure. This pattern persisted when we controlled for income, education, race/ethnicity, current health status, and weight at the start of the 12-mo period.

Although firm causal claims cannot be made using our research design, our results for weight change strengthen the circumstantial evidence that intermediate levels of household food insecurity contribute to weight gain and risk of obesity. The earlier cross-sectional research essentially looks at variation across individuals in household food security status and associates it with variation across individuals in weight and obesity status. Many unobserved individual characteristics confound a causal interpretation of that association. By contrast, the present analysis looks at variation across individuals in household food security status and compares it with weight changes in those same individuals during the same period. Unobserved individual characteristics that are fixed over time no longer confound a causal interpretation of this association because such characteristics were the same at the start and end of the period for which weight change was measured. Unobserved characteristics that change over time may still confound the causal interpretation. This analysis reduces, but does not eliminate the scope for confounding variables to bias estimates of the effect of food security on body weight.

The literature offers several possible mechanisms by which intermediate levels of food insecurity could contribute to weight gain and a higher risk of obesity. For individuals in households with intermediate levels of food insecurity, gradual weight gain could occur from either inconsistent access to food, leading to periods of underconsumption followed by compensatory overconsumption (1,7–10,21,22), or from consuming inexpensive foods with high energy density when money is less available to spend on food (1,8,21–23). For individuals in households with the most severe level of food insecurity, it seems possible that these mechanisms for weight gain could be offset in part by the more direct energy deficits that one might expect in conditions of hunger.

This research may have implications for federal food assistance and nutrition programs, which spend $42 billion/y to improve nutrition and household food security for low-income Americans (24). Previous research suggests that the cyclic availability of food stamps affects the purchasing patterns of some households, and that for at least some populations, there is a monthly cycle of consumption that follows allocation of food stamps (25). Further research is required to establish that cyclic purchasing does indeed lead to cyclic consumption patterns, and that the cyclic consumption patterns lead to gradual weight gain over time. Such a clearly established association could guide policy for federal food and nutrition programs that have the goal of reducing hunger, food insecurity, and related health conditions (26). For instance, the Food Stamp Program, which allocates benefits on a monthly basis, might be modified to provide benefits more frequently to even distribution (21,25). Nutrition programs, such as the Expanded Food and Nutrition Education program, could also target more thoroughly the management of resources to stabilize consumption (26). Additionally, the current structure of agriculture subsidies provides little support for fruits and vegetables (27), making them relatively more expensive than other foods with a higher energy density (28). The structure of subsidies could be evaluated, perhaps with the goal of making fruits and vegetables more available through food assistance programs, which would mean that they would be more accessible to people in food-insecure households during periods in which resources are low.

	FOOTNOTES

 
¹ Supported in part by a grant from the National Poverty Center at the University of Michigan.

² Supplemental Tables 1–3 are available with the online posting of this paper at www.nutrition.org.

Manuscript received 27 October 2005. Initial review completed 6 December 2005. Revision accepted 7 February 2006.

	LITERATURE CITED

TOP ABSTRACT SUBJECTS AND METHODS RESULTS DISCUSSION LITERATURE CITED

 

1. Dietz WH. Does hunger cause obesity? Pediatrics. 1995;95:766–7.[Abstract/Free Full Text]

2. Kendall A, Olson CM, Frongillo EA Jr. Validation of the Radimer/Cornell measures of hunger and food insecurity. J Nutr. 1995;125:2793–801.[Abstract/Free Full Text]

3. Rose D. Assessing food insecurity in the United States. USDA, Economic Research Service, Food and Consumer Economics Division No. 9706, Washington, DC; 1997.

4. Sobal J, Stunkard AJ. Socioeconomic status and obesity: a review of the literature. Psychol Bull. 1989;105:260–75.[Medline]

5. Stunkard AJ, Sorensen TI. Obesity and socioeconomic status—a complex relationship. N Engl J Med. 1993;329:1036–7.[Free Full Text]

6. Townsend MS, Murphy S, Peerson J, Rose D. Obesity in America: the role of income and related variables: summary report of the Socioeconomic Determinants of Overweight Status in the United States Project. Report prepared for the USDA, Economic Research Service, Washington, DC; 2000.

7. Townsend MS, Peerson J, Love B, Achterberg C, Murphy SP. Food insecurity is positively related to overweight in women. J Nutr. 2001;131:1738–45.[Abstract/Free Full Text]

8. Frongillo EA, Olson CM, Rauschenbach BS, Kendall A. Nutritional consequences of food insecurity in a rural New York State county. Discussion Paper no. 1120–97, Institute for Research on Poverty, University of Wisconsin-Madison, Madison, WI; 1997.

9. Olson CM. Nutrition and health outcomes associated with food insecurity and hunger. J Nutr. 1999;129:521S–4.[Medline]

10. Adams EJ, Grummer-Strawn L, Chavez G. Food insecurity is associated with increased risk of obesity in California women. J Nutr. 2003;133:1070–4.[Abstract/Free Full Text]

11. Nord M, Andrews M, Carlson S. Household food security in the United States, 2002. Food and Rural Economics Division, Economic Research Service 2003: Food Assistance and Nutrition Research Report No. (FANRR35).

12. Fisher JO, Birch LL. Restricting access to palatable foods affects children's behavioral response, food selection, and intake. Am J Clin Nutr. 1999;69:1264–72.[Abstract/Free Full Text]

13. Polivy J, Zeitlin SB, Herman CP, Beal AL. Food restriction and binge eating: a study of prisoners of war. J Abnorm Psychol. 1994;103:409–11.[Medline]

14. Polivy J. Psychological consequences of food restriction. J Am Diet Assoc. 1996;96:589–92.[Medline]

15. Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS). National Health and Nutrition Examination Survey General Data Release Documentation. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, 1999–2000, 2001–2002, cited 2005 October 26. Available from: http://www.cdc.gov/nchs/about/major/nhanes/datalink.htm.

16. Body Mass Index. BMI for Adults. Atlanta, GA: Centers for Disease Control and Prevention, U.S. Department of Health and Human Services; updated 2004 Dec 16; cited 2005 Oct 26. Available from: http://www.cdc.gov/nccdphp/dnpa/bmi/bmi-adult.htm.

17. Colditz GA, Willett WC, Rotnitzky A, Manson JE. Weight gain as a risk factor for clinical diabetes mellitus in women. Ann Intern Med. 1995;122:481–6.[Abstract/Free Full Text]

18. Rexrode KM, Hennekens CH, Willett WC, Colditz GA, Stampfer MJ, Rich-Edwards JW, Speizer FE, Manson JE. A prospective study of body mass index, weight change, and risk of stroke in women. JAMA. 1997;277:1539–45.[Abstract/Free Full Text]

19. Willett WC, Manson JE, Stampfer MJ, Colditz GA, Rosner B, Speizer FE, Hennekens CH. Weight, weight change, and coronary heart disease in women. Risk within the ‘normal’ weight range. Ann Intern Med. 1998;128:81–8.[Abstract/Free Full Text]

20. Rowland ML. Self-reported height and weight. Am J Clin Nutr. 1990;52:1125–33.[Abstract/Free Full Text]

21. Gibson D. Food Stamp Program participation is positively related to obesity in low income women. J Nutr. 2003;133:2225–31.[Abstract/Free Full Text]

22. Alaimo K, Olson CM, Frongillo EA Jr. Food insufficiency and American school-aged children's cognitive, academic, and psychosocial development. Pediatrics. 2001;108:44–53.[Abstract/Free Full Text]

23. Sarlio-Lahteenkorva S, Lahelma E. Food insecurity is associated with past and present economic disadvantage and body mass index. J Nutr. 2001;131:2880–4.[Abstract/Free Full Text]

24. Economic Research Service Briefing Room. Food and Nutrition Assistance Programs. Washington, DC; updated 2005 may 24- [cited 2005 October 26]. Available from: http://www.ers.usda.gov/Briefing/FoodNutritionAssistance/.

25. Wilde PE, Ranney CK. The monthly Food Stamp cycle: shopping frequency and food intake decisions in an endogenous switching regression framework. Am J Agric Econ. 2000;82:200–13.

26. U.S. Department of Agriculture, Food and Nutrition Service. Promoting healthy eating: an investment in the future. Report to Congress. 1999. Available from: http://www.fns.usda.gov/oane/MENU/Published/NutritionEducation/Files/nutedsum.htm.

27. Fields S. The fat of the land: do agricultural subsidies foster poorer health? Environ Health Perspect. 2004;112:A820–3.[Medline]

28. Drewnowski A, Darmon N. The economics of obesity: dietary energy and energy cost. Am J Clin Nutr. 2005;82:265S–73.[Abstract/Free Full Text]

This article has been cited by other articles:

				R. Widome, D. Neumark-Sztainer, P. J. Hannan, J. Haines, and M. Story Eating When There is Not Enough to Eat: Eating Behaviors and Perceptions of Food Among Food-Insecure Youths Am J Public Health, May 1, 2009; 99(5): 822 - 828. [Abstract] [Full Text] [PDF]

				X. Gao, T. Scott, L. M Falcon, P. E Wilde, and K. L Tucker Food insecurity and cognitive function in Puerto Rican adults Am. J. Clinical Nutrition, April 1, 2009; 89(4): 1197 - 1203. [Abstract] [Full Text] [PDF]

				S. Isanaka, M. Mora-Plazas, S. Lopez-Arana, A. Baylin, and E. Villamor Food Insecurity Is Highly Prevalent and Predicts Underweight but Not Overweight in Adults and School Children from Bogota, Colombia J. Nutr., December 1, 2007; 137(12): 2747 - 2755. [Abstract] [Full Text] [PDF]

				R. C. Whitaker and A. Sarin Change in Food Security Status and Change in Weight Are Not Associated in Urban Women with Preschool Children J. Nutr., September 1, 2007; 137(9): 2134 - 2139. [Abstract] [Full Text] [PDF]

				V. Tarasuk, L. McIntyre, and J. Li Low-Income Women's Dietary Intakes Are Sensitive to the Depletion of Household Resources in One Month J. Nutr., August 1, 2007; 137(8): 1980 - 1987. [Abstract] [Full Text] [PDF]

				P. E. Wilde Measuring the Effect of Food Stamps on Food Insecurity and Hunger: Research and Policy Considerations J. Nutr., February 1, 2007; 137(2): 307 - 310. [Abstract] [Full Text] [PDF]

				M. Nord and L. S. Kantor Seasonal Variation in Food Insecurity Is Associated with Heating and Cooling Costs among Low-Income Elderly Americans J. Nutr., November 1, 2006; 136(11): 2939 - 2944. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Online Supporting Material Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wilde, P. E. Articles by Peterman, J. N. Search for Related Content PubMed PubMed Citation Articles by Wilde, P. E. Articles by Peterman, J. N.