| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Department of Nutrition, University of California, Davis (LK and MT); the Department of Human Nutrition, Ohio State University, Columbus (HQ); the Department of Nutritional Sciences, University of California, Berkeley (PC); and the University of California Cooperative Extension, Contra Costa County, Pleasant Hill, CA (MF)
2 Supported by the University of California Division of Agriculture and Natural Resources (DANR) and the University of California, Davis, Economic Research Service Small Grants Program. 3 Address reprint requests to L Kaiser, Department of Nutrition, University of California, One Shields Avenue, Davis, CA 95616-8669. E-mail: llkaiser{at}ucdavis.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: The goal was to examine the relation of food insecurity to weight status in low-income Latino women.
Design: A cross-sectional survey was conducted among 559 low-income Latino women selected by convenience sampling. The survey included the 18-item US Household Food Security Scale, 2 items related to current and past food insufficiency, demographic information, and measured heights and weights. Data were collected between February and May 2001 in 6 California counties. The main outcomes were frequency of overweight and obesity, defined by a body mass index (in kg/m2) of 25–29.9 (overweight) and >30 (obese). Data analysis included analysis of variance, Mantel-Haenszel chi-square test, and logistic regression.
Results: The prevalence of food insecurity was 50–60%, and that of obesity was 37.4%. Controlling for years spent in the United States, per capita income, and parity, food insecurity with hunger, measured by the 10-item adult scale of the Food Security Scale, was significantly related to obesity (OR: 1.98; 95% CI: 1.14, 3.53). No interaction between years spent in the United States and current food insecurity was observed. Current food insufficiency, as measured by a single item, was not related to obesity. However, severe past food insufficiency was related to obesity in the US-born population only.
Conclusion: Food insecurity appears to be related to obesity in Latino women, but choice of instruments might influence the results.
Key Words: Food insecurity • food insufficiency • overweight • obesity • Latinos • low-income status
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The prevalence of overweight and obesity in the United States is greatest among persons of low socioeconomic status, who are the most likely to experience food insecurity (5, 6). Food choices or physiologic adaptations in response to episodic food shortages might increase body fat (7). However, studies examining the relation between food insecurity and obesity in adults produced conflicting results. In a national sample of the US population, women who experienced mild or moderate food insufficiency were more likely to be overweight than were women who were food sufficient, with no relation being observed in men (5). In a Canadian national sample, men in food-insufficient households were less likely to be overweight than were food-sufficient men (4). Two studies, using different instruments in different populations, found that body mass index (BMI; in kg/m2) tended to peak at mild levels of food insecurity and then declined as food insecurity became more severe (8, 9). Food insecurity with hunger, assessed by a 4-item subset of the FSS, was associated with increased risk of obesity in Latino, Asian, and African American women but not in non-Latino white women (10). The differences among these studies could be due to the use of different instruments, definitions of excess body weight, and cultural factors that influence response patterns to the questions.
More research on food insecurity and nutrition among Latino households is needed. This rapidly growing segment of the US population reports relatively high levels of food insecurity (11, 12) and is more likely to be obese than are non-Latino white adults [prevalence of obesity: 39.7% in Mexican American women and 30.1% in white women (13)]. The goal of this research was to examine the relation of food insecurity to overweight and obesity in low-income Latino women.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The study was a cross-sectional survey of low-income Latino households, carried out between February and May 2001 in 6 California counties. Three counties included rural areas in the Central Valley (Tulare, San Joaquin, and Stanislaus), 2 were from urban areas (Contra Costa and Sacramento), and 1 was from a rural agricultural region on the coast (San Benito).
The survey included the following instruments: 1) the 18-item US Department of Agriculture FSS (14); 2) 2 household food insufficiency items during the mother's childhood and the 12 mo preceding the study; 3) a 171-item self-reported household food inventory (15); 4) a 66-item food-frequency questionnaire for the children (16); and 5) a 16-item family demographic record form that recorded data on the age and sex of household members; mother's education, birth country, preferred language, and years of residence in the United States; household participation in food assistance programs; and income. In addition, the interviewers measured weights and heights of all the mothers and children in the study (17). Neither the children's anthropometry nor the food-frequency questionnaire data are discussed here. The following steps were taken to improve the clarity of the instruments: focus group testing; pilot-testing among 20 low-income Latino families recruited through the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC); revision and back-translation from Spanish to English; and cross-checking of the final translated instrument by 8 native Spanish speakers (from Mexico, Central America, and South America).
Each of the 6 counties hired bilingual staff members to recruit and interview the subjects. All staff members attended a 3-d training and interview certification program to standardize data collection procedures across the 6 counties. Because of the sensitive nature of the food-security questions, a particular emphasis was placed on interviewing techniques and maintaining confidentiality. The project director or the principal investigator visited each trainee to observe interviews and provide additional feedback. The immediate supervisors in each county reviewed survey forms for completeness and consistency before submission for data entry. The original protocol called for a minimum sample size of 500. The final sample was 561 families, with complete data available for the analysis from 559.
For a woman to be eligible for this study, the following criteria had to be met: 1) the subject identified herself and her child as Latino, Mexican, Mexican American, Chicano, or Hispanic; 2) the subject had at least one biological child born between June 1, 1995, and February 1, 1998; 3) the family had an income 
200% of the poverty level; 4) the subject was aged 
18 y (or was an emancipated minor); and 5) the subject was willing to sign the Human Subjects consent form. A subject was excluded from the study for any of the following reasons: 1) she was currently pregnant or had been pregnant within the previous 3 mo; 2) she planned to move within the next 2 wk; and 3) she self-reported alcohol or substance abuse, mental illness, or other illness that could affect accurate responses, body weight, or diet. If the family had more than one eligible child aged 3–6 y, interviewers randomly selected the target child by coin toss.
The interviewers recruited a convenience sample of subjects from a variety of community-based agencies, including the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC); Head Start; migrant camps; the local public health department; local health centers; and family resource centers. Some interviews were conducted in the subjects' homes but most were conducted in private rooms in health or WIC clinics. All subjects signed a Human Subjects Consent form approved by the Institutional Review Board of the University of California, Davis, or the Committee for the Protection of Human Subjects at the University of California, Berkeley. Interview length was between 45 and 60 min. At the end of the interview, the subjects received a gift certificate worth $20 at a local store.
The interviewers used the previous 3 mo as the frame of reference for 18 items on the FSS, because the population was expected to be most vulnerable to food insecurity during the winter months roughly coinciding with this time frame. For reasons relevant to another analysis, questions 14 through 17 were specifically referenced to the target child in this study. Responses to the food insecurity items were coded and scaled according to procedures described by Bickel et al (14). The food insufficiency question "Which of these statements best describes the food eaten in your household?" used the previous 12 mo as a frame of reference to allow examination of the relation of food insufficiency to body weight over a longer period of time (current food insufficiency). A second version of the food insufficiency question was reworded to capture the level of food insufficiency during the mother's childhood (past food insufficiency): "When you were a child, which of the following best describes your situation?" In our sample, this measure of past food insufficiency was negatively correlated with mother's education (r = –0.35, P < 0.0001), which indicated its validity in reflecting past deprivation. For some of the analyses, an additional hunger category was created, on the basis of the past food insufficiency question, in which "enough food of the kinds wanted" and "enough food but not always the kinds wanted" were grouped under "no hunger during childhood," and "sometimes not enough" and "often not enough" were grouped under "hunger during childhood." The exact wording of all questions is given in Table 1
, which also specifies the questions that were included in each of the subscales.
|
The bilingual interviewers weighed and measured the participants according to standard procedures (17). The participants removed their shoes, outdoor clothing, and any hair accessories that might interfere with measurement. Height measurements were taken with use of portable stadiometers fitted with a headpiece (stadiometers from Center for Health Education, Concord, CA; headpieces from Genetec, South San Francisco, CA). The reading was taken to the nearest 1/8 inch (1 mm) and recorded immediately. Participants were weighed on an electronic scale (SECA model 810/815; SECA Corp, Columbia, MD). Scales were calibrated daily before each use and set to zero. Participants stood still with both feet in the center of the platform. The measurement was read to the nearest 0.10 pound and recorded immediately. Overweight [body mass index (BMI; in kg/m2) 25.0 to < 30.0] and obesity (BMI 30.0) were defined for the mother.
These data were analyzed by using SAS for WINDOWS software (Release 8.01; SAS Institute, Cary NC). Descriptive statistics were calculated for demographic, acculturation, anthropometric, and food insecurity variables. The relation of maternal BMI to food insecurity and demographic variables was examined by using the Pearson product-moment correlation for continuous variables and the Mantel-Haenszel chi-square test for ordered categorical variables. Interaction effects between past and current hunger on maternal BMI were examined by using analysis of variance (PROC GLM) after control for parity and acculturation. Logistic regression procedures (PROC LOGISTIC) were used in the multivariate analysis to examine factors related to risk of overweight and obesity (18). We report results as significant at a P value 0.05. Alpha levels between P > 0.05 and 0.10 were considered of marginal statistical significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Because the most severe items in our version of the 18-item FSS (ie, items 14–17) referred to a target preschooler, the frequency of affirmative responses to these items should be lower than would be expected if they had referred to any children, particularly older ones, in the household. Instead, both the English- and Spanish-speaking families were more likely to respond affirmatively to some of these child items than to the adult items. This response pattern was particularly evident for the households interviewed in Spanish. These findings suggest that the use of the standard 18-item FSS, which includes the 10 adult and 8 child items, will not provide food insecurity prevalence statistics that are comparable to national data (Mark Nord, personal communication, 2003). Therefore, the 10-item adult and household scale might be more appropriate than the full 18-item FSS in characterizing food conditions among the Latino women in this study.
Characteristics of the 559 subjects are given in Table 2. Food insecurity with or without hunger occurred among 
50–60% of the sample, depending on the choice of measurement scale. Approximately 40% of the subjects reported that, sometimes or often, not enough food was available in their households when they were children. Overweight and obesity were prevalent.
|
and Table 4. Whether the entire 18-item FSS or the 10-item adult subscale is used, fewer of the overweight or obese women were food secure (Table 3). However, significance was attained only with the use of the 10-item subscale. As a point of comparison, the 4-item subscale used in the recent California study by Adams et al (10) was included, which confirms that these items are representative of the 2 longer scales. In contrast, the question related to current food insufficiency yielded a nonsignificant relation with body-weight status (Table 4). As indicated in Tables 1 and 2, the current food insufficiency item categorizes fewer subjects as having adequate access to food than are so characterized by the multi-item food insecurity measures (food insufficiency item: 35% have enough of the types of food wanted; 10-item adult subscale: 49% are food secure).
|
|
). Specifically, past food insufficiency with hunger was significantly related to greater prevalence of obesity in the US-born Latino women but not in the Mexican-born women. A similar relation did not emerge when the data were analyzed by either main language spoken at home or language of the interview. Because some less acculturated women in our study opted to be interviewed in English "just to practice their English," the language of the interview might not fully reflect language proficiency. We did not find differences in the prevalence of overweight by level of past food insufficiency when the data were stratified by either language of the interview, main language spoken at home, or birth country (data not shown). Other acculturation variables, including total years spent in the United States and proportion of life spent in the United States, were negatively correlated with past food insecurity (r = –0.20, P < 0.0001 and r = –0.24, P < 0.0001, respectively) and positively correlated with maternal BMI (r = 0.10, P < 0.01 and r = 0.08, P < 0.05, respectively). Thus, the acculturation variables reflect potentially confounding effects.
|
Food insecurity with hunger was significantly related to greater likelihood of maternal obesity in both the unadjusted and adjusted models, as shown in Table 6. However, food insecurity without hunger was not significantly related to maternal obesity, with the use of either the 10-item adult FSS subscale or the 4-item subscale. Food insecurity with or without hunger was not significantly related to risk of overweight in the unadjusted or adjusted models (data not shown).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Focus group research among low-income Latino families in California found that mild food insecurity is especially common during the winter (20). In these households, women reduce their own food intakes in an attempt to spare both the adult males and young children from experiencing hunger. Food insecurity in Latino households is associated with significantly lower household supplies of fruit and vegetables (19, 21). In times of food insecurity, Latino households of Mexican descent rely heavily on traditional foods, including corn tortillas, beans, rice, potatoes, eggs, tomatoes, and chili peppers (19). Matheson et al (21) also documented that periodic food shortages occurring as payday approaches are associated with lower meat and energy intakes among school-aged children in low-income Latino households. Although not measured in that study, energy intakes of women in those households are presumably also affected by periodic shortages of food. A cyclical or even seasonal pattern of involuntary food restriction could cause these women to overeat at a later time when more money is available to purchase food. An alternative explanation for a relation between food insecurity and obesity is that obese women apply different criteria than do normal-weight women in judging the adequacy of their food supplies. However, we found no evidence of an interaction between maternal obesity and household food inventory score that would support that hypothesis. Others have documented that immigration to the United States is associated with many other dietary and lifestyle changes that also increase the risk of obesity and chronic disease (22-24).
The use of different instruments for estimating food insecurity and food insufficiency in the same population can lead to different conclusions about an association with obesity. Although both measures used here are correlated with household food supplies in low-income Latino households (19), the measure of current food insufficiency classifies fewer households as having adequate food supplies than do the food insecurity scales, which are referenced to the previous 3 mo. Because this study was carried out during the winter, when many Latino families experience limited food access (20), the reverse should be true. One possibility for discrepancies between the instruments is that the distinction between "quantity" and "quality" of food is not as clear to the respondent when only a 1-item measure is used as it is when a multi-item scale is used. For these Latino households, having to rely on a monotonous diet of traditional Mexican foods might not be perceived as "sufficient" or "enough" food even when the quantities consumed approximate energy needs (20).
That past food insufficiency is significantly related to obesity only in the US-born Latino women only also suggests that differences might exist in the cultural interpretation of the questions. Alternatively, differences between strategies to cope with hunger and the environmental context during childhood in Mexico and in the United States could account for these differences. In either case, our retrospective measure of past food insufficiency does not perform better than other measures of current food insecurity in detecting a relation between adequacy of access to food and weight status in the total population. Improved methods and instruments to measure the level of past deprivation might be able to find a stronger relation between childhood experiences of hunger and adult weight.
This study had both strengths and limitations to consider in interpreting the data. A strength of our study was the use of measured weights and heights, which reduced random and systematic errors associated with reported weights and heights. A limitation is our use of a Spanish version of the 18-item FSS that preceded publication of the official translation (25). However, our version is similar to that translation and was tested for clarity in the target group. Higher consistency of responses among participants interviewed in Spanish rather than English could be because some who preferred to be interviewed in English might not have been truly proficient in the language. Because a convenience sample was recruited for this study, caution is needed in generalizing the findings to the larger population. However, regardless of the convenience sampling procedure, a main point of our article is that the choice of instruments can influence the association of food insecurity and obesity in a given population.
Although our study found an association between greater food insecurity and obesity, we cannot conclude that a cause-and-effect relation exists. Because food insecurity and obesity are not rare conditions in this population, finding an association between the 2 is not necessarily surprising. Improved retrospective measures and longitudinal study designs might be able to provide support for such a relation. When evaluating the nutritional consequences of inadequate access to food or the effect of food assistance and education interventions, researchers should carefully consider their choice of instruments and the acculturation level of the subjects.
|
ACKNOWLEDGMENTS |
|---|
MT, PC, and MF wrote the proposal submitted to the University of California Division of Agriculture and Natural Resources. HM and LK wrote the proposal funded by the Economic Research Service Small Grants Program. LK and HM analyzed the data. All authors participated in the design and conduct of the survey and in data interpretation. At the time of the research, all authors were in positions funded through the University of California Cooperative Extension. LK administered the University of California, Davis, Economic Research Service Small Grants Program.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. S Townsend, G. J Aaron, P. Monsivais, N. L Keim, and A. Drewnowski Less-energy-dense diets of low-income women in California are associated with higher energy-adjusted diet costs Am. J. Clinical Nutrition, April 1, 2009; 89(4): 1220 - 1226. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Chilton, M. M. Black, C. Berkowitz, P. H. Casey, J. Cook, D. Cutts, R. R. Jacobs, T. Heeren, S. E. de Cuba, S. Coleman, et al. Food Insecurity and Risk of Poor Health Among US-Born Children of Immigrants Am J Public Health, March 1, 2009; 99(3): 556 - 562. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A.-A. Lyons, J. Park, and C. H. Nelson Food Insecurity and Obesity: A Comparison of Self-Reported and Measured Height and Weight Am J Public Health, April 1, 2008; 98(4): 751 - 757. [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]
|
|
|
|
|
|
K. L. Hanson, J. Sobal, and E. A. Frongillo Gender and Marital Status Clarify Associations between Food Insecurity and Body Weight J. Nutr., June 1, 2007; 137(6): 1460 - 1465. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. McLaren Socioeconomic Status and Obesity Epidemiol. Rev., May 2, 2007; (2007) mxm001v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. K. Landy Children who had experienced family food insufficiency were more likely to be overweight at 4.5 years of age Evid. Based Nurs., April 1, 2007; 10(2): 58 - 58. [Full Text] [PDF]
|
|
|
|
 |
|
 R. C. Whitaker and S. M. Orzol Obesity Among US Urban Preschool Children: Relationships to Race, Ethnicity, and Socioeconomic Status Arch Pediatr Adolesc Med, June 1, 2006; 160(6): 578 - 584. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
