Abstract
Objectives. We used self-reported and measured height and weight data to examine the relationship between food insecurity and obesity.
Methods. We defined food insecurity according to 3 different models. We used self-reported and measured height and weight from 2 versions of the Canadian Community Health Survey to calculate obesity rates.
Results. When self-reported height and weight data were used in calculating obesity prevalence rates, rates were significantly higher among food-insecure respondents than among food-secure respondents; by contrast, when measured height and weight data were used, there were no significant differences. Female respondents classified as food insecure and experiencing mild hunger were at greater risk of obesity than were food-secure female respondents when measured height and weight were used.
Conclusions. Associations between obesity and food insecurity are more pronounced when self-reported data on height and weight are used than when measured height and weight data are used. Caution should be used when using self-reported data to examine the relationship between food insecurity and obesity.
Food security is recognized as an important determinant of health,1,2 and both food insecurity and obesity are emerging public health concerns. The relationships among inadequate nutrient intake, poor health, and food insecurity are well established in the literature.3–19 However, an evolving body of research investigating the association between body mass index (BMI; weight in kilograms divided by height in meters squared)—obesity in particular—and food insecurity has revealed contradictory findings in Canada, the United States, and other countries.2,8,12,20–22
For example, an analysis of 1998–1999 Canadian National Population Health Survey data revealed that residents of food-insecure households were 1.5 times more likely than residents of food-secure households to be obese (i.e., BMI ≥ 30 kg/m2) when age, gender, and income were taken into account.3 By contrast, in a later examination of food insufficiency, Vozoris and Tarasuk,17 using data from the 1996–1997 version of the same Canadian survey, found no significant association between household food insufficiency and BMI among female respondents and significantly decreased odds of overweight (i.e., BMI of 25.0–29.9) among male respondents after they controlled for age, education, and income adequacy.
Similarly conflicting findings have been uncovered in US studies. For instance, research involving a representative sample of US households revealed that, among female respondents, food insecurity was positively associated with being overweight.16 This finding was supported by another US study showing that women in food-insufficient households were more likely than women in food-sufficient households to be overweight and to consume poor diets.21 Conversely, in a study of adults residing in 2 US states, Laraia et al. concluded that there was no association between overweight and obesity with concerns about food sufficiency after they controlled for education, income, race/ethnicity, marital status, and general health.22
Further contradictory findings from an analysis of the Third National Health and Nutrition Examination Survey demonstrated that younger girls (aged 2 to 7 years) from all racial/ethnic groups who lived in food-insufficient households were at lower risk of being overweight than were younger girls who lived in food-sufficient households, whereas the opposite pattern was found among older (aged 8 to 16 years) non-Hispanic White girls.15 It is clear from the discrepancies just described that differences between studies in measurements of food insecurity and food insufficiency, definitions of overweight and obesity, and choices of confounding variables result in both confusion and complexity in terms of delineating the relationship between food insecurity and BMI.
The majority of previous research has relied on self-reported measures of height and weight to examine BMI and food insecurity.3,14,15,20,22 To our knowledge, no research in Canada to date has employed national data on measured height and weight to examine the relationship between food insecurity and obesity. Recent trends show that rates of obesity are higher when measured data on height and weight are used in calculating obesity than when self-reported height and weight are used.23 In addition, associations derived from most previous research involving self-reports (the Townsend et al.16 study is an exception) do not account for underreporting of weight or overreporting of height.
Given the potential consequences of both food insecurity and obesity on individual health and related health care system costs, it is important to understand how obesity relates to food insecurity. Therefore, we sought to provide a unique look at the association between obesity and food security by assessing data on measured height and weight as well as comparative data on self-reported height and weight.
METHODS
Data Sources
Data were derived from the Canadian Community Health Survey (CCHS), cycle 1.1 (CCHS 1.1; 2000–2001) and cycle 2.2 (CCHS 2.2; 2004). CCHS 1.1 was a general health survey of 135 535 individuals aged 12 years or older who were living in private households; residents of Indian reserves, Canadian Armed Forces bases, health care institutions, and some remote areas were excluded. The overall response rate was 85%. In this analysis, we excluded information from the 3 Canadian territories so that the data would be comparable with CCHS 2.2 data; as a result, the sample was reduced to 128 500 individuals.
CCHS 2.2, a survey involving 35107 individuals of all ages, was specifically designed to gather information about the nutritional status of the Canadian population. It did not include residents of the 3 Canadian territories, Indian reserves, and some remote areas or regular members of the Canadian Armed Forces. The overall response rate was 77%. In an effort to ensure that the data were comparable with those of CCHS 1.1, we included only people aged 12 years or older with measured height and weight information, which reduced the total to 16527 individuals. Table 1 ▶ presents weighted distributions of sample characteristics for CCHS 1.1 and CCHS 2.2.
TABLE 1—
Selected Characteristics of Among Respondents Aged 12 Years or Older: CCHS 1.1, 2000–2001, and CCHS 2.2, 2004
| CCHS 1.1 | CCHS 2.2 | |||
| Sample Size | Estimated Population,a No, Thousands (%) | Sample Size | Estimated Population,a No, Thousands (%) | |
| Total | 128 500 | 25 635 (100.0) | 16 527 | 26 500 (100.0) |
| Gender | ||||
| Male | 59 364 | 12 619 (49.2) | 7 300 | 13 181 (49.7) |
| Female | 69 136 | 13 016 (50.8) | 9 227 | 13 319 (50.3) |
| Age, y | ||||
| 12–19 | 17 002 | 3 193 (12.5) | 4 977 | 3 400 (12.8) |
| 20–34 | 25 985 | 6 269 (24.5) | 2 617 | 6 085 (23.0) |
| 35–44 | 24 872 | 5 294 (20.7) | 1 544 | 5 106 (19.3) |
| 45–64 | 36 676 | 7 253 (28.3) | 3 896 | 8 139 (30.7) |
| ≥ 65 | 23 965 | 3 626 (14.1) | 3 493 | 3 770 (14.2) |
| Household income category | ||||
| Low | 5 470 | 881 (3.4) | 626 | 735 (2.8) |
| Lower middle | 11 772 | 1 760 (6.9) | 1 308 | 1 534 (5.8) |
| Middle | 28 243 | 5 111 (19.9) | 3 598 | 5 116 (19.3) |
| Upper middle | 40 474 | 8 145 (31.8) | 5 236 | 8 694 (32.8) |
| High | 28 626 | 7 039 (27.5) | 3 987 | 8 009 (30.2) |
| Missing | 13 915 | 2 699 (10.5) | 1 772 | 2 412 (9.1) |
| Educational level | ||||
| Less than high school | 43 124 | 7 463 (29.1) | 7 283 | 6 901 (26.0) |
| High school | 22 668 | 4 758 (18.6) | 2 405 | 4 555 (17.2) |
| Some college | 9 712 | 2 102 (8.2) | 1 358 | 2 322 (8.8) |
| College or greater | 51 785 | 11 102 (43.3) | 5 393 | 12 468 (47.1) |
| Missing | 1 211 | 211 (0.8) | 88 | 254 (1.0) |
| Race/ethnicity | ||||
| White | 116 597 | 21 832 (85.2) | 14 676 | 21 924 (82.7) |
| Aboriginal | 2 146 | 238 (0.9) | 483 | 290 (1.1) |
| Black | 970 | 423 (1.7) | 146 | 450 (1.7) |
| Other | 7 781 | 2 949 (11.5) | 1 220 | 3 791 (14.3) |
| Missing | 1 006 | 193 (0.8) | 2 | 44 (0.2) |
| Marital status | ||||
| Previously married | 24 023 | 3 105 (12.1) | 3 017 | 3 293 (12.4) |
| Never married | 37 133 | 7 630 (29.8) | 6 064 | 7 866 (29.7) |
| Married | 67 197 | 14 880 (58.0) | 7 438 | 15 315 (57.8) |
| Missing | 147 | 21 (0.1) | 8 | 25 (0.1) |
| Area of residence | ||||
| Urban | 94 690 | 20 959 (81.8) | 12 546 | 21 624 (81.6) |
| Rural | 33 810 | 4 676 (18.2) | 3 981 | 4 876 (18.4) |
Note. CCHS 1.1 = Canadian Community Health Survey, cycle 1.1; CCHS 2.2 = Canadian Community Health Survey, cycle 2.2. As a result of rounding, totals may not sum to 100%.
aEstimated household Canadian population aged 12 years or older (number of people represented by given sample after application of sample weights).
Measurement of Obesity
BMI was used in calculating obesity. In CCHS 1.1, self-reported height and weight data were collected from all survey respondents. In CCHS 2.2, height and weight measurements were conducted for 62% of survey respondents aged 12 years or older. With a special sample weight applied, the estimates for this group represented the Canadian population.
We used these data to calculate BMIs for all respondents aged 12 years or older, excluding pregnant respondents. A BMI cutoff of 30 kg/m2 was used to classify adults (aged 18 years or older) as obese, in accord with the health risks associated with classification in this BMI category.25 Because there is uncertainty regarding the BMI levels associated with health risks at younger ages, we used the age- and gender-specific cutoff points defined by Cole et al.26 to classify obesity among youths aged 12 to 17 years. Midyear age points were assigned to all reported years of age (e.g., 15.5 for 15-year-olds).
Measurement of Food Insecurity
We formulated 3 different models of food insecurity in an effort to provide an enhanced understanding of the relationship between food insecurity and obesity when self-reported data on height and weight are used as well as when measured height and weight data are used.
Dimensional model of food insecurity.
In CCHS 1.1, food insecurity was determined with 3 questions. Respondents were asked how often, in the past 12 months, they or anyone else in their household (1) were concerned that there would not be enough to eat because of a lack of money, (2) did not have enough food to eat because of a lack of money, or (3) did not eat the quality or variety of foods that they wanted to eat because of a lack of money. Response options for the 3 questions were “often true,” “sometimes true,” and “never true.”
The definition of food insecurity was subsequently refined via modification of the 3 questions just described.3,4 We used this adaptation (relating to CCHS 1.1 only) to identify 3 dimensions of food insecurity for this study. Respondents who replied affirmatively (i.e., “often true” or “sometimes true”) to the first question were considered to have food anxiety, and those who replied negatively (never true) were considered to have no food anxiety. Respondents who replied affirmatively to the second question were considered to be experiencing food poverty, and those who replied negatively were classified as not experiencing food poverty. Finally, respondents who replied affirmatively to the third question were considered to have compromised diets, whereas those who replied negatively were considered to not have compromised diets.
In CCHS 2.2, food insecurity was determined through a more elaborate probing process based on 18 questions that assessed overall household food security on a continuous linear scale referred to as the Food Security Module, adopted from the United States Department of Agriculture model of food security.27 The first 3 questions of the Food Security Module were used to develop 3 dimensions of food insecurity comparable to those delineated in CCHS 1.1. Respondents were asked whether, in the past 12 months, (1) they or other household members worried that food would run out before they got money to buy more; (2) the food that they and other household members bought didn’t last, and there wasn’t any money to get more; and (3) they or other household members couldn’t afford to eat balanced meals. Again, response options for the 3 questions were “often true,” “sometimes true,” and “never true.”
We used the following parameters to define the 3 dimensions of food insecurity relating to CCHS 2.2: respondents who replied affirmatively (i.e., “often true” or “sometimes true”) to the first question were considered to have food anxiety; respondents who replied affirmatively to the second question were considered to be in a state of food poverty; and respondents who replied affirmatively to the third question were considered to have compromised diets. Similar percentages of respondents answered affirmatively to at least 1 of the 3 questions in each of the 2 surveys: 14.6% in CCHS 1.1 and 12.1% in CCHS 2.2.
Levels model of food insecurity.
Collectively, the 18 Food Security Module questions used in CCHS 2.2 represent a more sensitive measure of food insecurity than the dimensional model with its discrete categories. The Food Security Module includes questions that provide a unique look at how food decisions involving children relate to different degrees of food insecurity and more in-depth insights into the severity of food insecurity. The instrument can be used to determine household food security status through coding responses to each question. Here we coded responses such as “yes,” “often true,” and “almost every month” as 1 (“affirmative”) and responses such as “no” and “never true” as 0 (“negative”). Resulting Food Security Module scale values ranged from 0 to 18 for households with children and from 0 to 10 for households without children. In our analysis, we classified Food Security Module scores into 4 household food security status groups, as follows.27
Food secure: These households showed no or minimal evidence of food insecurity (Food Security Module score of 0–2).
Food insecure without hunger: Food insecurity was evident in household members’ concerns about the adequacy of their food supply and in adjustments to household food management, including reduced quality of food and increased atypical coping patterns. Little or no reduction in household members’ food intake was reported (Food Security Module scores of 3–7 for households with children and 3–5 for households without children).
Food insecure with mild hunger: Food intakes for adult household members were reduced to an extent implying that these individuals had repeatedly experienced the physical sensation of hunger. In most (but not all) food-insecure households with children, such reductions are not observed among children at this stage (Food Security Module scores of 8–12 for households with children and 6–8 for households without children).
Food insecure with severe hunger: These households reduced children’s food intake to an extent that the children experienced hunger. Adults in households both with and without children repeatedly experienced more-extensive reductions in food intakes (Food Security Module scores of 13–18 for households with children and 9–10 for households without children).
Overall food insecurity model.
In CCHS 1.1, respondents who replied “often” or “sometimes” to at least 1 question were considered to have experienced food insecurity. Therefore, those who replied “never” to all 3 questions were categorized as food secure. In CCHS 2.2, respondents whose Food Security Module scores were greater than 2 were considered to reside in food-insecure households.
Sociodemographic Variables
Household income was defined according to total income from all sources in the 12-month period before the survey. Educational level was grouped into 4 categories: less than high school, high school, some post–high school education, and college degree. Marital status was classified as (1) single or never married; (2) divorced, separated, or widowed; or (3) married or in a common-law relationship.
Ethnicity was grouped into 4 categories (White, Aboriginal, Black, or other) on the basis of responses to a question on ethnic/ racial origin. On the basis of Statistics Canada’s geographical classifications for the 1996 census, area of residence was categorized as urban or rural. We defined urban areas as those with a minimum population of 1000 residents and a population density of 400 or more people per square kilometer. Areas that did not meet these criteria were classified as rural.
Data Analysis
We used weighted cross-tabulations to estimate obesity prevalence rates for all 3 models of food insecurity. Also, we conducted multivariate logistic regression analyses to determine the odds of being obese for individuals living in households characterized as food insecure. We controlled for confounding effects by including sociodemographic variables in each regression model. To account for survey design effects, we estimated the variance used in calculating coefficients of variation (i.e., standard errors of estimates) and confidence intervals using the bootstrap technique.28,29
RESULTS
In comparisons involving the dimensional model of food insecurity, obesity prevalence rates were significantly higher among individuals living in households with any of the dimensions of food insecurity—food anxiety, compromised diets, or food poverty—than among the overall study population when self-reported height and weight data were used (Table 2 ▶). Conversely, rates were not different on any of these dimensions when measured height and weight data were used. In an attempt to gain a more sensitive perspective on relationships between obesity and severity of food insecurity according to measured height and weight, we conducted additional analyses using the levels model of food insecurity but were able to do so only for the CCHS 2.2 data. The results of these analyses showed a significant association between obesity and food insecurity with mild hunger.
TABLE 2—
Obesity Prevalence Results, by Dimensional and Level Models of Food Insecurity and Overall, Among Respondents Aged 12 Years or Older: CCHS 1.1, 2000–2001, and CCHS 2.2, 2004
| Obesity Rate, % | ||
| Food Insecurity Model | Self-Reported Height and Weight | Measured Height and Weight |
| Dimensional model | ||
| Food anxiety | 19.0*** | 22.5 |
| Compromised diet | 19.3*** | 23.6 |
| Food poverty | 20.1*** | 23.8 |
| Levels model | ||
| Food insecurity without hunger | . . . | 19.3 |
| Food insecurity with mild hunger | . . . | 34.2* |
| Food insecurity with severe hunger | . . . | 30.6a |
| Overall model | 15.1 | 21.8 |
Note. CCHS 1.1 = Canadian Community Health Survey, cycle 1.1; CCHS 2.2 = Canadian Community Health Survey, cycle 2.2. See “Methods” section for information about how models were created.
aCoefficient of variation (standard error of the estimate) between 16.6% and 33.3%.
* P<.05; ***P<.001, for differences from total population.
Further gender-specific analyses using the overall food insecurity model classifications demonstrated that rates of obesity were significantly higher among respondents who were food insecure than among respondents who were food secure when self-reported height and weight data from CCHS 1.1 were used (Figure 1 ▶). However, when measured height and weight data from CCHS 2.2 were used to determine obesity, there were no significant differences in obesity according to food security status for either male or female respondents. These results are consistent with those shown in Table 2 ▶. The data in Figure 1 ▶ were based on the overall food insecurity model, and thus, food insecurity status reflected all 18 Food Security Module questions. In other words, the lack of association between food insecurity and obesity when measured data were used did not appear to be explained by the decision to use only 3 questions of the Food Security Module.
FIGURE 1—
Obesity prevalence rates among respondents aged 12 years or older by self-reported height and weight (a) and measured height and weight (b): CCHS 1.1, 2000–2001, and CCHS 2.2, 2004.
Note. CCHS 1.1 = Canadian Community Health Survey, cycle 1.1; CCHS 2.2 = Canadian Community Health Survey, 2.2; E = coefficient of variation (standard error of the estimate) between 16.6% and 33.3%. aFrom CCHS 1.1. bFrom CCHS 2.2. * P < .05; **P < .01; ***P < .001, for differences from food-secure households.
We further examined the association between obesity and overall food insecurity and dimensions and levels of food insecurity for both self-reported height and weight and measured height and weight while controlling for sociodemographic variables such as household income, education, ethnicity, marital status, and area of residence (Table 3 ▶). The overall association between food insecurity and obesity was significant for both male and female respondents in the case of self-reported but not measured data. No significant association between any dimension of food insecurity and obesity was evident among men when either self-reported or measured data were used.
TABLE 3—
Logistic Regression Analysis of Obesity Associated With 3 Different Models of Food Insecurity Among Respondents Aged 12 Years or Older: CCHS 1.1, 2000–2001, and CCHS 2.2, 2004
| Adjusted Obesity Odds Ratio (95% Confidence Interval) | ||
| Self-Reported Height and Weighta | Measured Height and Weightb | |
| Men | ||
| Overall food insecurity model | ||
| Yes | 1.22* (1.20, 1.34) | 1.23 (0.72, 2.07) |
| No (reference) | 1.00 | 1.00 |
| Dimensional food insecurity model, food anxiety | ||
| Yes | 1.03 (0.88, 1.22) | 1.30 (0.78, 2.17) |
| No (reference) | 1.00 | 1.00 |
| Dimensional food insecurity model, compromised diet | ||
| Yes | 1.12 (0.97, 1.29) | 0.92 (0.56, 1.52) |
| No (reference) | 1.00 | 1.00 |
| Dimensional food insecurity model, food poverty | ||
| Yes | 1.17 (0.96, 1.43) | 0.89 (0.45, 1.74) |
| No (reference) | 1.00 | 1.00 |
| Levels of food insecurity model | 1.02 (0.55, 1.88) | |
| Insecurity without hunger | . . . | |
| Insecurity with mild hunger | . . . | 1.77 (0.74, 4.27) |
| Insecurity with severe hunger | . . . | 1.18 (0.31, 4.48) |
| Food security (reference) | . . . | 1.00 |
| Women | ||
| Overall food insecurity model | ||
| Yes | 1.66* (1.53, 1.80) | 1.35 (0.90, 2.00) |
| No (reference) | 1.00 | 1.00 |
| Dimensional food insecurity model, food anxiety | ||
| Yes | 1.10 (0.95, 1.28) | 0.88 (0.52, 1.49) |
| No (reference) | 1.00 | 1.00 |
| Dimensional food insecurity model, compromised diet | ||
| Yes | 1.47* (1.31, 1.66) | 1.35 (0.81, 2.23) |
| No (reference) | 1.00 | 1.00 |
| Dimensional food insecurity model, food poverty | ||
| Yes | 1.19* (1.02, 1.41) | 1.38 (0.76, 2.52) |
| No (reference) | 1.00 | 1.00 |
| Levels of food insecurity model | ||
| Insecurity without hunger | . . . | 0.95 (0.60, 1.50) |
| Insecurity with mild hunger | . . . | 2.72* (1.43, 5.16) |
| Insecurity with severe hunger | . . . | 2.04 (0.74, 5.59) |
| Food security (reference) | . . . | 1.00 |
Note. CCHS 1.1 = Canadian Community Health Survey, cycle 1.1; CCHS 2.2 = Canadian Community Health Survey, cycle 2.2. Each model was adjusted for age, household income, education, ethnicity, marital status, and place of residence. Missing categories for household income, education, ethnicity, and marital status were included in each model to maximize the sample size. See “Methods” section for information about how models were created. Obesity defined as a body mass index (weight in kilograms divided by height in meters squared) of 30 kg/m2 or greater.
aIn CCHS 1.1.
bIn CCHS 2.2.
dReference category.
* P < .05, for differences from total population.
Among female respondents, no specific dimension of food insecurity was associated with obesity when measured height and weight were used. However, in general, higher odds of obesity were found among female respondents when self-reported height and weight were used. Specifically, when self-reported data on height and weight were used, female respondents who had a compromised diet or were experiencing food poverty were more likely to be classified as obese.
When we used the levels model of food insecurity, we obtained quite different results for obesity among female respondents. Interestingly, when measured height and weight were used, female respondents living in households classified as food insecure with mild hunger were almost 3 times more likely to be obese than were female respondents living in food-secure households. Household food insecurity in the absence of hunger or severe hunger was not associated with obesity among female respondents.
DISCUSSION
The fact that CCHS 1.1 and CCHS 2.2 did not include homeless people or Aboriginal peoples living on Indian reserves may have been problematic, given that they tend to be at higher risk of food insecurity and obesity than other groups.4,13,24 Furthermore, because analyses were based on cross-sectional data, no conclusions about cause-and-effect relationships can be made.
Previous research has shown that different instruments used to measure food insecurity can affect the relationship between food insecurity and obesity.20,23 Although there may be some justification for caution because of the use of different instruments to measure food insecurity in CCHS 1.1 and CCHS 2.2, we attempted to compensate for this limitation through the use of 3 different models of food insecurity to add clarity and strength to our conclusions.
Our comparative analysis between self-reported height and weight data from CCHS 1.1 and measured height and weight data from CCHS 2.2 provides new and meaningful information on the association between obesity and food insecurity. When self-reported height and weight were used, prevalence rates of obesity were significantly higher among food-insecure male and female respondents than among food-secure male and female respondents; by contrast, when measured height and weight were used, there were no significant differences between the 2 groups.
Although the questions used to measure food insecurity in CCHS 1.1 and 2.2 were similar, differences in wording may have contributed to varied interpretations by participants, altering some of our results. However, when Rowland’s equations30 were used to adjust the self-reported data from CCHS 1.1 to allow estimates of measured data from CCHS 2.2, associations between obesity and food insecurity appeared to be generally consistent in the 2 surveys (data not shown).
Although the sample sizes for most of our models were sufficiently large, those for household insecurity with severe hunger in the food security levels model had a relatively small sample (n = 170). This may have resulted in a failure to achieve statistically significant findings that would have emerged had the sample been larger.
Previous research has shown that self-reported height and weight data are unreliable.31 Moreover, studies have consistently shown that women underestimate their weight to a greater extent than do men and that discrepancies in self-reported weight are greater among overweight and obese individuals (again, more so for women).30,31 For example, Bostrom and Diderichsen31 found that when self-reported height and weight were used to determine BMI, overall, 81% of men and 78% of women were classified accurately, as compared with only 61% of obese men and 55% of obese women. This suggests that individuals, especially women, who are classified as obese according to measured height and weight may be classified as simply overweight or even as normal weight if self-reported height and weight are used. As a result of such discrepancies, the relationship between obesity and food insecurity observed when self-reported data are used may in fact represent a relationship between food insecurity and severe or morbid obesity.
Regardless of whether food insecurity was measured using 1 of the dimensions of food anxiety or compromised diet or the more extreme situation of food poverty, obesity prevalence rates were significantly higher among food-insecure individuals than among those who were food secure when self-reported data were used. When measured height and weight were used, most significant differences between food insecurity and obesity disappeared, but a more in-depth examination of the levels model reveals that female respondents who were food insecure with mild hunger were almost 3 times more likely to be obese than were female respondents living in households without food insecurity. This demonstrates the intricate effects of food insecurity on food consumption patterns.
The reason that the outcome just described was significant only among women may be that women are more likely than men to be “emotional eaters,” under stress, consuming increased amounts of sweet, fatty, and energy-dense foods.32 Furthermore, the finding that mild stress may be associated with increased eating and severe stress with decreased consumption might explain why obesity is more prevalent among those in households at mild hunger levels.13,17,33,34
In contrast to the female respondents, the male respondents in food-secure households were no more likely than were male respondents in food-insecure households to be obese when measured height and weight were used. Moreover, obesity among male respondents was not associated with any dimension of food insecurity even when self-reported height and weight were used. This finding may support evidence showing that food insecurity is experienced differently by different members of a household.35 Nutrient intake studies have shown that women may be the first to compromise their diet when the household is food insecure.6,10 Also, research has shown that food insufficiency is associated with the consumption of fewer than 5 fruits or vegetables per day among women.36
As a household’s concerns about food adequacy escalate, adjustments to food management take place, such as reduced quality of food purchased. One hypothesis suggests that decreased food purchasing power results in the purchasing of energy-dense foods and thus increased energy intakes.37 If a woman is more likely to compromise her own nutrient intake for the sake of other household members, then she may become more vulnerable to consumption of energy-dense, nutrient-poor foods and hence obesity. Moreover, these consumption patterns may be transferred to other family members.38 Child welfare laws may further contribute to women decreasing their food intake to ensure that their children’s food intakes are adequate, given that insufficient diets among children may suggest parental neglect.39
Until further research is completed, associations between obesity and food insecurity should be cautiously interpreted. All of the instruments used to measure food security status have changed over time, and to date, they have neglected to describe frequencies and durations of food insecurity, which are both likely to influence BMIs.40 Moreover, no tool has been validated to measure food security among Aboriginal populations in Canada.41 Various types of geographical and cultural differences exist and should be recognized and considered in measurements of food security.
The present results substantiate the need for further research that would provide a better understanding of how health is influenced by the association between BMI and food insecurity. Our findings suggest that a more rigorous instrument needs to be developed to measure food insecurity in unison with measured height and weight. Furthermore, longitudinal studies incorporating consistent measures of food security would aid in delineating the association between obesity and food insecurity.15,40
Peer Reviewed
Contributors A.-A. Lyons originated the study, J. Park was the primary statistician, and C. H. Nelson supervised all aspects of its implementation. All of the authors helped to conceptualize ideas, interpret findings, and write and review drafts of the article.
Human Participant Protection No protocol approval was needed for this study.
References
- 1.McIntyre L. Food insecurity. In: Raphael D, ed. Social Determinants of Health: Canadian Perspectives. Toronto, Ontario: Canadian Scholars Press Inc; 2004:173.
- 2.Public Health Agency of Canada. Food security as a determinant of health. Available at: http://www.phacaspc.gc.ca/phsp/phdd/overview_implications/08_food.html. Accessed June 29, 2007.
- 3.Rainville D, Brink S. Food Insecurity in Canada, 1998–1999. Ottawa, Ontario: Human Resources Development Canada; 2001.
- 4.Che J, Chen J. Food insecurity in Canadian households. Health Rep. 2001;12:11–22. [PubMed] [Google Scholar]
- 5.Kendall A, Olson C, Frongillo E. Relationship of hunger and food insecurity to food availability and consumption. J Am Diet Assoc. 1996;96:1019–1024. [DOI] [PubMed] [Google Scholar]
- 6.Rose D, Oliveira V. Nutrient intakes of individuals from food-insufficient households in the United States. Am J Public Health. 1997;87:1956–1961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Rose D. Economic determinants and dietary consequences of food insecurity in the United States. J Nutr. 1999;129:517–520. [DOI] [PubMed] [Google Scholar]
- 8.Tarasuk V, Beaton GH. Women’s dietary intakes in the context of food security. J Nutr. 1999;129:672–679. [DOI] [PubMed] [Google Scholar]
- 9.Klesges LM, Pahor M, Shorr R, Wan JY. Financial difficulty in acquiring food among elderly disabled women: results from the Women’s Health and Aging Study. Am J Public Health. 2001;91:68–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.McIntyre L, Glanville T, Raine K. Do low-income mothers compromise their nutrition to feed their children? Can Med Assoc J. 2003;168:686. [PMC free article] [PubMed] [Google Scholar]
- 11.Olson C. Nutrition and health outcomes associated with food insecurity and hunger. J Nutr. 1999;129: 521–524. [DOI] [PubMed] [Google Scholar]
- 12.Hamelin AM, Habicht JP, Beaudry M. Food insecurity: consequences for the household and broader social implications. J Nutr. 1999;129:525–528. [DOI] [PubMed] [Google Scholar]
- 13.McIntyre L, Connor SK, Warren J. Child hunger in Canada: results of the 1994 National Longitudinal Survey of Children and Youth. Can Med Assoc J. 2001; 163:961–965. [PMC free article] [PubMed] [Google Scholar]
- 14.Sarlio-Lahteenkorva S, Lahelma E. Food insecurity is associated with past and present economic disadvantage and body mass index. J Nutr. 2001;131: 2880–2884. [DOI] [PubMed] [Google Scholar]
- 15.Alaimo K, Olson C, Frongillo E. Low family income and food insufficiency in relation to overweight in US children: is there a paradox? Arch Pediatr Adolesc Med. 2001;155:1161–1167. [DOI] [PubMed] [Google Scholar]
- 16.Townsend MS, Peerson J, Love B, Achterberg C, Murphy SP. Food insecurity is positively related to overweight in women. J Nutr. 2001;131:1738–1745. [DOI] [PubMed] [Google Scholar]
- 17.Vozoris N, Tarasuk V. Household food insufficiency is associated with poorer health. J Nutr. 2003; 133:120–126. [DOI] [PubMed] [Google Scholar]
- 18.Stuff JE, Casey PH, Szeto KL, et al. Household food insecurity is associated with adult health status. J Nutr. 2004;134:2330–2335. [DOI] [PubMed] [Google Scholar]
- 19.Casey PH, Szeto KL, Robbins JM, et al. Child health related quality of life and household food security. Arch Pediatr Adolesc Med. 2005;159:51–56. [DOI] [PubMed] [Google Scholar]
- 20.Adams EJ, Grummer-Strawn L, Chavez G. Food insecurity is associated with increased risk of obesity in California women. J Nutr. 2003;133:1070–1074. [DOI] [PubMed] [Google Scholar]
- 21.Basiotis P, Lino M. Food insufficiency and prevalence of overweight among adult women. Fam Econ Nutr Rev. 2003;15:55. [Google Scholar]
- 22.Laraia BA, Siega-Riz AM, Evenson KR. Self-reported overweight and obesity are not associated with concern about enough food among adults in New York and Louisiana. Prev Med. 2004;38:175–181. [DOI] [PubMed] [Google Scholar]
- 23.Kaiser L, Townsend MS, Melgar-Quinonez HR, Fujii ML, Crawford PB. Choice of instrument influences relations between food insecurity and obesity in Latino women. Am J Clin Nutr. 2004;80:1372–1378. [DOI] [PubMed] [Google Scholar]
- 24.Gundersen C, Weinreb L, Wehler C, Hosmer D. Homelessness and food insecurity. J Housing Econ. 2003;12:250–277. [Google Scholar]
- 25.Canadian Guidelines for Body Weight Classification in Adults. Ottawa, Ontario: Health Canada; 2003.
- 26.Cole TJ, Bellizzi MC, Flegal KM, et al. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ. 2000;320:1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Bickel G, Nord M, Price C, Hamilton CW, Cook J. Measuring Food Security in the United States: Guide to Measuring Household Food Security. Washington, DC: US Department of Agriculture; 2000.
- 28.Rao JNK, Wu CFJ, Yue K. Some recent work on resampling methods for complex surveys. Surv Methodol. 1992;18:209–217. [Google Scholar]
- 29.Rust KF, Rao JN. Variance estimation for complex surveys using replication techniques. Stat Methods Med Res. 1996;5:283–310. [DOI] [PubMed] [Google Scholar]
- 30.Bostrom G, Diderichsen F. Socioeconomic differentials in misclassification of height, weight and body mass index based on questionnaire data. Int J Epidemiol. 1997;26:860–866. [DOI] [PubMed] [Google Scholar]
- 31.Rowland ML. Self-reported weight and height. Am J Clin Nutr. 1990;52:1125–1133. [DOI] [PubMed] [Google Scholar]
- 32.Oliver G, Wardle J, Leigh Gibson L. Stress and food choice: a laboratory study. Psychosom Med. 2000; 62:853–865. [DOI] [PubMed] [Google Scholar]
- 33.Greeno CG, Wing RR. Stress-induced eating. Psychol Bull. 1994;115:444–464. [DOI] [PubMed] [Google Scholar]
- 34.Rutledge T, Linden W. To eat or not to eat: affective and physiological mechanisms in the stress-eating relationship. J Behav Med. 1998;21:221–240. [DOI] [PubMed] [Google Scholar]
- 35.Tarasuk VS. Discussion Paper on Household and Individual Food Insecurity. Ottawa, Ontario: Office of Nutrition Policy, Health Canada; 2001.
- 36.Roberts M, Innes M, Lyons C. Insight on Cancer: News and Information on Nutrition and Cancer Prevention. Ottawa, Ontario: Cancer Care Ontario; 2005.
- 37.Improving the Health of Canadians: Promoting Healthy Weights. Ottawa, Ontario: Canadian Population Health Initiative; 2006.
- 38.Gibson D. Long-term food stamp program participation is positively related to simultaneous overweight in young daughters and obesity in mothers. J Nutr. 2006;136:1081–1085. [DOI] [PubMed] [Google Scholar]
- 39.Biggerstaff MA, Morris PM, Nichols-Casebolt A. Living on the edge: examination of people attending food pantries and soup kitchens. Soc Work. 2002;47: 267–277. [DOI] [PubMed] [Google Scholar]
- 40.Tarasuk VS. Health implications of food insecurity. In: Raphael D, ed. Social Determinants of Health: Canadian Perspectives. Toronto, Ontario: Canadian Scholars Press Inc; 2004:97–100.
- 41.Lawn J, Harvey D. Nutrition and Food Security in Kugaaruk, Nunavut: Baseline Survey for the Food Mail Pilot Project. Ottawa, Ontario: Indian and Northern Affairs Canada; 2004.