Food insecurity is defined as inconsistent or unpredictable access to affordable, nutritionally adequate food(1). Food insecurity has historically been associated with underweight status, but is now increasingly associated with obesity(Reference Larson and Story2–Reference Nettle, Andrews and Bateson4). This association with obesity may be explained by the fact that low-nutrient, low-fibre, energy-dense foods tend to be cheaper than nutrient-dense, fibre-rich, lower-energy foods(Reference Monsivais and Drewnowski5,Reference Drewnowski and Specter6) . Thus, food insecure (FI) individuals may disproportionately depend on these low-cost, less nutritious foods, contributing to their risk for obesity. Another possible explanation for the food insecurity–obesity association is that food insecurity is linked to fluctuating food availability, which may increase the risk for cycles of energetic restriction followed by binge eating, a pattern which, if repeated over time, may increase the propensity for weight gain(Reference Dinour, Bergen and Yeh7).
Several epidemiological studies have investigated the relationship between food insecurity and weight status in adults and children. A 2017 meta-analysis included 123 cross-sectional and seven longitudinal studies and found an overall positive association between food insecurity and elevated weight status in analyses unadjusted and adjusted for socio-economic status(Reference Nettle, Andrews and Bateson4). When looking at specific populations, this meta-analysis found that the food insecurity–obesity association was stronger in adult females than in children (defined as <16 years of age) or adult males.
Food insecurity is of particular concern for adolescents because they have high energy and nutrient needs to fuel their rapid growth and development. Reliable access to food that is familiar, of acceptable quality, and nutritionally adequate is foundational to healthful eating patterns for adolescents(Reference Satter8,Reference Satter9) . Consequently, any disruption to that foundation – including food insecurity – may put adolescents at increased risk for appetite dysregulation, inconsistent eating patterns and disruption to predictable growth and development patterns, including accelerated weight velocity(Reference Satter8,Reference Spear10) . When investigating the food insecurity–weight status relationship in adolescents, findings have been mixed(Reference Widome, Neumark-Sztainer and Hannan11–Reference Lohman, Neppl and Lee15). A notable study by Lohman and colleagues examined the association of food insecurity with BMI gain over a 16-year period through prospective growth curve analysis(Reference Lohman, Neppl and Lee15). At baseline, participants were 15 years old, primarily non-Hispanic white and residing in the rural Midwest. Investigators found that household food insecurity at 15 years of age predicted more rapid BMI gain from 16 to 31 years of age. However at baseline, there was no cross-sectional association between household food insecurity and BMI(Reference Lohman, Neppl and Lee15).
Food insecurity may also be linked to disordered eating. Cross-sectional studies in adults have found positive associations between food insecurity and disordered eating(Reference Bruening, MacLehose and Loth16–Reference Lydecker and Grilo21). One US-based study of adults found that food insecurity was more prevalent in the group with bulimia nervosa pathology in comparison with binge eating and control groups(Reference Lydecker and Grilo21). Another study in adult patrons of food pantries in San Antonio, TX found that participants with the least food security reported higher levels of binge eating and overall eating disorder pathology compared with more food secure (FS) participants(Reference Becker, Middlemass and Taylor20). Despite the high prevalence of disordered eating behaviours in adolescents(Reference Neumark-Sztainer, Wall and Larson22–Reference Eaton, Kann and Kinchen25), studies examining the relationship between food insecurity and disordered eating behaviours and attitudes in this population are limited(Reference Tester, Lang and Laraia26,Reference West, Goldschmidt and Mason27) . We identified only three studies, a qualitative study(Reference Tester, Lang and Laraia26) and a study which examined food insecurity as one potential risk factor for disordered eating(Reference West, Goldschmidt and Mason27), and one which examined food insecurity and body dissatisfaction(Reference Altman, Ritchie and Frongillo28). These studies found that food insecurity was associated with disordered eating or body dissatisfaction in adolescents. Our study builds on these data by specifically investigating whether household food insecurity, reported by parents/guardians, is associated with measured adolescent weight status and a comprehensive panel of problematic eating behaviours, reported by adolescents, in an ethnically diverse sample. Household food insecurity reported by parents/guardians was chosen as the exposure variable because of our goal to improve understanding of how the home food environment, not just the individual perceptions of adolescents regarding their food insecurity, may support or disrupt the foundation of healthful eating patterns for adolescents. We chose parent/guardian report because it is likely the more accurate measure of household-level food security status.
The current study expands upon the extant literature on food insecurity and weight status and disordered eating behaviours to explore these associations in a large population-based sample of adolescents. The study explores cross-sectional associations both with and without adjustment for socio-economic status and ethnicity/race, given the interconnectedness of food insecurity with household income(Reference Bhargava, Jolliffe and Howard29). We hypothesised that food insecurity would be associated with higher prevalence of overweight status, meal skipping, binge eating and unhealthy weight control behaviours (i.e. fasting, eating very little food, using food substitutes, skipping meals, smoking more cigarettes, volitional vomiting, laxative, diet pill or diuretic use) in unadjusted analyses. Lower socio-economic status is associated with higher prevalence of food insecurity(Reference Coleman-Jensen, Rabbitt and Gregory30), less frequent breakfast intake(Reference Larson, Story and Eisenberg31,Reference Watts, Mason and Loth32) , higher prevalence of disordered eating(Reference Nagata, Garber and Tabler33) and higher BMI(Reference Watts, Mason and Loth32,Reference Sherwood, Wall and Neumark-Sztainer34,Reference O’Dea and Caputi35) . Additionally, the existing literature suggests that persons who identified with an ethnicity/race other than non-Hispanic white experience higher prevalence of food insecurity(Reference Coleman-Jensen, Rabbitt and Gregory30) and tend to have less frequent breakfast intake(Reference Larson, Story and Eisenberg31), higher prevalence of disordered eating(Reference Rodgers, Watts and Austin36) and higher BMI(Reference Rodgers, Watts and Austin36). We were interested in understanding the independent effect of food insecurity on disordered eating and BMI; thus, we examined models adjusted for ethnicity/race and our measure of socio-economic status and parental education. We hypothesised that parental education and ethnicity/race would account for some of the association between food insecurity and weight status, meal skipping and disordered eating, and therefore, associations observed in unadjusted models would be attenuated in the adjusted models. To the best of our knowledge, this study is the first to examine associations between food insecurity, using a strong measure of household food insecurity, with both weight status and disordered eating in a population-based sample of urban adolescents.
Methods
Study design and sample
EAT 2010 (Eating and Activity in Teens) was designed to examine dietary intake, physical activity, weight control behaviours, weight status and factors associated with these outcomes in adolescents(Reference Berge, Wall and Larson37–Reference Larson, Wall and Story39). Project F-EAT (Families and Eating and Activity Among Teens) was designed to examine factors within the family and home environment of potential relevance to these weight-related behaviours(Reference Bruening, MacLehose and Loth16,Reference Bauer, MacLehose and Loth40,Reference Berge, MacLehose and Loth41) . Data for this analysis are from these two coordinated, population-based studies.
For EAT 2010, surveys and anthropometric measures were completed by 2793 adolescents during the 2009–2010 academic year. The study population included adolescents from twenty public middle and high schools in the Minneapolis–St. Paul metropolitan area of Minnesota, which serve ethnically/racially diverse and largely low-income communities.
For Project F-EAT, data were collected by surveying up to two parents/guardians (n 3709) of the adolescents in EAT 2010. For the purpose of simplicity throughout the remainder of this article, both parents and guardians will be referred to as ‘parents.’ Approximately 70 % of adolescents provided information for two parents and 30 % provided contact information for one parent. In total, 2382 adolescent participants in EAT 2010 (85·3 %) had at least one parent respondent and there were two parent respondents for 1327 adolescents. Because our analyses depended on household variables obtained from the parents, adolescent participants whose parents did not participate in the study were excluded. Only data from the adolescent’s primary parent were used in the current analysis to ensure the most accurate information on the usual home environment. When two parents responded, primary parent status was determined using an algorithm that accounted for the family living situation (preference to parents who lived with their child more than half the time), relationship to the adolescent (preference to biological and adoptive parents over step-parents) and the parent’s sex (preference to female because of literature indicating that female parents are more likely to be responsible for household food purchasing and preparation)(Reference Larson, Wall and Story39). Pairing adolescents with the primary parent resulted in a final analytic sample of 2285 adolescent–parent dyads.
Adolescent survey development and measures
The EAT 2010 survey is a 235-item self-report instrument which assesses a range of factors of potential relevance to weight status and weight-related behaviours among adolescents. Survey development was guided by a review of previous Project EAT surveys(Reference Neumark-Sztainer, Croll and Story42,Reference Neumark-Sztainer, Story and Perry43) to identify the most salient items; a theoretical framework, which integrates an ecological perspective with Social Cognitive Theory(Reference Sallis, Owen, Fisher, Glanz, Rimer and Viswanath44,Reference Bandura45) ; expert review by professionals from different disciplines and extensive pilot testing with adolescents(Reference Berge, Wall and Larson37,Reference Larson, Wall and Story39) . The test–retest reliability of measures over a one-week period was also examined in a separate sample of 129 middle and high school students(Reference Neumark-Sztainer, Wall and Larson22,Reference Goldschmidt, Loth and MacLehose46) .
Adolescent socio-demographic characteristics
Age, sex and ethnicity/race were self-reported by adolescents. Ethnicity/race was assessed with the following question: ‘Do you think of yourself as…? (1) White, (2) Black or African American, (3) Hispanic or Latino, (4) Asian American, (5) Native Hawaiian or Pacific Islander, (6) American Indian or Native American or (7) Other.’ Only thirty-five adolescents reported ‘Hawaiian or Pacific Islander’; therefore, they were coded as ‘Mixed or Other Race.’
Adolescent weight status
Trained research staff measured adolescents’ height and weight using standardised procedures during selected health, physical education and science classes. Measurements were completed in a private area(Reference Gibson47). Height was measured to the nearest 0·1 cm using a Shorr board and weight to the nearest 0·1 kg using a calibrated scale. Centers for Disease Control and Prevention guidelines were used to calculate BMI and BMI percentile for each adolescent(Reference Kuczmarski, Ogden and Grummer-Strawn48,Reference Bauer, Marcus and Larson49) . Adolescents’ BMI was then dichotomised to represent weight status; a BMI ≥ 85th percentile was categorised as overweight, and a BMI < 85th percentile was categorised as not overweight.
Adolescent meal frequency and disordered eating behaviours
Adolescents were asked to report how often they ate breakfast, lunch and dinner. For example, they were asked, ‘During the past week, how many days did you eat breakfast?’ Response options ranged from 0 to 7 d, and participants selected one response (test–retest r = 0·76). The same questions and responses were repeated for lunch (test–retest r = 0·47) and dinner (test–retest r = 0·56).
Based on extensive pilot testing with adolescents and expert review to determine face validity(Reference Berge, Wall and Larson37,Reference Neumark-Sztainer, Croll and Story42) , unhealthy weight control behaviours were assessed with the following question, ‘Have you done any of the following things in order to lose weight or keep from gaining weight during the past year?’ Responses included fasted, ate very little food, used a food substitute, skipped meals, smoked more cigarettes, took diet pills, made myself vomit, used laxatives and used diuretics. In addition to reporting ‘yes/no’ in response to each of the individual unhealthy weight control behaviours, a positive response for one or more behaviour was coded as use of any unhealthy weight control behaviour. Those unhealthy weight control behaviours further categorised as extreme included took diet pills, made myself vomit, used laxatives and used diuretics. Study test–retest agreement was 85 % for unhealthy weight control behaviours and 96 % for the extreme weight control behaviour subset(Reference Goldschmidt, Loth and MacLehose46).
Binge eating with and without loss of control was assessed using two questions adapted from the adult version of the Questionnaire on Eating and Weight Patterns-Revised(Reference Yanovski50). This questionnaire has good psychometric properties in adolescents(Reference Johnson, Grieve and Adams51). Binge eating was assessed by asking, ‘In the past year, have you ever eaten so much food in a short period of time that you would be embarrassed if others saw you (binge-eating)?’ Loss of control was assessed next by asking, ‘During the times when you ate this way, did you feel you couldn’t stop eating or control what or how much you were eating?’ Test–retest agreement for the binge question was 90, and 75 % for the loss of control question(Reference Goldschmidt, Loth and MacLehose46).
Parent survey development and measures
The Project F-EAT survey was designed to assess food-specific parenting practices and the home food environment. A multi-disciplinary research team developed the Project F-EAT survey with guidance from an ecological framework(Reference Sallis, Owen, Fisher, Glanz, Rimer and Viswanath44,Reference Story, Kaphingst and Robinson-O’Brien52) , a comprehensive review of the literature, and extensive pilot testing with parents of adolescents. Test–retest reliability was assessed in a subsample of 102 parent respondents who completed the parent survey twice within a 2-week period(Reference Bruening, MacLehose and Loth16).
Parental education
Parent-reported household educational attainment was used to measure socio-economic status(Reference Bruening, MacLehose and Loth16,Reference Sherwood, Wall and Neumark-Sztainer34,Reference Berge, Wall and Larson38,Reference Neumark-Sztainer, Story and Hannan53) . Parents were asked, ‘What is the highest grade or year of school that you have completed?’ The same question was asked about their spouse or partner. Responses included did not finish high school, finished high school or got General Educational Development, some college or training after high school, finished college, and advanced degree (e.g. Master’s degree, PhD, MD).
Household food insecurity
Past-year food security was measured as a part of the parent survey using the previously validated six-item US Household Food Security Survey Module modified for self-administration which assesses food security over the past 12 months. This scale has been shown to correctly classify 97·7 % of families when compared with the full eighteen-item scale included in the Current Population Survey(Reference Blumberg, Bialostosky and Hamilton54,55) . The survey module includes the following items: ‘Is this statement true?: “We couldn’t afford to eat balanced meals”’; ‘In the past 12 months, did you or other adults in your household ever cut the size of your meals or skip meals because there wasn’t enough money for food?’; ‘In the past 12 months, did you ever eat less than you felt you should because there wasn’t enough money for food?’ and ‘In the past 12 months, were you ever hungry but didn’t eat because there was not enough money for food?’ Affirmative responses to the items were summed according to the established guidelines, and this raw score (test–retest r = 0·77) was used to categorise households as FS (score = 0–1) or FI (score ≥ 2)(Reference Bruening, MacLehose and Loth16,55) . We used the less severe cut-point of low food security (score ≥ 2) because we were most interested in understanding, from a population perspective, how more prevalent exposures to food insecurity were linked to more common public health problems (i.e. disordered eating behaviours). Additionally, in a supplemental sensitivity analysis, we used the more severe cut-point of very low food security (score ≥ 5)(Reference Bruening, MacLehose and Loth16,56) .
Statistical analysis
Bivariate associations between household food insecurity status and key variables including socio-demographic characteristics were assessed using t tests or χ 2 tests, as appropriate. We were interested in understanding the specific effect of food insecurity on disordered eating and weight status above and beyond the effect of socio-economic status. Therefore, crude and adjusted (parental education, race/ethnicity, sex and age) logistic regression models with robust standard errors were used to estimate marginal probabilities and 95 % CI for each binary outcome (e.g. overweight status). Similarly, crude and adjusted (parental education, race/ethnicity, sex and age) linear regression models with robust standard errors were used to estimate marginal means and 95 % CI for each ordinal outcome (e.g. frequency of breakfast). Adjusted models used Huber–White robust sandwich estimators to adjust for potential clustering by school. Linear regression models are appropriate for ordinal and other non-normally distributed outcomes in large sample sizes(Reference Lumley, Diehr and Emerson57). All analyses were conducted using SAS 9.4 (copyright 2002–2012).
Results
Descriptive findings of food insecurity by socio-demographic characteristics
Participating adolescents were ethnically/racially diverse and were primarily living in low socio-economic status homes (Table 1). The mean age of adolescent participants was 14·4 years (age range 10–22 years); 45·8 % were identified as male and 54·2 % as female. Parent participants had a mean age of 42·3 (sd 8·6) years. The majority of parent respondents were mothers or other female guardians (62·0 %). Household food insecurity was experienced in the past year by 38·9 % of this ethnically/racially diverse, largely low-income, urban adolescent sample. Non-white adolescent race, lower household educational attainment and lower household income were associated with greater household food insecurity (P < 0·001). The data did not suggest differences in household food security status based on age or sex (Table 1).
Table 1 Socio-demographic characteristics of a sample of adolescents in Minneapolis–St. Paul, Minnesota by household food security status in 2009–2010*
P-values are bolded if statistically significant at the 0·05 significance level.
* χ 2 test: % and n. T Test: mean and sd. Parental education was reported by parents and refers to the highest educational attainment of any parent or guardian in the household. All other socio-demographic characteristics were self-reported by adolescents.
Eating behaviours and weight status by food insecurity status in adolescents
In the unadjusted model, household food insecurity was associated with higher prevalence of overweight status (FI: 42·3 % v. FS: 37·9 %, P = 0·039). Adjusting for ethnicity/race, parental education, sex and age attenuated this difference, and it was no longer statistically significant (Tables 2 and 3).
Table 2 Eating behaviours and weight status by household food security status in adolescents, unadjusted estimates*
UWCB, Unhealthy Weight Control Behaviour.
P-values are bolded if statistically significant at the 0·05 significance level.
* χ 2 test: % and n. T test: Mean and sd.
Table 3 Eating behaviours and weight status by household food security status in adolescents, adjusted estimates*
UWCB, Unhealthy Weight Control Behaviour.
P-values are bolded if statistically significant at the 0·05 significance level.
* Logistic regression models with robust standard errors were used to estimate marginal probabilities and 95 % CI for each binary outcome variable adjusted for parental education, race/ethnicity, age, sex and potential clustering by school. Similarly adjusted multiple regression models with robust standard errors were used to estimate means and 95 % CI for mean number of breakfast, lunch and dinner outcomes.
† Prevalence is too low to estimate result.
Household food insecurity was associated in unadjusted models with less frequent adolescent breakfast consumption (FI: 4·1 times/week v. FS: 4·4 times/week, P = 0·005), greater likelihood of engaging in any unhealthy weight control behaviour (FI: 49·0 % v. FS: 39·5 %, P < 0·001) and the following specific unhealthy weight control behaviours: fasting, eating very little food, meal skipping, laxative use and diuretic use to control weight. There was no association between food insecurity and volitional vomiting, diet pill use, food substitute use or smoking more cigarettes nor was there an association with binge eating (Table 2).
After adjusting for race/ethnicity, parental education, sex and age, associations remained statistically significant for food insecurity with any unhealthy weight control behaviour (FI: 44·5 % v. FS: 37·8 %, P = 0·007), including the following specific unhealthy weight control behaviours: fasting (FI: 13·4 % v. FS: 10·0 %, P = 0·047) and laxative use (FI: 0·8 % v. FS: 0·3 %, P = 0·005) (Table 3). Observed relationships were similar when we used the more severe cut-point value of very low food security from the six-item US Household Food Security Survey Module (data not shown).
Discussion
In this population-based sample of ethnically diverse, largely low-income, urban adolescents, we found that household food insecurity was associated with higher prevalence of overweight and several unhealthy eating behaviours in unadjusted models. After adjustment for ethnicity/race, parental education, sex and age, associations remained between food insecurity and unhealthy weight control behaviours such as fasting and laxative use. Household food insecurity, overweight and disordered eating behaviours were all common in this adolescent sample, with 43 % reporting unhealthy weight control behaviours, 40 % experiencing overweight and 39 % living in FI homes. These findings highlight the complex nature of prevalent nutrition-related problems experienced by adolescents.
Studies examining the food insecurity-eating behaviour relationship have generally not focused on the adolescent age group. To our knowledge, only three studies have investigated associations between food insecurity and disordered eating behaviours and attitudes in adolescents(Reference Tester, Lang and Laraia26–Reference Altman, Ritchie and Frongillo28). The first was a qualitative study by Tester and colleagues in low-income children and adolescents enrolled in an outpatient weight management programme (mean age: 11 years old)(Reference Tester, Lang and Laraia26). Using the US Core Food Security Model to characterise food insecurity and Diagnostic and Statistical Manual of Mental Disorders, 5th Edition criteria for binge eating disorder and other specified feeding and eating disorders to characterise disordered eating, participants’ disordered eating emerged as a theme in the FI group but not in the FS group(Reference Tester, Lang and Laraia26). The second study by West and colleagues was a longitudinal population-based study in adolescents that explored various risk factors for disordered eating, one of which was food insecurity. This study found that self-reported food insecurity during adolescence independently predicted binge eating 5 years later, but only in adolescents from low socio-economic status backgrounds(Reference West, Goldschmidt and Mason27). Our results are consistent with both of these studies in that we found positive associations between household food insecurity and disordered eating behaviours in adolescents. A third study by Altman and colleagues found that food insecurity was cross-sectionally associated with higher odds of body dissatisfaction in children and adolescents (n 14 768, mean age = 10·2 years)(Reference Altman, Ritchie and Frongillo28). Their results were consistent across all racial/ethnic groups and BMI categories even after adjusting for age, sex, free/reduced price meal eligibility, BMI and race/ethnicity(Reference Altman, Ritchie and Frongillo28). More studies are needed to further understand the relationships between food insecurity and eating behaviours, body satisfaction and weight status in young people.
When investigating the food insecurity-eating behaviour relationship in adolescents, it is important to consider the complex nature of the food environments that FI adolescents occupy. For example, FI adolescents may be attempting to control their weight while also navigating various food environments at home, school, with peers, at work and/or extracurricular activities. Furthermore, adolescents may engage not only in disordered eating to control weight but also for other reasons. For example, a 2016 report found that in FI households, parents generally try to protect both children and adolescents from experiencing hunger. Adolescents also routinely take on this role by restricting their food intake to ensure younger siblings have enough to eat(Reference Waxman, Popkin and Galvez58). Therefore, future research on this topic should investigate competing motivations for adolescent disordered eating behaviours. It may also be useful to explore whether birth order and/or the specific age during which food insecurity is most acute plays a role in eating behaviour within the context of a FI household.
Household food insecurity, overweight and disordered eating behaviours were each highly prevalent in our sample of adolescents from diverse ethnic/racial and socio-economic backgrounds. These findings are concerning, given the well-established individual health risks associated with obesity(Reference Ogden, Carroll and Fryar59,Reference Dietz60) , food insecurity(Reference Olson, Bove and Miller61–Reference Seligman, Laraia and Kushel64) and disordered eating(Reference Lock, Reisel and Steiner65–Reference Hudson, Hiripi and Pope69). Previous results reported by our team and others reveal that both binge eating(Reference Nagata, Garber and Tabler33,Reference Bauer, Marcus and Larson49,Reference Field, Camargo and Taylor70) and restrictive eating(Reference Nagata, Garber and Tabler33,Reference Rodgers, Watts and Austin36,Reference Field, Austin and Taylor71) are positively associated with overweight. Despite these findings, food insecurity, disordered eating and overweight are most often investigated and addressed as separate and unrelated(Reference Nagata, Garber and Tabler33,Reference Neumark-Sztainer72–Reference Kennedy, Forman and Woods77) . In recent years, some attention has been paid to the overlap between overweight and disordered eating in adolescents, but the role of food insecurity is generally not included in such discussions(Reference Nagata, Garber and Tabler33,Reference Neumark-Sztainer72–Reference Kennedy, Forman and Woods77) . Healthcare providers who work with adolescents should be aware that disordered eating, food insecurity and overweight are prevalent problems that can co-occur and may perhaps exacerbate one another. Given the high prevalence of food insecurity, disordered eating and overweight, and their potential to negatively impact health outcomes for adolescents, understanding how to best address all three of these problems in a comprehensive manner should be the focus of future research.
Our study has several strengths. To our knowledge, it is the first quantitative study to investigate the relationship between household food insecurity and a range of weight-related outcomes in adolescents: meal frequency, disordered eating and weight status. Our population-based sample was large and ethnically diverse. Adolescents’ height and weight were measured by trained research staff using standardised procedures. We were able to survey adolescents about their own eating behaviours, survey parents about household food security status and then link parent data to adolescent data. We investigated a wide variety of eating behaviour and socio-demographic variables in the adolescents. Finally, the survey’s extensive pilot testing and test–retest reliability testing ensured that questions were developmentally appropriate for adolescents. Despite our study’s strengths, it also has important limitations. Because the causes of disordered eating and food insecurity are multifactorial, there may have been confounding or causal variables that we did not include in our statistical models. Our study relied on several self-reported measures, so recall bias may have affected our results. Due to our cross-sectional design, we were unable to draw conclusions about a temporal or causal relationship between household food insecurity and adolescent eating behaviours or weight status. It is possible that parent-reported food security may not accurately reflect the food security experienced by the adolescent. Carlos Chavez and colleagues found discordance in reports of food insecurity in 51 % of 70 Latino adolescent–parent dyads(Reference Carlos Chavez, Hernandez and Harris78). However, it is worth considering that developmental factors may play a role in such discordance. Furthermore, our specific research question focused on the home food environment, rather than adolescent perceptions of food security; thus, we determined parental report of food security was the most appropriate exposure for our study. Finally, these data were collected in 2010, shortly after the Great Recession of 2008, and the median household income range in this sample ($20 000–34 999) was much lower than that of households with families in 2010 in Minneapolis and St. Paul of $61 725 and $55 254, respectively(79). These factors may explain the high prevalence of food insecurity in our sample and may impact the generalisability of our results. However, the high prevalence of food insecurity in this sample provided a unique opportunity to improve understanding of the role of food insecurity in adolescent nutrition-related health outcomes.
Our study builds on a small amount of published literature showing a positive association between household food insecurity and adolescent disordered eating(Reference Tester, Lang and Laraia26–Reference Altman, Ritchie and Frongillo28). Our findings point to the need for well-designed longitudinal studies to investigate the aetiological nature of these relationships. Furthermore, these results draw attention to health disparities that may exist for youth who are overweight and/or FI. Clinicians and public health practitioners should be aware that all adolescents, including those who are overweight or FI, are at high risk for disordered eating behaviours and the accompanying negative health consequences. Given the high prevalence and poor health outcomes associated with food insecurity, overweight and disordered eating in adolescents, future research should focus on elucidating these relationships and understanding how to intervene with individuals who are at the highest risk. This study highlights the varied and intersecting barriers adolescents face to achieving nutritionally adequate, balanced and healthful eating patterns. More research is needed to confirm these findings and further understand these multifaceted problems.
Acknowledgements
Acknowledgements: None. Financial support: This study was supported by grant numbers R01HL127077 and R35HL139853 from the National Heart, Lung, and Blood Institute (PI: D.N.-S.). L.H. was supported by grant number 5 T79MC00007-31-00 from Health Resources and Services Administration Maternal and Child Health Bureau. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Heart, Lung, and Blood Institute, the National Institutes of Health, or the Health Resources and Services Administration Maternal and Child Health Bureau. Conflicts of interest: None. Authorship: All authors helped conceptualise the analysis plan, assisted with interpretation of the results, contributed to writing the manuscript and thoroughly reviewed the final manuscript. In addition, L.H. formulated the research question, conducted statistical analysis and drafted the manuscript. S.T. oversaw the statistical analysis. N.L. helped with statistical analysis. S.M.M. helped formulate the research question. D.N.-S. conceptualised the larger Project EAT study design and oversaw data collection. Ethics of human subject participation: This study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving research study participants were approved by the University of Minnesota’s Institutional Review Board Human Subjects Committee. Written informed assent or consent was obtained from all participants.
