Food Insecurity and Binge Eating: A Systematic Review and Meta-Analysis

Abstract

Objective:

This review synthesized literature on the relationship between food insecurity and binge eating.

Methods:

Relevant studies were identified by searching PubMed, CINAHL, PsycINFO, and grey literature from inception to October 2022. Eligible studies included primary research that assessed the relationship between food insecurity and binge eating. Data extraction was performed independently by 2 reviewers. Pooled odds ratios and 95% confidence intervals (CI) were obtained from random effect models with the R package meta. Analyses were stratified by binge eating vs binge eating disorder (BED), study type (cross-sectional vs longitudinal), and age (adults vs adolescents).

Results:

We included 24 articles that reported on 20 studies and 13 articles were included in the meta-analysis. Based on the random effects meta-analysis, the odds of adults in the food insecure group having binge eating were 1.66 (95% CI=1.42, 1.93) times the odds of adults in the food secure group having binge eating. The odds of adults in the food insecure group having BED were 2.70 (95% CI=1.47, 4.96) times the odds of adults in the food secure group having BED. Insufficient data were available for a meta-analysis on adolescents or longitudinal relationships.

Conclusions:

These findings support that food insecurity is associated with binge eating in adults. There is a need for research to investigate mechanisms underlying this relationship. Results highlight the importance of screening participants with food insecurity for disordered eating behaviors and vice versa. Future research is needed to examine whether interventions targeting food insecurity may help to mitigate disordered eating behaviors.

Introduction

Conceptual and theoretical models of binge eating, defined as eating an amount of food that is objectively larger than most people would eat in a similar period of time under similar circumstances while feeling a loss of control over eating, have predominantly focused on psychological contributors to this condition (Burton & Abbott, 2017). A widely cited factor related to the development and maintenance of binge eating is dietary restraint. Dietary Restraint Theory was proposed by Herman and Mack in 1975 and emphasized that restricting intake to the point of becoming chronically hungry made people more susceptible to overeating (Herman & Mack, 1975). This model was further supported by Polivy and Herman in 1985 when they presented evidence that food restriction or dieting precedes binge eating, and that people who binge eat may be trapped in a cycle of restricting food intake and binge eating (Polivy & Herman, 1985). They postulated that people who engaged in restricted eating would cognitively regulate their eating as they no longer were relying on physiological cues to eat. Preclinical and clinical data have shown that restrictive eating is a precipitating factor for some individuals to develop binge eating, and cognitive dietary restraint remains a core component of modern models of binge eating (Williamson et al., 2004). However, most models of binge eating have focused on internal, cognitive, behavioral, and psychological influences and little has been synthesized about how external social determinants of health, such as food insecurity, may contribute to binge eating.

Food insecurity is the limited or uncertain availability of nutritionally adequate and safe foods or limited or uncertain ability to acquire foods in socially acceptable ways (USDA Economic Research Service, 2022a). Food security status lies on a continuum from high food security to marginal food security to low food security to very low food security (USDA Economic Research Service, 2022a). Low food security is defined as reduced quality, variety, or desirability of diet with little to no indication of reduced food intake. Very low food security is classified as disruptions in eating patterns and reduced intake. In the United States, the prevalence of household food insecurity was 10.2% in 2021 (USDA Economic Research Service, 2022b).

Several processes may underlie the relationship between food insecurity and binge eating. The episodic nature of food insecurity is characterized by periods of decreased intake when an individual’s food supply is low followed by overconsumption when food is available (Dinour et al., 2007; Hamrick & Andrews, 2016; Hazzard et al., 2023). This creates a cycle of increased food availability and restriction; when food runs low or runs out, individuals are forced to reduce the quantity of food consumed or skip meals (Dinour et al., 2007; Hamrick & Andrews, 2016). People with food insecurity may also need to have greater cognitive dietary restraint to spread their resources for other household members (i.e., children) or stretch food to make it last longer (Middlemass et al., 2021), which may increase their likelihood of binge eating. In addition, it has been proposed that the stress of being food insecure contributes to increased chronic activation of cortisol production pathways, an important risk factor for binge eating as well as consumption of highly palatable foods (Gluck et al., 2004; Walker et al., 2018). Numerous primary studies have similarly investigated the link between food insecurity and binge eating. A review by Hazzard et al., 2020 identified 14 studies that quantitatively examined the relationship between food insecurity and overall eating disorder psychopathology, disordered eating behaviors, or eating disorder diagnosis, including binge eating (Hazzard et al., 2020). They found 8 papers that assessed the food insecurity-binge eating relationship and these generally supported the relationship between these variables. There has been an increase in the number of studies that have examined the relationship between food insecurity and binge eating. The present paper expands on the work by Hazzard et al., 2020 by providing an updated review of the literature and synthesizing this literature in a meta-analysis. The purpose of this paper was to perform a systematic review and meta-analysis to assess the relationship between food insecurity and binge eating.

Method

This review was planned and conducted under the direction of the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for systematic reviews (Page, McKenzie, Bossuyt, Boutron, Hoffmann, Mulrow, Shamseer, Tetzlaff, Akl, et al., 2021; Page, McKenzie, Bossuyt, Boutron, Hoffmann, Mulrow, Shamseer, Tetzlaff, & Moher, 2021). This review was not pre-registered.

Literature Search

A search strategy was developed in consultation with a research librarian at the University of Pennsylvania. PubMed, CINAHL, and PsycINFO were searched for relevant articles published between inception to December 2021. The search was updated in October 2022 during the peer review process to identify any additional articles. Search terms were selected by referring to similar literature and discussion among the authors. Search terms included (“binge-eating disorder”[Mesh] OR “binge eating” OR binge eat* OR eating disorder* OR disordered eat* OR eating pathology OR “loss of control eat*” OR “emotional eat*”) AND (“food insecurity”[Mesh] OR food desert* OR food insecur* OR food secur* OR food neglect). A supplemental search of grey literature was performed in MedNar, Web of Science, Scopus, and ProQuest. Terms for the grey literature search included “binge eating” and “food insecurity.” The reference lists of relevant review articles and of all eligible studies were manually examined for additional articles not identified by the search.

Inclusion/Exclusion Criteria

Inclusion criteria were that studies evaluated individual or household food insecurity or a synonymous concept using a scale, questionnaire, or indicator; evaluated binge eating or binge-eating disorder (BED) using a scale, questionnaire, or interview; and assessed the relationship between food insecurity and binge eating/BED. Studies involving participants of any age were included. Searches were limited to human studies using quantitative methods published in English.

Study Selection

Two authors independently screened titles and abstracts for possible eligibility. All duplicate articles were removed. Articles found to be potentially eligible by at least 1 reviewer were moved to full-text review. At the full-text review level, 2 authors independently screened articles based on the inclusion/exclusion criteria. The other authors helped reach consensus on any articles that had discrepancies.

Data Extraction and Quality Assessment

A data extraction tool was created for this study. The form was tested with 3 studies for refinement. Data were independently extracted by 2 authors and then fully checked by a third author for completeness and accuracy. Disagreements were resolved by discussion among authors, if necessary. Information was extracted on study characteristics (e.g., authors, year of publication, sample size), sample characteristics (e.g., age, BMI, sex, race/ethnicity), methods to assess food insecurity and binge eating, and outcomes related to binge eating, food insecurity, and their relationship. Two reviewers independently assessed the risk of bias of the included studies using a modified version of the Newcastle-Ottawa scale (Herzog et al., 2013; Modesti et al., 2016). The scale assessed representativeness of the sample, sample size, non-respondents, ascertaining of exposure, comparability based on design and analysis, assessment of outcome, and statistical test.

Data Synthesis and Meta-Analysis

Data were grouped by condition examined (binge eating and BED), study design (cross-sectional and longitudinal), and age (adults and adolescents). We performed a meta-analysis that included articles that reported the frequency of binge eating and/or BED among people with and without food security. To be included in the meta-analysis, studies had to report on the percent of individuals with binge eating and/or BED (binary) among individuals with and without food security. For papers that were included in the systematic review but did not contain this data, we contacted corresponding authors of the studies by email to request this information. Results were stratified by design (cross-sectional versus longitudinal) and by age (adults versus adolescents) given the importance of considering development context in binge eating and BED (Tanofsky-Kraff et al., 2020). Data analysis was performed using R, version 4.2.1. Groups that had more than 2 studies were pooled using odds ratios (OR) and 95% confidence intervals (CI) were obtained with both fixed effect and random effect models with R package meta (Schwarzer, 2007). The between-study heterogeneity was measured using the Cochran Q test and $I^2$ statistic. A Cochran Q test p<0.05 was considered significant for heterogeneity. An $I^2$ value of 0% to 25% represents insignificant heterogeneity, 26% to 50% represents low heterogeneity, 51% to 75% represents moderate heterogeneity, and >75% represents high heterogeneity.

Results

A total of 1244 records were identified during the initial electronic search. After removal of 759 duplicates and the title and abstract review, 53 articles were assessed at full-text level (Figure 1). Another 29 articles were excluded. This resulted in a total of 24 articles included in this review.

Figure 1.

PRISMA 2020 flow diagram for new systematic reviews

Reason 1 Does not focus on binge eating or loss of control eating

Reason 2 Wrong study design

Reason 3 Does not link binge eating with food insecurity

Study Design

All articles (N=24) were published in or after 2012 and report on 20 different studies. Information about study characteristics, measurement of food insecurity and binge eating, results, and quality are presented in Tables 1, 2, 3, and 4, respectively. Although studies outside of the US were not intentionally excluded, the results of the literature review only yielded studies based in the US. Of the 24 articles, 7 (29.2%) used data from Project EAT (Eating and Activity over Time), a program of studies that assessed weight-related problems among young adults across the life course. This included cohorts from Project EAT I-III, Project EAT 2010-2018, and Project F-EAT (Families and Eating and Activity in Teens). Out of the 24 included articles, 21 (87.5%) were cross-sectional analyses and 3 (12.5%) were longitudinal (all from Project EAT).

Table 1.

Study Characteristics

Citation	N	Mean (Range) Age, years	Mean BMI/BMI Percentile or % Overweight	% Female	Race		Ethnicity	Education				% Receiving Federal Benefits
					%White	%Black	%Hispanic/Latino	% < High School	% High School	% Some College	% College Degree or Higher
Ash et al., 2014	276	34.9 (>18)	28.5	83.0	73.6	8.0	9.4	1.5	11.2	33.3	53.6	NR
Becker et al., 2017	503	NR (NR)	NR	76.5	11.3	16.5	64.6	41.7	28.4	23.5	5.4	NR
Becker et al., 2019	891	42.1 (NR)	30.7	67.3	10.1	5.9	76.2	36.5	31.5	26.5	5.4	NR
Bruening et al., 2012	2095	41.0 (NR)	28.5	92.5	29.4	27.8	16.9	29.5	21.0	26.8	21.7	NR
Bruening et al., 2017	55 (mothers)	NR (NR)	NR	100	NR	NR	85.3	NR	50.9			85
	55 (adolescents)	14.6 (NR)	NR	43.6	NR	NR	83.6	NR	NA	NA	NA	NR
Cheung et al., 2022	369	42.7 (18-72)	NR	79.7	94.9	3.5	0.5	3.8	26.8	38.2	31.1	19.0
Christensen et al., 2021	579	21.8 (18-78)	25.1	76.3	84.1	3.5	9.0	0	100		NR	NR
Coffino et al., 2020	36145	46.5 (≥18)	NR	51.9	66.2	11.8	14.8	13.0	25.8	61.2	NR	16.3
Goode et al., 2021	4553	42 (NR)	NR	63	0	100	0	24	35	24	17	53
Hazzard, Barry, et al., 2022	2914	44.9 (NR)	NR	53.4	74.3	11.5	10.2	14.2	29.5	30.9	25.4	NR
Hazzard, Hooper, et al., 2022	1813	EAT-I: 14.9 (11-18) EAT-II: 19.5 (16-23) EAT-III: 24.9 (21-31)	NR	57.1	50.7	17.8	5.7	NR	NR	NR	NR	NR
Hooper et al., 2020	2285	14.5 (10-22)	39.6% overweight	54.2	21.0	28.0	17.2	NA	NA	NA	NA	NR
Hooper et al., 2022	1340	EAT-2010: 14.5 (NR) EAT-2018: 22.0 (NR)	Elevated BMI during adolescence: 22.7% BMI≥30 kg/m2 during young adulthood: 28.4%	53.0	20.3	28.6	16.9	NR	NR	NR	NR	NR
Horvath et al., 2022	366	41.5 (19-67)	BMI ≥35 kg/m2: 100%	85.5	94.8	2.7	1.1	4.4	26.0	26.2	43.4	NR
Kim et al., 2021	58	15.2 (11.1-18.9)	BMI percentile: 80.5%	62	15	50	34.5	NA	NA	NA	NA	NR
Larson et al., 2020	1568	Baseline: 14.4 (NR) Follow up: 22.0 (NR)	NR	57.9	18.8	29.0	16.9	48.8	35.6	37.4	8.6	24.2
Linsenmeyer et al., 2021	164	17.0 (12-23)	BMI: 25.1 BMI percentile: 60.3 (for consistent with gender identity)	17.1(transgender female)	NR	NR	NR	NR	NR	NR	NR	NR
Poll et al., 2020	111	21 (19-23)	NR	0	56.8	34.2	0.9	0	NR	NR	NR	NR
Rasmusson et al., 2018	1251	35.4 (NR)	28.0	67.8	67.1	NR	20.6	NR	12.8	34.1	53.1	NR
Simone et al., 2021	720	24.7 (NR)	NR	62.1	29.6	18.2	16.5	NR	NR	NR	NR	NR
Stinson et al., 2018	82	38 (18-65)	29	35.4	34	11	11	Mean=12.4	NR	NR	NR	NR
West et al., 2019	2179	14.9 (NR)	22.0	52.8	63.4	10.0	3.9	NR	NR	NR	NR	NR
West et al., 2021	60	13.9 (12-17)	BMI Percentile: 65.5	53.3	73.1	9	1.7	NR	NR	NR	NR	NR
Zickgraf et al., 2019	240	41.1 (17-70)	48.3	74	71	NR	NR	NR	NR	NR	22.9	NR

Note. NR=Not reported. NA=Not applicable.

Table 2.

Measurements of Food Insecurity and Binge Eating

Study	Food Insecurity								Binge Eating
	Measure	Current or Past	Duration	% of Sample	Level of Food Security				Measure	Binge eating, BED, or LOC?	Duration	% of Sample
					% Very Low	% Low	%Marginal	% High
Ash et al., 2014	6-item US Household Food Security Survey Module	Current	12 months	48.2	27.5	20.7	NR	NR	Eating Disorder Examination Questionnaire	Binge eating	28 days	NR
Becker et al., 2017	Radimer/Cornell Food Insecurity Measure	Current	NR	91.8	45.1a	37.8b	8.9c	8.2d	Eating Disorder Diagnostic Scale for DSM 5	Binge eating	NR	12.8
Becker et al., 2019	Radimer/Cornell Food Insecurity Measure	Current	NR	90.3	21.6a	27.8b	41.0c	9.7d	Eating Disorder Diagnostic Scale for DSM 5	Binge eating	NR	16.9
Bruening et al., 2012	6-item US Household Food Security Survey Module	Current	12 months	38.9	13.3	NR	NR	NR	“In the past year, have you ever eaten so much food in a short period of time that you would be embarrassed if other saw you (binge eating)?”	Binge eating	Past year	8.2
Bruening et al., 2017	6-item US Household Food Security Survey Module	Current	12 months	Mothers: 65.4 Adolescents: 43.6	Mothers: 34.5 Adolescents: 14.5	NR	NR	NR	One question from The Questionnaire on Eating and Weight Patterns-Revised (QEWP-R)	Binge eating	12 months	Mothers: 29.1 Adolescents: 30.1
Cheung et al., 2022	10-item US Adult Food Security Survey Module	Current	12 months	20.9	7.9 (“high FI”)		13.0	79.1	Gormally Binge Eating Scale	Binge eating symptom score	NR	27.1 (≥mild symptoms)
Christensen et al., 2021	Radimer/Cornell Food Insecurity Measure	Current	NR	52.8	NR	NR	NR	NR	Eating Disorder Diagnostic Scale 5 and Clinical Impairment Assessment	BED	NR	2.1
Coffino et al., 2020	“Did a parent or other adult living in your home make you go hungry or not prepare regular meals?” *Study assessed food neglect	Past (Before 18 years of age)	NR	1.5	NR	NR	NR	NR	NIAAA Alcohol Use Disorder and Associated Disabilities Interview Schedule 5 (AUDADIS-5)	BED	Lifetime	NR
Goode et al., 2021	“In the past 12 months, in your household, there was: enough food to eat; sometimes not enough food to eat; often not enough food to eat”	Current	12 months	11	NR	NR	NR	NR	World Mental Health Composite International Diagnostic Interview	Recurrent binge eating	Lifetime/12 months	Past 12 months: 2 Lifetime: 3
Hazzard, Barry, et al., 2022	5-items modified from US Household Food Security Survey Module	Current	12 months	11.1	NR	NR	NR	NR	World Health Organization Composite International Diagnostic Interview	BED	12 month	1.2
Hazzard, Hooper, et al., 2022	1-item adapted from US Household Food Security Survey Module: “How often during the last 12 months have you been hungry because your family [you for young adults] couldn’t afford food?”	Current and past	12 months	Eat I: 8.9 Eat II: 20.9 Eat III: NA	NR	NR	NR	NR	Two questions from the Questionnaire on Eating and Weight Patterns-Revised	Binge eating	Past year	Baseline: 8.6 5-year: 6.9 10-year: 12
Hooper et al., 2020	6-item US Household Food Security Survey Module (completed by parent)	Current	12 months	38.9	NR	NR	NR	NR	Two questions adapted from adult version of the Questionnaire on Eating and Weight Patterns-Revised	Binge eating	Past year	7.8
Hooper et al., 2022	6-item U.S. Household Food Security Survey Module (completed by parent)	Past	12 months	Baseline: 37.8 Follow up: NR	NR	NR	NR	NR	1 question adapted from the Questionnaire on Eating and Weight Patterns-Revised	Binge eating	Past year	Prevalence Baseline: 13.6 Follow up: 20.2 Incidence: 17.3 Persistence: 38.1
Horvath et al., 2022	10-item U.S. Household Food Security Survey Module	Current	12 months	17.2	NR	NR	NR	NR	Gormally Binge Eating Scale	Binge eating score	NR	Mean BES score 12.6 Mild to moderate: 23 Severe: 4.3
Kim et al., 2021	Food Security Survey Module for Youth Ages 12 and Older	Current	1 month	41	NR	NR	NR	NR	Eating Disorders Examination Questionnaire Adolescent Version (EDE-A)	Binge eating	28 days	NR
Larson et al., 2020	2 items from US Household Food Security Survey Module	Current and past (only current reported in relation to binge eating)	12 months	Past: NR Current: 23.3	NR	NR	NR	NR	“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)?”	Binge eating	Past year	NR
Linsenmeyer et al., 2021	Hunger Vital Sign	Current	12 months	21.2	NR	NR	NR	NR	Adolescent Binge Eating Disorder Questionnaire (ADO-BED)	BED score	NR	9.1
Poll et al., 2020	US Adult Food Security Survey Module	Current and past	Current: 9-12 months Past: Senior year of high school	High school: 18.0 College: 19.8	High school: 6.3 College: 7.2	High school: 7.2 College: 2.7	High School: 4.5 College: 9.0	High School: 82.0 College: 80.2	Two questions from the Questionnaire on Eating and Weight Patterns- Revised	Binge eating	9-12 months	NR
Rasmusson et al., 2018	6-item US Household Food Security Survey Module	Current	12 months	33.7	15.3	18.5	NR	NR	Questionnaire on Eating and Weight Patterns-5	BED	3 months	6.8
Simone et al., 2021	2 items modified from US Household Food Security Survey Module	Current	Past month	13.0	NR	NR	NR	NR	2 questions adapted from the Questionnaire on Eating and Weight Patterns-Revised	Binge eating	Past month	14.1
Stinson et al., 2018	6-item US Household Food Security Survey Module	Current	12 months	56	NR	NR	NR	NR	Gormally Binge Eating Scale	Binge eating score	NR	Mean BES score 5.4
West et al., 2019	1-item adapted from US Household Food Security Survey Module	Current	12 months	6.6	NR	NR	NR	NR	Two questions adapted from adult version of the Questionnaire on Eating and Weight Patterns-Revised	Binge eating	Past year	5.7 (excluded those with baseline BE)
West et al., 2021	18-item US Household Food Security Survey Module (completed by parent)	Current	12 months	21.6	3.3	18.3	13.3	65.0	Eating Disorders Examination Questionnaire	LOC eating	28 days	11.7
Zickgraf et al., 2019	10-item US Adult Food Security Survey Module	Current	NR	15.8	15.8 (Collapsed as “food insecure”)		25.8	58.3	Gormally Binge Eating Scale	Binge eating and BED	NR	Clinically significant BED: 6.3

Note.

^aChild hunger;

^bIndividual;

^cHousehold;

^dNot food insecure.

BED=binge eating disorder. LOC eating=Loss of control eating. NR=Not reported.

Table 3.

Results of the Relationship between Food Insecurity and Binge Eating

Study	Age Group	Design for Assessment of Food Insecurity and Binge Eating	Recruitment Setting	Summary of Relationship between Food Insecurity and Binge Eating	Results of Unadjusted Analyses	Results of Adjusted Analyses	Covariates
Food insecure vs food secure: Binge eating (Yes vs No)
Bruening et al., 2012	Adults	Cross-sectional	Parents of students (Project F-EAT)	↑	Food secure: 6.4 Food insecure: 11.3 p<0.001	Food secure: 6.3 (5.0, 7.8) Food insecure: 11.1 (8.9, 13.7) Adjusted difference: 4.8 (2.1, 7.7) p< 0.01	Age, gender, race/ethnicity, marital status, employment status, and highest household education
Bruening et al., 2017	Adults	Cross-sectional	Public housing sites	=	p=0.43	NR	None
Christensen et al., 2021	Adults	Cross-sectional	University	↑	Food secure: 28.9 Food insecure: 38.9 p=0.02	NR	None
Goode et al., 2021	Adults	Cross-sectional	Adults from the National Survey of American Life	↑	Recurrent in past 12 months OR: 1.87 (1.10, 3.17); p=0.02 Lifetime: p=0.006	Recurrent OR: 1.58 (0.93, 2.69); p=0.09	Age, marital status, education, employment status
Hazzard, Hooper et al., 2022	Adults	Cross-sectional	Project EAT-II	Not calculated	Food secure: 6.4% Food insecure: 10.6%	NR	NR
Horvath et al., 2022	Adults	Cross-sectional	Bariatric surgery candidates	Not calculated	Food secure: 96.7% Food insecure: 100% *Horvath et al., defined binge eating behavior as at least 1 on the BES	NR	None
Poll et al., 2020	Adults	Cross-sectional	College athletic teams	=	Food secure: 9.2% Food insecure: 0% p=0.94	NR	None
Simone et al., 2021	Adults	Cross-sectional	Project C-EAT (COVID-19 EAT from 2010-2018 cohort)	=	NR	Full sample OR: 1.22 (0.64-2.34) First incidence of binge eating OR: 1.33 (0.54-3.25)	Gender, race/ethnicity, socioeconomic status
Larson et al., 2020	Adolescents to adults	Cross-sectional/Longitudinal	Project EAT 2010-2018	↑	Food secure: 18.4 Food insecure: 25.3 p=0.005	Food secure: 17.8 Food insecure: 25.6 p=0.003	Gender identity, ethnicity/race, parent socioeconomic status, student status, employment status, receipt of public assistance, living situation
Bruening et al., 2017	Adolescents	Cross-sectional	Public housing sites	=	p=0.57	NR	None
Hazzard, Hooper et al., 2022	Adolescents	Cross-sectional	Project EAT-I	Not calculated	Food secure: 7.0% Food insecure: 10.3%	NR	NR
Hooper et al., 2020	Adolescents	Cross-sectional	Project EAT 2010/Project F-EAT	=	p=0.72	Food secure: 7.5 (6.4, 8.7) Food insecure: 6.9 (5.2, 9.1) p=0.65	Parental education, race/ethnicity, age, sex
Kim et al., 2021	Adolescents	Cross-sectional	Clinical sample with loss of control or binge eating, and community sample	=	p=0.60	NR	None
West et al., 2019	Adolescents	Longitudinal	Project EAT I-II	↑	NR	Food insecurity at EAT-I was associated with an increased risk for endorsing binge eating at EAT-II in the low-SES group RR: 1.4 (0.7, 2.8) Not able to calculate for high-SES group due to sparse data	Age, gender, race, and BMI at baseline
Hazzard, Hooper, et al., 2022 Longitudinal: Past food insecurity→future BE	Adolescents to adults	Longitudinal	Project EAT I-III	↑	NR	Cross sectional PR: 1.49; 95% CI: 1.04 - 2.12; p<0.05 Longitudinal PR: 1.41; 95% CI: 1.05 - 1.90; p<0.05	Cross-sectional: Age, sex, ethnicity/race, socioeconomic status Longitudinal: Age, sex, ethnicity/race, socioeconomic status, and disordered eating behavior at the time of food insecurity assessment
Hooper et al., 2022 Longitudinal: Past food insecurity→future BE
Prevalence	Adolescents to adults	Longitudinal	Project EAT 2010-2018/Project F-EAT	=	Food secure: 18.9 Food insecure: 22.4 p=0.13	Food secure: 19.9 (16.9-23.2) Food insecure: 23.7 (19.6-28.4) p=0.11	Ethnicity/race and parental education
Incident BE	Adolescents to adults	Longitudinal	Project EAT 2010-2018/Project F-EAT	↑	p=0.04	Food secure: 16.2 (13.3-19.5) Food insecure: 21.3 (17.1-26.3) p=0.04	Ethnicity/race and parental education
Persistent BE	Adolescents to adults	Longitudinal	Project EAT 2010-2018/Project F-EAT	=	p=0.67	Food secure: 43.0 (32.6-54.1) Food insecure: 40.3 (27.2-54.9) p=0.74	Ethnicity/race and parental education
Food insecure vs food secure: BED (Yes vs No)
Coffino et al., 2020 Cross-sectional: Retrospective recall of past food neglect/lifetime BED	Adults	Cross-sectional	Adults from National Epidemiological Survey on Alcohol and Related Conditions III	↑	p<0.01	OR Model 1: 3.89 (2.33-6.49) OR Model 2: 2.95 (1.73-5.03)	Model 1: Age, sex, race/ethnicity, income, education Model 2: Model 1, adverse childhood experiences, and history of being on government assistance programs before age 18
Hazzard, Barry, et al., 2022	Adults	Cross-sectional	Adults from National Comorbidity Survey-Replication	↑	Food secure: 1.2% Food insecure: 5.3% p<0.001	PR: 3.49 (1.84-6.62)	Age, sex, race/ethnicity, education, and income-to-poverty ratio
Christensen et al., 2021	Adults	Cross-sectional	University	=	No significant difference	NR	None
Food insecure vs food secure: Binge eating (Continuous)
Ash et al., 2014	Adults	Cross-sectional	Online	=	Binge eating episodes Food secure: 3.3 Food insecure: 3.7 p=0.72	NR	None
Cheung et al., 2022	Adults	Cross-sectional	Bariatric surgery candidates	↑	BES score: r=0.14, p<0.001 BES scores Food secure: 12.1 Food insecure: 14.6	NR	None
Stinson et al., 2018	Adults	Cross-sectional	Community	↑	NR	Food insecure>food secure; p=0.01	Age, sex, race, and ethnicity, and subjective social status
Zickgraf et al., 2019	Adults	Cross-sectional	Bariatric surgery candidates	↑	NR	B=0.82 (0.28, 1.36), p=0.003	Sex, ethnicity, education attainment, age
Kim et al., 2021	Adolescents	Cross-sectional	Clinical sample with loss of control or binge eating, and community sample	↑	NR	Among those with at least 1 BE episode, mean episodes in past month: FI= 6.8 ± 5.4 Not FI = 3.3 ± 4.6 p<0.01	Site
Linsenmeyer et al., 2021	Adolescents or young adults	Cross-sectional	Adolescent gender clinic	↑	r=0.25, p<0.05	NR	None
Level of food insecurity: Binge eating (Yes/No)
Becker et al., 2017	Adults	Cross-sectional	Food pantries	NR	Not food insecure: 7.9 Household food insecure: 2.4 Individual food insecure: 11.7 Child hunger food insecure: 16.4 *No statistical test performed	NR	None
Becker et al., 2019	Adults	Cross-sectional	Food pantries	NR	Not food insecure: 11.9 Household food insecure: 13.4 Individual food insecure: 16.3 Child hunger food insecure: 26.6 *No statistical test performed	NR	None
Level of food insecurity: Binge eating (Continuous)
Ash et al., 2014	Adults	Cross-sectional	Online	=	NR	Food insecurity score (continuous)*binge eating frequency (continuous): r=0.01, p>0.05	Gender, race, body mass index
Cheung et al., 2022	Adults	Cross-sectional	Bariatric surgery candidates	=	BES score: r=0.08, NS Food secure: 12.1 Marginally food secure: 15.0 Food insecure: 13.8	NR	None
Horvath et al., 2022	Adults	Cross-sectional	Bariatric surgery candidates	↑	r=0.15, p<0.01	NR	None
Zickgraf et al., 2019	Adults	Cross-sectional	Bariatric surgery candidates	=	BES scores: p=0.06	NR	None
Christensen et al., 2021	Adults	Cross-sectional	University	↑	No food insecurity: mean=1.6 Food insecurity: mean=2.5 p=0.01	NR	None
West et al., 2021	Adolescents	Cross-sectional	Community	↑	r=0.56, p<0.01	b = 0.66, p < .01	Age, sex, race, ethnicity, and BMI percentile
Level of food insecurity: BED (Yes/No)
Rasmusson et al., 2018	Adults	Cross-sectional	Amazon’s Mechanical Turk	↑	Low food security and very low food security were associated with classification in the BED group relative to healthy weight and obesity groups (p=0.001)	Healthy weight vs BED Low food security OR: 2.45 (1.42, 4.24) Very low food security OR: 1.91 (1.01-3.61)	Sex, age, education level, and ethnicity
Zickgraf et al., 2019	Adults	Cross-sectional	Bariatric surgery candidates	↑	p=0.002	NR	None

Note. Numbers in paratheses are 95% confidence intervals. OR=odds ratio. PR=prevalence ratio. RR=relative risk.

Table 4.

Results of the Critical Appraisal of the Included Studies

Citation	Selection (Maximum 5 Stars)				Comparability (Maximum 2 Stars)	Outcome (Maximum 3 Stars)		Total Score
	Representativeness of the sample (Maximum 1 Star)	Sample size (Maximum 1 Star)	Non-respondents (Maximum 1 Star)	Ascertainment of exposure (Maximum 2 Stars)	Based on design and analysis	Assessment of outcome (Maximum 2 Stars)	Statistical test (Maximum 1 Star)
Ash et al., 2014				**	*	*	*	5
Becker et al., 2017	*			**	*	*	*	6
Becker et al., 2019	*			**	*	*	*	6
Bruening et al., 2012	*		*	**	*	*	*	7
Bruening et al., 2017	*			**	*	*	*	6
Cheung et al., 2022	*			**	*	*	*	6
Christensen et al., 2021			*	**	*	*	*	6
Coffino et al., 2020	*		*	*	*	**	*	7
Goode et al., 2021	*		*	*	*	**	*	7
Hazzard, Barry, et al., 2022	*		*	**	*	**	*	8
Hazzard, Hooper, et al., 2022	*		*	*	*	*	*	6
Hooper et al., 2020	*		*	**	*	*	*	7
Hooper et al., 2022	*		*	**	*	*	*	7
Horvath et al., 2022	*		*	**	*	*	*	7
Kim et al., 2021	*			**	*	*	*	6
Larson et al., 2020	*		*	**	*	*	*	7
Linsenmeyer et al., 2021	*		*	**		*	*	6
Poll et al., 2020				**		*	*	4
Rasmusson et al., 2018	*			**	*	*	*	6
Simone et al., 2021	*			**	*	*	*	6
Stinson et al., 2018				**	*	*	*	5
West et al., 2019	*		*	*	*	*	*	6
West et al., 2021				**	*	*	*	5
Zickgraf et al., 2019				**		*	*	4

Sample Characteristics

There was a wide range of sample sizes ranging from 58 to 36,145 participants (Table 1). The sample recruitment settings were diverse (Table 3). Settings included middle and high schools (n=7; 29.2%; all from Project EAT); bariatric surgery clinics (n=3; 12.5%); national surveys (n=3; 12.5%); food pantries (n=2; 8.3%); community (n=2; 8.3%); universities (n=2; 8.3%); online (n=2; 8.3%); public housing sites (n=1; 4.2%); and an adolescent gender clinic (n=1; 4.2%). One additional study (4.2%) included participants recruited from clinic and community settings. Of the 24 articles, 13 (54.2%) assessed adults, 5 (20.8%) examined adolescents, 3 (12.5%) studied longitudinal changes from adolescence to adulthood, 2 (8.3%) measured college students, and 1 (4.2%) examined both parents and their children (Tables 1 and 3). The majority (n=21; 87.5%) of reports included samples that were ≥50% female. In 12 (50%) articles 50% or more of the sample were White, in 2 (8.3%) reports 50% or more of the sample were Black, and in 3 (12.5%) papers 50% or more identified as Hispanic/Latino. Over half of the articles (n=16; 66.7%) collected information on educational attainment. Only 5 (20.8%) articles reported data about the percentage of participants receiving federal government benefits.

Study Methodology

Seven articles (29.2%) used the 6-item version of the US Household Food Security Survey (United States Department of Agriculture Economic Research Service, 2021), 3 (12.5%) used the Radimer/Cornell Food Insecurity Measure (Kendall et al., 1995), 3 (12.5%) used the 10-item version of the US Household Food Security Survey (United States Department of Agriculture Economic Research Service, 2021), 2 (8.3%) reports used 2-items from the US Household Food Security Survey Module (United States Department of Agriculture Economic Research Service, 2021), 1 (4.2%) used the Food Security Survey Module for Youth Ages 12 and Older (USDA Economic Research Service, 2006), 1 (4.2%) article used the 18-item US Household Food Security Survey (United States Department of Agriculture Economic Research Service, 2021), 1 (4.2%) used 5 items modified from the US Household Food Insecurity Survey Module (United States Department of Agriculture Economic Research Service, 2021), 1 (4.2%) used the Hunger Vital Sign (Gundersen et al., 2017), and 1 (4.2%) used an unspecified version of the US Adult Food Security Survey Module (Table 2). Four articles (16.7%) used 1-item, including one study that assessed food neglect using the question “Did a parent or other adult living in your home make you go hungry or not prepare regular meals?” (Coffino et al., 2020). Most articles assessed current food insecurity (n=19; 79.2%) and 3 (12.5%) also assessed past food insecurity. There were 2 (8.3%) studies that measured past food insecurity alone. The prevalence of food insecurity in studies ranged from 1.5% to 91.8%. Level of food security (e.g., very low, low) was reported in 8 (33.3%) articles and 2 (8.3%) additional articles reported household, individual, and child hunger.

Most articles assessed binge-eating behavior (n=14; 58.3%) or BED (n=4; 16.7%). Other studies assessed a binge-eating symptom score (n=3; 12.5%), BED score (n=1; 4.2%), loss of control eating (n=1; 4.2%), or both binge eating and BED (n=1; 4.2%). Of the articles that reported the prevalence of binge eating in their sample, the percent endorsing this behavior ranged from 2% to 30.1%. Eight (33.3%) articles used The Questionnaire on Eating and Weight Patterns 5 (Yanovski et al., 2015) or questions adapted from this measure, 4 (16.7%) used the Gormally Binge Eating Scale (Gormally et al., 1982), 3 (12.5%) articles used Eating Disorder Diagnostic Scale for DSM 5 (Stice et al., 2000), 3 (12.5%) used the Eating Disorders Examination Questionnaire (1 adolescent and 2 adult version) (Mond et al., 2004), 2 (8.3%) used the World Mental Health Composite International Diagnostic Interview (Kessler & Üstün, 2004), 1 (4.2%) article used the NIAAA Alcohol Use Disorder and Associated Disabilities Interview Schedule 5 (Grant et al., 2015), and 1 (4.2%) used the Adolescent Binge Eating Disorder Questionnaire (Chamay-Weber et al., 2017). Two (8.3%) of the articles asked: “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)?”

Quality Appraisal

Table 4 shows the results of the quality appraisal of eligible studies. The average quality rating on a scale of 0 (low) to 10 (highest) was 6.1. The scores of included studies ranged from 4 (low) to 8 (high).

Meta-Analysis

From the identified papers, 13 articles were eligible for meta-analysis. Of the articles that included adults, 9 studies examined binge-eating behavior and 5 assessed BED. For adults, the odds ratio of binge eating between food insecure and secure groups was 1.66 (95% CI=1.42, 1.93; $I^2$=0%; Figure 2). The odds ratio of BED between food insecure and secure groups was 2.70 (95% CI=1.47, 4.96; $I^2$=73%; Figure 3). A sensitivity analysis was conducted that removed the study by Coffino et al., which assessed food neglect. By excluding this study, the odds ratio decreased but remained statistically significant (OR=2.33, 95% CI=1.06, 5.11, $I^2$=74%; Figure S1). There were not sufficient studies to pool data from adolescents or longitudinal designs in a meta-analysis.

Figure 2.

Meta-analysis of adult, cross-sectional studies investigating the relationship between binge eating and food insecurity.

Note. Becker et al., 2017 and Becker et al., 2019 reported results clustering for “food insecurity” based on combining the responses for household/individual/child hunger scores. For Poll et al., 2020 which included a zero cell, a continuity correction of 0.5 was used.

Figure 3.

Meta-analysis of adult, cross-sectional studies investigating the relationship between binge eating disorder and food insecurity.

Note. Rasmusson et al., 2018 reported results in a combined group of low/very low food security. For Zickgraf et al., 2019, we did not include results on marginal food insecurity.

Narrative Synthesis

While there were not sufficient studies in adolescents that reported on the percent of individuals with and without binge eating/BED by food insecurity status to pool in a meta-analysis, three cross-sectional studies conducted in adolescents showed no statistically significant association between food insecurity and binge eating (Table 3). However, a cross-sectional analysis of Project EAT I-II found that severe food insecurity was associated with a greater prevalence of binge eating (PR=1.49, 95% CI=1.04, 2.12) (Hazzard, Hooper, et al., 2022). Three articles evaluated the relationship between a history of food insecurity and binge eating. Two were longitudinal cohort studies that followed individuals from adolescence to adulthood, Project EAT I-III (Hazzard, Hooper, et al., 2022) and Project EAT 2010-2018 (Hooper et al., 2022), and one was a cross-sectional study that used retrospective recall for past food insecurity (Coffino et al., 2020). All three studies supported that the experience of food insecurity in childhood was a significant predictor of the development of binge eating in later life. The relationship between food insecurity and BED in adolescents was not assessed in any studies.

Six studies examined whether individuals with food insecurity, compared to those with food security, had higher levels of binge eating on a continuous scale (Table 3). All but one study (Ash, 2014) found that those with food insecurity, relative to those without food insecurity, had higher scores on binge eating. Three of these studies used a continuous score on the Gormally Binge Eating Scale (Cheung et al., 2022; Stinson et al., 2018; Zickgraf et al., 2019), 3 used the Eating Disorder Examination Questionnaire (two with the adult version (Fairburn & Beglin, 1994) and one was the Adolescent version) and 1 used the Adolescent Binge Eating Disorder Questionnaire. Six studies examined whether level of food insecurity was associated with continuous binge eating scores; 3 found that higher food insecurity was associated with higher continuous scores and 3 did not.

Discussion

Food insecurity is associated with reduced dietary quality (Leung et al., 2014), overweight/obesity (Moradi et al., 2019), and mood disorders (Pourmotabbed et al., 2020). This is the first meta-analysis to demonstrate that food insecurity is also associated with binge eating. Based on the random effects meta-analysis, the odds of adults in the food insecure group having binge eating were 1.66 times the odds of adults in the food secure group having binge eating. The odds of adults in the food insecure group having BED were 2.70 times the odds of adults in the food secure group having BED. Dietary restraint due to shape, weight, and eating concerns is a well-documented contributor to binge eating (Herman & Mack, 1975) and target for binge-eating interventions (Wilson & Fairburn, 1993). The current results demonstrate that involuntary, externally imposed dietary restraint, due to food insecurity, is an important factor to consider in the etiology and treatment of binge eating.

These findings are congruent with dietary restraint theories of binge eating and highlight that external dietary restraint due to financial restraints should be considered in the prevention and treatment of binge eating (Herman & Mack, 1975). There are several mechanisms that likely contribute to the relationship between food insecurity and binge eating (Hazzard et al., 2020). Insufficient food may contribute to feelings of dietary restraint and deprivation (Middlemass et al., 2021), restricted food choices, anxiety about eating, and preoccupation with food (Rosa et al., 2018). Ultimately, these feelings can contribute to overeating and feelings of loss of control when food is available (Olson et al., 2007). The primary source of nutrition assistance for individuals in the US with a low-income is the Supplemental Nutrition Assistance Program (SNAP), and 28.2% of households exhaust their SNAP benefits within a week of receipt and 53.3% within the first two weeks (Insight Policy Research Group, 2021). SNAP beneficiaries have a significant decline in dietary quality in the final 10 days of the benefit cycle (Whiteman et al., 2018). Food insecurity is associated with negative affect including depressive symptoms and stress (Pourmotabbed et al., 2020). Negative affect is a well-recognized antecedent to binge eating (Dingemans et al., 2017), as shown in affect regulation models (Haedt-Matt & Keel, 2011). Binge eating may serve as a coping mechanism to downregulate negative emotions (Brockmeyer et al., 2014). Food insecurity is also highly associated with socioeconomic status and poverty, which also may contribute to these relationships. For example, Simone et al., controlled for overall financial difficulties in their analysis (Simone et al., 2021). In adjusted analyses, they found that food insecurity was not significantly associated with binge eating. Examination of how general financial difficulties may also contribute to binge eating and BED are important to consider when developing novel interventions for this population. For example, low income may prevent patients from being able to own a car to shop for food at a grocery store.

Individuals with food insecurity are more likely to live in “food deserts,” defined as low access to outlets selling healthy food such as supermarkets, supercenters, or large grocery stores (Ver Ploeg et al., 2009). At the same time, these individuals are also more likely to live in “food swamps,” characterized by a greater density of stores that sell nutrient-poor, energy-dense food (e.g., fast-food restaurants and convenience stores) than healthy food options (Rose et al., 2009; Rummo et al., 2017), which are the types of foods that individuals are most likely to binge eat (Allison & Timmerman, 2007). As posited by conditioning models of binge eating, exposure to such a food environment may prompt food cravings and binge eating (Jansen, 1998; Meule et al., 2018). High-calorie and highly-palatable foods are often less expensive than lower calorie foods and cost of food is a primary determinant of dietary intake, especially for people with low incomes (Darmon & Drewnowski, 2015).

There was a moderate level of heterogeneity among the five studies that assessed the relationship between food insecurity and BED in adults. Each of these studies had very different sample characteristics and recruitment methods that ranged from a university to national samples to bariatric surgery candidates to Amazon Turk. In addition, one of the studies examined food neglect (Coffino et al., 2020). “Food neglect” and “food insecurity” are conceptually different, but the way “food neglect” is measured is frequently overlapping with food insecurity. In the article by Coffino et al., food neglect was quantified by the question “Did a parent or other adult living in your home make you go hungry or not prepare regular meals?” (Coffino et al., 2020). We chose to retain this study as we believe it fits conceptually into food insecurity, defined as a lack of consistent access to enough food to live an active, healthy life (albeit a severe form). Results remained statistically significant when we removed this study. More studies are needed that measure the relationships among food insecurity, food neglect, and BED.

We were not able to synthesize results from adolescents in a meta-analysis. However, we found evidence that the experience of food insecurity in childhood is a predictor of the development of binge eating in later life. This suggests that children experiencing food insecurity should be carefully monitored for the development of binge eating as they age. However, socioeconomic challenges and food insecurity are often chronic and these associations may be reflective of the relationships between current food insecurity and binge eating. Future longitudinal studies are needed to examine the longitudinal nature and course of the relationships between food insecurity and binge eating. In addition, further study is needed to examine the association between food insecurity and loss of control eating. Given the challenges of defining a large amount of food in adolescents as well as the relative lack of autonomy over food decisions (Byrne et al., 2019; Moustafa et al., 2021), loss of control eating may be an important behavior to examine in relationship to food insecurity in youth and may better capture disordered eating in this population.

This study has several limitations. We only assessed studies written in English. This review included studies that examined binge-eating behavior. Due to the commonality of subthreshold levels of binge eating, we did not limit studies to those that compared individuals with and without BED. Further, we did not review eating disorder psychopathology or other disordered eating behaviors (e.g., self-induced vomiting or diuretic use) or eating disorders (e.g., bulimia nervosa or anorexia nervosa-binge purge). Future studies are needed to understand whether associations between food insecurity and binge eating extend to these other eating disorders that feature binge eating as a core symptom.

Despite these limitations, this review underscores the importance of considering broader structural and social determinants of disordered eating. Importantly, patients with binge eating should be screened for food insecurity and vice versa. If patients with binge eating screen positive for food insecurity, they should be actively referred to food resources. Patients with food insecurity should also be screened for binge eating. Further, patients with food insecurity may have challenges with access to eating disorder treatment and care. Interventions that improve food insecurity may help mitigate risk of binge eating. Future studies are needed to test interventions tailored to participants with food insecurity and binge eating.

Public Significance Statement.

Food insecurity is a common but underrecognized contributor to binge eating. In this article, we systematically reviewed research that has been published on the relationship between food insecurity and binge eating. We found support that food insecurity should be considered in the prevention and treatment of binge eating.

Data availability:

Data covered in this systematic review and meta-analysis are available from the corresponding author, upon reasonable request.