Eating Disorders in U.S. Active Duty Military Members and Veterans: A Systematic Review

ABSTRACT

Introduction

Feeding and eating disorders can be difficult to treat and frequently co-occur with other mental health conditions. The last systematic review of eating disorders in a military and veteran population was published in 2015. An updated review is warranted to re-examine the current literature on eating disorders in the active duty and veteran populations.

Materials and Methods

A systematic review that described the prevalence, co-occurrence of other disorders and/or events, and health care utilization of U.S. active duty members and veterans was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses. Databases and Medical Subject Headings (MeSH) terms used are listed in  Appendix A. Each category of the literature was extracted and graded using the Oxford Centre for Evidence-Based Medicine Levels of Evidence.

Results

Twenty-one studies revealed prevalence estimates with varying rates based on demographic information. Trauma exposure is consistently associated with eating disorder development. Individuals diagnosed with eating disorders had greater health care utilization.

Conclusions

Research on eating disorders in the military and veteran populations has expanded in recent years. Limitations of the evidence included in this review stem from the use of self-reported questionnaires, changes to medical record systems, and limited generalizability to the overall population of patients with eating disorders. Further research should investigate the impact of demographic factors and trauma exposure on the development of an eating disorder within the military and veteran populations.

INTRODUCTION

Feeding and eating disorders are manifestations of disordered eating behaviors that impair psychological and physical health.¹ Binge eating disorder is characterized by recurrent symptoms of out-of-control eating and consuming a large amount of food.¹ Patients with anorexia nervosa obsessively desire to lose weight and have a low Body Mass Index (BMI).¹ Bulimia nervosa is defined as episodes of binging with purging or methods to avoid weight gain.¹ Patients with eating disorders report severe psychological distress in response to the loss of control of their eating patterns and associated social isolation.² Comorbidities include bipolar II disorder, major depressive disorder (MDD), obsessive-compulsive disorder, PTSD, and substance use disorders (SUD).^1,3 Patients frequently experience potentially irreversible medical complications.¹ Worldwide, 8.4% (3.3%-18.6%) of women and 2.2% (0.8%-6.5%) of men will have an eating disorder. In America, at any given time, 4.6% of the population has an eating disorder.⁴ Between 2018 and 2019, the cost of eating disorders in America was estimated at $64.7 billion.⁵

Members of the U.S. military experience eating disorders at comparable rates to civilians.⁶ The crude overall incidence rate of any eating disorder was 2.7 cases per 10,000 person-years in U.S. active duty service members.⁶ However, this is likely an underrepresentation because eating disorders are understudied in U.S. active duty, veteran, and male populations.^4,7 Underreporting due to career concerns may affect the true prevalence. History of an eating disorder is a disqualifying condition for entry into any branch of the U.S. military.⁸ Once in the military, eating disorders interfering with duty performance are grounds for separation.^8,9 For these reasons, the actual prevalence of eating disorders in active duty and veteran populations may be higher than recent studies suggest.

What is especially concerning is that some aspects of U.S. military service have the potential to increase the possibility of developing an eating disorder.⁷ Military culture emphasizes sacrifice and service and is defined by stressful events like deployments and exposure to violence.¹⁰ Stressful life events can lead to the development of an eating disorder.⁷ Active duty service members are required to maintain strict, service-specific physical fitness and weight standards that may be related to the development of an eating disorder.^7,11 Furthermore, higher rates of PTSD and MDD found in these populations may predispose patients to eating disorders.¹² Women in the U.S. military who experience sexual harassment or assault are at an increased risk of developing eating disorders.¹³

In addition to the negative effects an eating disorder has on a service member’s physical and mental health, the disorder can negatively impact readiness. Patients with anorexia nervosa have nonspecific symptoms including difficulty concentrating, dizziness, and fatigue.¹ Patients with bulimia nervosa who participate in purging behavior may have electrolyte imbalances and acidosis.¹ Such symptoms in a deployed setting would be difficult to evaluate and manage, potentially requiring medical evacuation. Even in robust clinical settings, these conditions can consume a great deal of medical resources.

Given the impact of eating disorders on patients, the likelihood that the true prevalence is higher in U.S. military and veteran populations, the risk factors associated with developing an eating disorder inherent to military service, and their impact on readiness, a comprehensive understanding of eating disorders in the U.S. military and veteran population is needed.

The last systematic review of eating disorders in active duty and veteran populations was published in 2015.⁷ Since then, several studies were published.^6,14,15 Additionally, the Diagnostic and Statistical Manual of Mental Health Disorders (DSM) was updated in 2013, which included clarifications to anorexia nervosa and bulimia nervosa as well as introducing binge eating disorder.¹ These changes may affect how prevalence is recorded, and resources are allocated to patients with diagnosed eating disorders. For example, amenorrhea is no longer a criterion for the diagnosis of anorexia nervosa, which may be due to the use of hormonal contraception and diagnoses in men.¹ Of note, the 2022 DSM update did not include any changes to eating disorders.¹⁶ In addition to new diagnostic criteria, the treatment of eating disorders has become more holistic in recent years.¹⁷ Comprehensive understanding of the neuroscience behind the development of eating disorders drives the need for an updated review.¹⁸

The purpose of this study is to systematically review the recent published literature on the prevalence, co-occurring disorders and events, and health care utilization of individuals with eating disorders in active duty U.S. military and veteran populations.

METHODS

A systematic review of eating disorders in U.S. military and veteran populations was conducted using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).¹⁹

Search Strategy

Databases included PubMed, PsycINFO, CINAHL, Embase, and the Defense Technical Information Center (DTIC). MeSH Terms used can be found in  Appendix A. Searches were completed in November 2020.

Screening

Original studies completed after the previous review⁷ on the prevalence, co-occurrence of other disorders/events, and health care utilization of patients with eating disorders or behaviors suggesting eating disorder pathology in all branches of the U.S. military and veteran population were included. Case studies, opinion pieces, and studies on non-U.S. populations were excluded. Review articles were excluded, but the articles included in these reviews were cross-referenced with our search to ensure all primary literature was included. Preliminary screening for inclusion focused on titles and abstracts and then subsequently on full-length articles and was conducted by two reviewers (DT and MQ), and any discrepancies were resolved by a third reviewer (EM). Results of each reviewer’s findings were tracked using the Covidence Systematic Review software package.²⁰ Cohen’s Kappa was reported for each level of screening.

Data Extraction

Included articles related to active duty or veteran patients with an eating disorder were divided into three categories—incidence, co-occurring conditions or events, or health care utilization. Articles that reported incidence as a demographic description in describing co-occurring conditions/events or health care utilization were categorized as the outcomes related to these latter findings.

Each article was extracted by the lead author (DT) and one of the other authors (MQ or VF). A separate reviewer (EM) was available to resolve any conflicts. Extraction included first describing the specific research question(s) the article was attempting to answer within its respective category. For articles describing incidence, descriptive statistics of incidence or prevalence were reported. For articles describing co-occurring disorders/events, descriptive statistics, odds ratios, and risk ratios were reported. Lastly, for articles that described health care utilization, descriptive statistics describing hospitalizations, appointments, and costs were reported. Study sample, study size, demographic information, and clinical assessment tools were also noted (Table I).

TABLE I.

Prevalence of Eating Disorders in the Military

Study and quality of evidence	Sample	Study aim(s)	Eating disorder assessment	Findings
Williams, V. F. (2018)6 Population: level 1	Active duty members with eating disorder diagnoses (n = 1788) receiving care at fixed MHS facilities or civilian facilities via TriCare from 2013 to 2017 Demographics: 67% Female, 59.5% White	To describe the comparative incidence of diagnoses of AN, BN, and OUED among active duty service members during 2013-2017	ICD-9/10 CM diagnostic codes in the EMR for AN, BN, and OUED	In both genders, OUED was the most prevalent ED in active duty service members. Overall incidence rates were higher in the youngest age group (≤29), non-Hispanic white servicewomen, Marine Corps members, junior enlisted or officers, and combat-specific occupations.
Bankoff, S. M. and Wolf, E. J. (2016)34 Population: level 1	Male veterans (n = 642) with trauma history in the Knowledge Networks-GfK Research Panel Demographics: Avg age: 64.11 (SD = 11.20) (22-89 range); 85.5% Caucasian; 72.2% Married; 44.1% Bachelor’s degree or higher	To examine disordered eating among nonheterosexual male veterans and the association between BMI, symptoms of AN, BN, and BED, and food addiction	EDDS, YFAS, and self-reported height and weight	Sexual minority orientation was not significantly associated with BMI when controlling for age but was significantly associated with higher EDDS scores. Sexual minority orientation was significantly associated with higher YFAS scores.
Buchholz, L. J. (2018)35 Population: level 2	Women ages 18-65 within the Upstate New York VA system who attended at least one primary care MH appointment in FY 2015 (n = 176) Demographics: NR	To evaluate the prevalence of eating disorder symptoms in women veterans receiving mental health care	EDE-Q	MST is unrelated to disordered eating, restraint, eating concerns, weight concerns, and shape concerns; PTSD and anxiety are related. Women veterans’ most reported objective and subjective binge episodes.
Slane, J. D. (2016)36 Population: level 2	Veterans from the conflicts in Iraq and Afghanistan who participated in the Women Veterans Cohort Study and completed the primary eating measure (n = 662) Demographics: 54.9% Female, 66.5% White	To document the rates of eating disorders, disordered eating behaviors, such as binge eating, and associated psychopathology among veterans returning from Iraq and Afghanistan and examine sex differences in these behaviors and co-occurring conditions	ESRE scale; self-reported history of AN, BN, or BED	Female veterans self-reported disordered eating at higher rates compared to male veterans. Both genders who engage in disordered eating had higher rates of PTSD and MDD. Women engaging in disordered eating had greater rates of alcohol abuse compared to women without disordered eating patterns.

Abbreviations: MHS, military health system; AN, anorexia nervosa; BN, bulimia nervosa; OUED, other unspecified eating disorder; ICD, International Classification of Disease; EMR, electronic medical record; BED, binge-eating disorder; MST, military sexual trauma.

Categorization of Studies

Based on extracted data, each article was collaboratively graded by all authors using the Oxford Centre for Evidence-Based Medicine Levels of Evidence v2.1/2011 (CEBM).²¹ These guidelines define the best methods or evidence for different types of research questions. Evidence types are graded using a 5-point scale, with level 1 indicating the best evidence. For example, for studies assessing how common a problem is, a level 1 study would employ “Local and current random sample surveys (or censuses).” A level 1 study examining screening efficacy would require “Systematic review of randomized trials.” An example of a study employing a lower level of evidence (level 3) to examine screening efficacy includes “Non-randomized controlled cohort/follow-up stud[ies].” To rate each study’s level of evidence, we first determined what type of question the study was “answering” (Column 1 on the CEBM Guide) and then assigned the levels based on the CEBM descriptions for that type of research question.²¹

RESULTS

Here, 173 articles were identified after deduplication (Fig. 1).¹⁹ In all, 141 articles were excluded based on their titles and abstracts. Cohen’s Kappa coefficient for this screening was 0.77. The full text of the remaining 34 articles was screened. Ultimately 22 studies were included in this review. Reasons for exclusion in the full-text review included failure to describe an eating disorder,^22,24–28 limited outcomes,²⁹ being a review article,^30,31 studying a population that did not meet the inclusion criteria,^32,33 or being a case report.³⁴ Included articles can be found by prevalence or incidence of eating disorders (Table I), co-occurring conditions/events (Table II), and health care costs and utilization (Table III). Within each table, articles are stratified based on the quality of CEBM evidence.²¹

FIGURE 1.

PRISMA 2020 flow diagram for new systematic reviews.

TABLE II.

Relationship between Eating Disorders and Other Events/Disorders

Study and quality of evidence	Sample	Study aim(s)	Eating disorder assessment	Findings
Arditte Hall, K. A. (2017)38 Population: level 1	Veterans (n = 585) with trauma exposure in Knowledge Networks-GfK Research Panel Demographics: Avg age: 64.11 (SD = 11.20), 85.5% White, non-Hispanic, avg BMI: 26.35 (SD = 5.57)	To examine the associations between eating disorder symptoms and specific types of trauma (i.e., childhood physical abuse, adult physical assault, childhood sexual abuse, adult sexual assault, and military-related trauma)	EDDS	In male veterans, military-related trauma was associated with ED symptoms severity while combat exposure was not.
Arditte Hall, K. A. (2018)39 Population: level 1	Randomly selected woman VA patients (n = 186) from the Northeastern United States Demographics: Avg age: 53.51 (SD = 14.29), 86.6% Caucasian, avg BMI: 28.77 (SD = 6.93)	To examine if five trauma types—childhood physical abuse, adult physical assault, childhood sexual abuse, adult sexual assault, and military-related trauma—were associated with EDs	EDDS	EDDS scores were associated with a history of adult physical assault, sexual assault, and military-related trauma. Military-related trauma was the only type of trauma that was significantly associated with ED severity.
Bartlett, B. A. et al. (2018)22 Population: level 1	Veterans (n = 642 [male] and n = 198 [female]) who endorsed trauma exposure Demographics: 85.51% White, non-Hispanic, 4.52% Black, 4.36% Hispanic, 3.43% mixed race, and 2.18% other race. Avg age = 64.11 (SD = 11.20), 81.78% married/cohabitating	To examine the associations between past-year IPV and DE	EDDS	Physical, sexual, and psychological/emotional forms of IPV were significantly associated with DE.
Breland, J. Y. et al. (2018)40 Population: level 1	Women VA patients (n = 407) aged 18-70 without a 5-year history of suicidal ideation or psychotic disorders Demographics: Avg age: 49 (SD = 13), 40% veterans of color	To investigate MST and combat exposure as independent predictors of eating disorders in female military veterans	Self-report questionnaires based on the SCOFF questionnaire; EDE-Q	Women veterans with MST had twice the likelihood of having an eating disorder compared to women who did not; combat exposure was not associated with eating disorders while military sexual trauma was.
Livingston, W. S. (2019)15 Population: level 1	Veterans (n = 265,806) who separated between 2004 and 2011, deployed to post 9/11 conflicts in Iraq and Iran, and had 5 years of administrative follow-up with the VA Demographics: NR	To examine the association between homelessness and eating disorders in veterans	ICD-9 CM diagnostic codes for AN, BN, and OUED	Compared to domiciled veterans, homeless veterans were more likely to have an eating disorder.
Masheb, R. M. (2019)41 Population: level 1	Overweight/obese veterans (n = 120) seeking weight management treatment and attended an in-person MOVE! Orientation Demographics: 89.2% male; avg age: 61.7 (SD = 8.65); 74.2% Caucasian; avg BMI: 38.0 (SD = 7.30)	To determine internal consistency of the MWI and prevalence of making weight behaviors; to compare those who screened positive and negative on the MWI on measures of weight, binge eating, and related eating pathology	MWI, EDE-Q, YEOQ, mYFAS, NEQ, ISI, QEWP-R, and VA-BES	One-third of veterans who were overweight/obese screened positive for engaging in making weight behaviors during military service
Mitchell, K. S. et al. (2016)42 Population: level 1	The first three cohorts (n = 33,937) enrolled in the Millennium Cohort Study Demographics: 82% male, avg age 36.0 (SD = 8.9), 72% White, 46% Army, 70% enlisted	To investigate longitudinal associations between PTSD, disordered eating, and weight gain	PHQ	PTSD was associated with weight change from time 2 to time 3 via disordered eating. In Caucasian men, there was a significant association between PTSD and weight gain from compensatory behaviors.
Mitchell, K. S. and Wolf, E. J. (2016)43 Population: level 1	Veterans (n = 697) in the Knowledge Networks-GfK Research Panel Demographics: 92% male, avg age: 62.99 (SD = 12.03), 84.6% Caucasian, 76.8% married, 87.0% college educated	To record rates of PTSD, EDs, and food addiction and assess associations between PTSD, food addiction, and ED symptoms in males; to explore the role of emotion regulation in linking PTSD to ED symptoms and food addiction symptoms	EDDS and YFAS	PTSD was directly associated with ED symptoms, food addiction, and expressive suppression in the sample.
Rosenbaum, D. L. (2016)44 Population: level 1	Women veteran established primary care patients (n = 484) with at least 2 visits in the past year termed “mental health stakeholders” Demographics: 2.5% Asian or NH/PI/AI/AN, 24.8% Black, 61.2% White, 7.4% Hispanic, 38.5% married, 37.7% divorced	To investigate prioritization of BED and weight-related influences on treatment services. To examine potential treatment indicators and patient-driven priorities for mental health care, in women with and without BED	PHQ; BED screening, height, and weight information from medical records	Within the BED group, there was a greater frequency of obesity, depression, PTSD, mood disorders, sleep disorders, hypertension, chronic pain, and lipid disorders. Treatment priorities for this group included mood concerns, weight loss, and body image/food issues.
Hoerster, K. D. (2015)45 Population: level 2	Veterans (n = 332) from Iraq/Afghanistan at the VA Puget Sound Health Care System, Seattle Demographics: 91.5% male, avg age: 31.1 (SD = 8.5), 72.6% Caucasian, 75.4% army	To examine the role of PTSD and depression symptoms in association with binge eating in Iraq and Afghanistan Veterans	PHQ	Participants meeting depression and PTSD criteria as well as symptom severity were more likely to meet binge eating screening criteria.
Huston, J. C. (2019)46 Population: level 2	Women Veterans (n = 190) who participated in KnowledgePanel at three timepoints Demographics: Avg age: 54.06 (SD = 14.00), 70.16% White, non-Hispanic, 18.98% Black, non-Hispanic, 52.59% married, 25.02% divorced	To examine the direct impact of IPV on ED symptoms, as well as the mediating effects of depression and PTSD symptoms, in a longitudinal cohort of female veterans.	EDDS	IPV was indirectly related to EDDS scores when measuring PTSD and depression symptoms. IPV, PTSD, and depression symptoms were significantly associated with eating disorder symptoms.
Blais, R. K. (2017)47 Population: level 3	Veterans from the VA clinical data 1- (n = 595,525) or 5-year follow-up (n = 265,806) Demographics: data from the 2011 OEF/OIF roster file and VA clinical data	To evaluate the association between military sexual trauma and an eating disorder diagnosis in veterans of OIF and OEF	ICD-9 CM codes for AN, BN, and OUED	In both cohorts, a history of MST in both male and female veterans was associated with the co-occurrence of an ED.
Dorflinger, L. M. (2017)48 Screening: Level 3	Veterans referred to a weight management program at Connecticut VA (n = 110)	To examine NES among veterans seeking weight management treatment.	NEQ, ISI, PHQ-2, PC-PTSD, EDE-Q, EOQ, MOVE!23, and QEWP-R	In veterans, NES was associated with binge eating and overeating in response to emotions.
	Demographics: 90% male, 75% Caucasian, 18.3% Black, avg. age: 61.6 (SD = 8.5), avg. BMI: 38.0 (SD = 7.4)
Dorflinger, L. M. (2017)49 Screening: level 3	Veterans (n = 116) referred to a primary care weight management program Demographics: 88.8% male, avg age: 61.66 (SD = 8.73), 93.1% non-Hispanic, 75% Caucasian, 20.9% black, avg BMI 37.90 (SD = 7.35)	To analyze the validity of the BES	QEWP-R	The single-item measure of BES is comparable to established screening tools in sensitivity, specificity, and negative predictive value and can be used to screen BED in a primary care setting.
Breland, J. Y. et al. (2018)2 Prognosis: level 4	Women veterans (n = 20) aged 18-70 with DE Demographics: Avg age: 48 (SD = 15); 55% women of color	To understand the relationship between disordered eating and trauma exposure	SCOFF questionnaire	Disordered eating causes short-term relief from the negative affect and maladaptive thoughts and is used as a mechanism to avoid potential trauma.

Abbreviations: IPV, intimate partner violence; DE, disordered eating; SCOFF, sick, control, one, fat, food; YEOQ, yale emotional overeating questionnaire; mYFAS, modified yale food addiction scale 2.0; YFAS, yale food addiction scale; NEQ, night eating questionnaire; ISI, insomnia severity index; PHQ, patient health questionnaire; OEF/OIF, Operation Enduring Freedom/Operation Iraqi Freedom; NES, night eating syndrome.

TABLE III.

Health Care Costs and Utilization of Patients with Eating Disorders

Study and quality of evidence	Sample	Study aim(s)	Eating disorder assessment	Findings
Bellows, B. K. (2015)14 Population: level 1	VA patients with BED diagnosis from 2000 to 2011 (n = 1487) Demographics: NR	To compare the one-year health care costs and utilization of patients with BED to patients with EDNOS or NED	NLP-identified diagnosis of BED in EMR EDNOS ICD-9 code 307.5 on two or more separate encounters	Higher costs were associated with patients with BED, costing $5,589 more than EDNOS-only and $18,152 more than patients with no ED.
Bellows, B. K. (2016)37 Population: level 1	VA patients between 2000 and 2011 with at least 1 year of pre/post index diagnosis care and no history of eating disorder diagnosis other than BED or EDNOS (n = 3842) Demographics: 68.5% male, Avg age: 49.2 (SD = 11.0), 70.7% White, avg BMI: 38.8 (SD = 10.2), 79.7% BMI ≥ 30	To describe the characteristics of treatments of patients with BED and to compare treatment modalities to EDNOS-only and NED patients	ICD-9 CM diagnostic code in the EMR, NLP-identified diagnosis	Treatment modalities between EDNOS-only patients and NED patients differed, as did demographic information, indicating differences in populations.
Huston, J. C. (2018)50 Population: level 1	Women VA patients 18 years or older in New England (n = 198) Demographics: Avg age 54.09 (SD = 14.83); 87.37% White	To examine VA health care use among female veterans with an eating disorder diagnosis	EDDS	Female veterans with EDs reported significantly greater use of all types of mental health care services, and significantly worse physical and mental health functioning, than female veterans without EDs.

Abbreviations: NED, no eating disorder; NLP, natural language processing.

Prevalence of Eating Disorders in the U.S. Military

Four articles discussed the prevalence of eating disorders in U.S. active duty and veteran populations (Table I). Two articles employed level 1 CEBM²¹ evidence,^6,35 and two articles employed level 2 CEBM evidence. Both level 2 articles were retrospective reviews of survey data.^36,37

Williams et al.⁶ reported the incidence of eating disorders in the U.S. military as 2.7%, with the most common diagnosis being other specified feeding or eating disorder (46.4%) followed by bulimia nervosa (41.8%) and anorexia nervosa (11.9%). The overall incidence rates were higher in the younger age group (≤29), non-Hispanic white servicewomen, Marine Corps members, junior enlisted or junior officers, and combat-specific occupations.⁶ Patients with binge eating disorder were more likely to be younger, male, and obese compared to patients with eating disorder not otherwise specified.³⁸

Slane et al.³⁷ found female veterans returning from Iraq and Afghanistan self-reported higher rates of disordered eating compared to their male counterparts (18.6% and 7.9%, respectively). Binge eating disorder diagnoses were lower for both male and female veterans compared to civilian samples (0.4% and 1.2%, respectively, compared to 2.0% and 3.5%, respectively).³⁷ Gay, bisexual, or other sexual minority veteran men had higher Yale Food Addiction Scale (YFAS) scores, higher BMI, and Eating Disorder Diagnostic Scale (EDDS) scores compared to heterosexual men. Younger age and higher BMI predicted higher EDDS scores in nonheterosexual veterans.³⁵

The most common occurrence of eating disorder symptomatology was objective binge episodes (16.5%) followed by subjective binge episodes (13.6%), dietary restraint (10.2%), laxative misuse (2.3%), self-induced vomiting (0.6%), and excessive exercise (4.0%) in a sample of veteran women.³⁶ Within this population, 9.7% of participants scored in the clinically significant range on the Eating Disorder Examination Questionnaire (EDE-Q).³⁶ In a nationwide sample of veterans, common eating disorder behavior were binge eating (21.4%), fasting (11.7%), excessive exercise (6.7%), use of laxatives or diuretics (2.3%), and vomiting (1%).³⁹

Relationship between Eating Disorders and Other Events/disorders

Fifteen articles discussed comorbidities of individuals with eating disorders in active duty and veteran populations (Table II). Nine articles had level 1 CEBM²⁰ evidence,^{15,23,39–45} two had level 2 evidence,^46,47 three had level 3 evidence,^48–50 and one had level 4 evidence.² Non-level 1 articles were cross-sectional surveys or retrospective studies. Comorbid disorders or events included PTSD, MDD, trauma, homelessness, and intimate partner violence.

Only military-related trauma was significantly associated with eating disorder symptomatology and severity in male and female veterans.^39,40 However, in women, history of adult physical assault and adult sexual assault were each associated with more severe eating disorder symptomatology.⁴⁰ Veteran women who reported military sexual trauma had twice the likelihood of having an eating disorder compared to women who did not.^41,47 Eating disorder likelihood based on military sexual trauma was higher among male veterans of Operation Enduring and Iraqi Freedom compared to female veterans.⁴⁸ Combat exposure defined as accidents, traumatic loss, or moral injury was not associated with eating disorder symptom severity.^39,41,47 Participants who reported a greater number of trauma events, those who were younger, and those with higher BMIs reported more severe eating disorder symptomatology.³⁹ Other associations with an eating disorder included identifying as Asian and age.⁴¹

In a sample of Veteran Health Administration (VHA)-enrolled veterans, the odds of having an eating disorder diagnosis were 59% higher among veterans experiencing homelessness, and 0.6% of homeless veterans had an eating disorder diagnosis.¹⁵ Eating disorder not otherwise specified was the most common (n = 24) followed by bulimia nervosa (n = 22).¹⁵ Veterans with an eating disorder were more likely to have PTSD, MDD, alcohol use disorder, and SUD.¹⁵ These veterans were most frequently young, never married, and white women with comorbid depression, PTSD, or substance use.¹⁵

Hoerster et al.⁴⁶ examined PTSD and MDD symptoms in connection with binge eating in Iraq and Afghanistan veterans. Participants who met MDD and PTSD screening criteria were more likely to meet binge eating criteria.⁴⁶ In both genders, veterans with self-reported disordered eating had a significantly higher rate of PTSD and alcohol abuse than those without.³⁷ PTSD was directly associated with eating disorder symptoms and weight change as well as indirectly associated with eating disorder symptoms mediated by expressive suppression.^43,44

Veteran women in a primary care setting with binge eating disorder had a greater frequency of obesity, depression, PTSD, mood disorders, sleep disorders, hypertension, chronic pain, and lipid disorders.⁴⁵ Eating disorder symptoms were significantly associated with PTSD and depression symptoms as well as lifetime intimate partner violence.^23,47

In veterans, Emotion- and Stress-Related Eating (ESRE) scores and BMI were significantly correlated.³⁷ Night eating syndrome was associated with a higher BMI as well as higher scores on measures of binge eating, emotional overeating, and eating disorder symptomatology in veterans.⁴⁹ Veterans screened for PTSD and depression were more likely to endorse eating to get back to sleep after waking at night.⁴⁹ Rates of unhealthy “making weight behaviors” were higher in younger, ethnic/racial minority (60%), female (53.8%) participants.⁴² These participants were more likely to engage in binge eating (84.2%) and compensatory behavior (53.8%).⁴¹

Dorflinger et al.⁵⁰ examined the validity of the Veteran Affairs Binge Eating Screener (VA-BES) in identifying binge eating disorder in weight loss seeking veterans. VA-BES demonstrated good sensitivity [88.9 (8/9)] and specificity [83.2 (89/107)] with a negative predictive value of >98% (89/90) when a cutoff of ≥ 2 episodes of binge eating a week was applied.⁵⁰ When compared to the established screening tool Questionnaire on Eating and Weight Patterns-Revised (QEWP-R), the positive predictive value was low [30.8 (8/26)].⁵⁰

Breland et al.² noted that trauma-related disordered eating is often a mechanism to address maladaptive thoughts and negative affect and often used as a defense mechanism to avoid unwanted attention.

Health Care Utilization by Individuals with Eating Disorders

Three articles discussing the health care utilization of individuals with eating disorders in active duty and veteran populations (Table III) had level 1 CEBM²¹ evidence and represented cohort studies.^14,38,51 Of veteran affairs (VA) patients, health care costs for patients with binge eating disorder were $5,589 higher than patients with eating disorder not otherwise specified (EDNOS) and $18,152 higher than patients without an eating disorder.¹⁴ Patients with eating disorder not otherwise specified and binge eating disorder used at least one treatment modality including psychotherapy and pharmacotherapy compared to patients without an eating disorder.³⁸ Compared to peers without eating disorders, 1 in 10 female veterans were found to have an eating disorder with a higher likelihood of using VA mental health services (including outpatient services, medication, and SUD).⁵¹

DISCUSSION

We performed a systematic literature review on eating disorders in U.S. military active duty service members and veterans. Twenty-two articles have been published on this topic since the last literature review,⁷ and 13 were constructed using the highest level of evidence. These recent efforts have expanded our understanding of eating disorders in active duty and veteran populations. Since the last review,⁷ there have been published studies that used investigations of eating disorders among nationally representative samples of veterans.^{14,35,37–39,44,48}

Prevalence of Eating Disorders

Active duty service members remain understudied population as only two articles were not focused on veterans.^6,43 Since most studies used a retrospective chart review,^{6,14,15,38,48} the results are likely more indicative of rates of help seeking behavior or clinician likelihood to diagnose an eating disorder than the actual prevalence of eating disorders in the U.S. military and veteran population. It is unclear if the branch of service, gender, age, and sexual orientation increases the likelihood of developing an eating disorder. Future prospective work examining a representative sample of service members and veterans is needed to get the true prevalence of eating disorders in these populations based on demographics and to determine if there are differences in help seeking or diagnoses as compared to these true prevalence rates.

Overview of Comorbid Disorders or Events

Eating disorders are commonly associated with PTSD, MDD, and sexual trauma in active duty/veteran populations.⁴⁸ Noncombat-related military trauma was associated with disordered eating.² Therefore, clinicians caring for active duty or veteran patients with any of these conditions/situations should carefully assess for concomitant conditions. This is especially true for men with a history of military sexual trauma, as they have twice as likely to have an eating disorder compared to female veterans.^41,48 Screening should be repeated throughout treatment, as individuals with disordered eating frequently do not seek care for these symptoms until later in the course of their concomitant disorders.^15,43,47 This may be due to patients viewing their eating disorder symptomatology as a method for taking control. Specifically, mental health clinicians caring for active duty or veteran patients with PTSD may want to screen patients who engage in maladaptive suppression of emotions for an eating disorder as the findings of multiple studies suggest the role of disordered eating as a mode of emotional regulation in those with history of trauma exposure.^2,41,44,49

Efforts to “make weight” (i.e., weigh less than a specific target based on gender/age) to meet U.S. military standards are related to disordered eating.⁴² However, cross-sectional study designs limit the examination of causal relationships between “making weight” and disordered eating.^35,46 Since the previous review in 2015,⁷ several changes implemented to military fitness testing include a new Army fitness test,⁵² alternative cardio options in the Navy,⁵³ and removal of waist circumference measurements in the Air Force.⁵⁴ Recent research⁵⁵ highlighted how physical fitness standards may have unintended consequences in the Marines.

One way to help identify active duty patients with eating disorders is to provide routine screening for populations at risk. Our review, along with the previous review,⁷ identified several concomitant conditions with eating disorders.^{2,36,37,39–41,43–46,48,51} Routine screening for MDD, PTSD, and suicide are standard in the U.S. military. Adding screening methods for eating disorders will help identify patients with eating disorders and determine the level of care needed for these patients (e.g., VA-BES, ESRE, making weight inventory (MWI), and QEWP-R). Screening tools were shown to be a valid single-item measure of screening for binge eating in primary care with good sensitivity, specificity, and negative predictive value.⁵⁰ Masheb et al.⁴² indicated that the “Making Weight Inventory” had good internal consistency and could be used to assess for service member’s use of unhealthy weight loss strategies to meet weight requirements.

Health Care Utilization Related to Eating Disorders

Veteran women with binge eating disorder experience a combination of mood symptoms, dietary concerns, substance use difficulties, and difficulties with body.^45,51 These findings highlight potentially complex clinical profiles of veteran women with binge eating disorders and the resources needed to support them. Patients with binge eating disorder have higher 1-year total health care costs compared to patients without eating disorders,¹⁴ but these data may represent inadequate treatment (and thus the need for ongoing follow-up). Furthermore, these data may underrepresent the treatment needs of these patients, as treatment for eating disorders is often nonexistent in the rural communities many veterans live in.

Access to treatment for eating disorders is limited for active duty service members and their family members. Outside organizations, such as the Eating Disorders Coalition, have raised concern about the access to specialty care for eating disorders within the U.S. military health system.⁵⁶ Only 35% of U.S. civilian eating disorder treatment facilities accept TriCare, and facilities may not be located near large military populations.⁵⁶ Despite being identified as a crucial part of treatment and recovery, the TriCare network does not currently reimburse for dietician services (medical nutrition therapy) for any eating disorders.⁵⁶

Limitations Found in the Literature

In addition to the limitations specific to each of these three above-mentioned research areas, there were common methodological limitations found in the studies included in this review. Many of the studies relied on self-reporting questionnaires,^{35–37,39–47,49–51} which can suffer from subjectivity.⁵⁷ In addition to a paucity of work on active duty populations, the included studies may have limited generalizability to the overall active duty/veteran population due to the narrow sample of participants.^36,40,49,51 Other limitations include high dropout rates^36,37,47 and small sample size.^{2,15,35,40,42,45,48,49,51}

Several studies were challenged by their screening methods. The ESRE scale makes it unclear if eating behaviors are pathological.³⁷ In studies that measure military-related trauma, the exact nature of trauma is vague. One article described military-related trauma as accidents, traumatic loss, or moral injury,³⁹ while another article defined it as “seeing something horrible or being badly scared” during military service.⁴⁰ Intimate partner violence was assessed using a different measure than typically used in research.⁴⁷ Dorflinger et al.⁵⁰ used the QEWP-R to detect binge eating instead of the Eating Disorder Examination interview.

Challenges of Studying Eating Disorders in the Military

Few clinicians in the U.S. military have the required specialized training to treat severe eating disorders.⁵⁸ As a result, U.S. military patients with complicated eating disorders are referred to civilian medical facilities, which use different documentation. Studies that evaluated the rate of eating disorder diagnosis by looking at the military outpatient medical record coding likely missed cases that were too acute or severe to be managed within the Department of Defense (DoD).^{6,14,15,38,48} Even if documentation was sent to the DoD, it would not be coded within the DoD medical record and thus would not appear in any database query. Additionally, while active duty service members have access to free health care, they may avoid seeking treatment since an eating disorder diagnosis can result in medical separation.

The U.S. military uses a variety of different medical records systems. Currently, data are not automatically shared between inpatient and outpatient medical records. In some facilities, records from emergency department visits are stored in an entirely separate record system. None of the studies that were included mentioned reviewing any medical records systems other than the one that is routinely used for standard outpatient care resulting in underreporting of care occurring in emergency rooms or inpatient units.

Limitations of Our Review

We did not cross-reference all the citations in the articles that met inclusion criteria (“snowballing”), potentially missing studies not indexed in the databases searched. Our Kappa for inclusion was lower than desired; however, our use of a third reviewer for any discrepancies likely mitigated this potential shortcoming. Because literature reviews were excluded, two articles^30,31 may have had higher levels of CEBM evidence for research questions not related to incidence or concominence.²¹ For the two review articles,^30,31 we cross-referenced the articles included in each review to confirm all primary literature was included in our search.

CONCLUSION

Diagnosing and treating patients with eating disorders present unique challenges in the U.S. military. Continued efforts to improve care for these individuals are critical. The next step in this investigation should include population-based studies assessing prevalence and comorbidities along with randomized control trials of assessment tools and treatment options. This could be accomplished by adding screening for eating disorders to the U.S. military’s annual health screening or by including cutoffs for low BMIs on annual physical fitness testing—like current upper BMI limits. Improved detection of an eating disorder in patients with comorbid conditions can assist with individualized patient care, further improving health care utilization and possibly reducing costs in this population.

ACKNOWLEDGMENTS

Special thanks to the contributions of Rhonda J. Allard, MLIS at the James A. Zimble Learning Resource Center.

FUNDING

None declared.

CONFLICT OF INTEREST STATEMENT

None declared.

APPENDIX A MeSH Terms

Total Results: 427 - 254 duplicates removed = 173 results

PubMed, 11/05/2020 – 112 results:

(“Military Personnel”[mh] OR “Veterans”[mh] OR active-duty[tw] OR air-force[tw] OR army[tw] OR deployed[tw] OR navy[tw] OR military[tw] OR servicemember*[tw] OR service-member*[tw] OR soldier*[tw] OR veteran*[tw]) AND (“Bulimia“[Mesh] OR ”Feeding and Eating Disorders”[Mesh] OR abnormal-eating[tw] OR anorex*[tw] OR binge-eating[tw] OR bulimi*[tw] OR dietary-behavior*[tw] OR disordered-eating[tw] OR eating-behavior*[tw] OR eating-disorder*[tw] OR eating-pattern*[tw] OR food-addiction[tw] OR night-eating-syndrome[tw] OR pica[tw]) AND english[lang] AND (2015/2/19:3000/12/12[pdat]).

Embase, 11/05/2020 – 127 results:

(‘military personnel‘/exp OR ‘veteran’/exp OR (active-duty OR air-force OR army OR deployed OR navy OR military OR servicemember* OR service-member* OR soldier* OR veteran*):ab, ti, kw) AND (’eating disorder’/exp OR (abnormal-eating OR anorex* OR binge-eating OR bulimi* OR dietary-behavior* OR disordered-eating OR eating-behavior* OR eating-disorder* OR eating-pattern* OR food-addiction OR night-eating-syndrome OR pica):ab, ti, kw) AND [2015-2020]/py AND [19-2-2015]/sd NOT [6-11-2020]/sd AND [english]/lim AND ([article]/lim OR [article in press]/lim OR [review]/lim)

CINAHL, 11/05/2020 – 98 results:

(MH “Military Personnel+ “OR MH “Military Deployment+” OR MH “Military Recruits”

OR MH “Veterans+“ OR (active-duty OR air-force OR army OR deployed OR navy OR military OR servicemember* OR service-member* OR soldier* OR veteran*)) AND (MH ”Eating Disorders+” OR (abnormal-eating OR anorex* OR binge-eating OR bulimi* OR dietary-behavior* OR disordered-eating OR eating-behavior* OR eating-disorder* OR eating-pattern* OR food-addiction OR night-eating-syndrome OR pica))

Limiters - Published Date: 20,150,201-; English Language; Peer Reviewed.

PsycINFO, 11/05/2020 – 64 results:

1. (exp military personnel/ OR exp Military Veterans/ OR (active-duty OR air-force OR army OR deployed OR navy OR military OR servicemember* OR service-member* OR soldier* OR veteran*).tw.) AND (exp eating disorders/ or binge eating/ or food addiction/ OR (abnormal-eating OR anorex* OR binge-eating OR bulimi* OR dietary-behavior* OR disordered-eating OR eating-behavior* OR eating-disorder* OR eating-pattern* OR food-addiction OR night-eating-syndrome OR pica).tw.)

2. limit 1 to (peer reviewed journal and english language and yr=“2015 - 2021”)

DTIC, 11/09/2020 – 26 Results:

1. (active-duty OR air-force OR army OR deployed OR navy OR military OR servicemember* OR service-member* OR soldier* OR veteran*) AND (abnormal-eating OR anorex* OR binge-eating OR bulimi* OR dietary-behavior* OR disordered-eating OR eating-behavior* OR eating-disorder* OR eating-pattern* OR food-addiction OR night-eating-syndrome OR pica).

Author notes

The views expressed are those of the authors and do not necessarily reflect the views of the U.S. Government, the U.S. Department of Defense, the Uniformed Services University, the Department of Health and Human Services, or the U.S. Public Health Service.