ABSTRACT
Introduction
Eating disorders are often under-detected, which poses a serious threat to the health of individuals with eating disorder symptoms. There is evidence to suggest that the military represents a subpopulation that may be susceptible to high prevalence of eating disorders and vulnerable to their underdiagnosis. Underreporting of eating disorder symptoms in the military could lead to this underdiagnosis of individuals with eating disorder symptoms. The purpose of this study was to examine the association between military affiliation and eating disorder symptoms among college students and the likelihood of eating disorder diagnosis among those with eating disorder symptoms using a large, diverse college-aged sample of both military-involved and civilian students.
Materials and Methods
Participants for this study were from the 2015-2016, 2016-2017, and 2017-2018 Healthy Minds Study (HMS). Healthy Minds Study is a large, cross-sectional cohort study of both undergraduate and graduate students from universities and colleges across the United States and Canada. The Healthy Minds Study survey questions include assessment of demographic information, military status, self-reported eating disorder symptoms using the SCOFF questionnaire, and self-reported eating disorder diagnosis. Univariate analysis, chi-square analysis, and logistic regression with an unadjusted and covariate adjusted model were used to examine the association between eating disorder symptoms and military affiliation. These analyses were also used to examine the association between eating disorder diagnosis among those with eating disorder symptoms and military affiliation. All analyses were conducted using SPSS.
Results
The prevalence of eating disorder symptoms was high among both the civilian (20.4%) and military-involved (14.4%) students. Among females, there was a significantly higher (P value = .041) prevalence of eating disorder symptoms among civilian college students (24.7%) compared to military-involved students (21.3%). Among those with eating disorder symptoms, the prevalence of diagnosis was low in both military and civilian students. Specifically, the prevalence of diagnosis was significantly lower (P value = .032) among military-involved college students (10.8%) compared to civilian college students (16.4%). Differences in sociodemographic characteristics (e.g., gender, race/ethnicity, and age) among military-involved and civilian college students appear to explain this association.
Conclusions
The underdiagnosis of eating disorders is far too common, and this represents a threat to the health of military and civilian populations alike. Underdiagnosis of eating disorders within military environments may be due to underreporting, particularly among men and racial/ethnic minority groups.
INTRODUCTION
Eating disorders are a serious and growing health concern among the general population, affecting both men and women.1,2 Contrary to popular belief, can affect a wide range of individuals from different races/ethnicities,3 socioeconomic backgrounds,3 and levels of body satisfaction.4 The median age of onset of eating disorders is 18-21 years of age.5 The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), defines full-syndrome eating disorders (anorexia nervosa, bulimia nervosa, binge eating disorder, other specified feeding or eating disorder, and unspecified feeding or eating disorder) and outlines the diagnostic criteria, risk factors, comorbidities, and other characteristics associated with each disorder.6 Full-syndrome eating disorders involve a combination of symptoms that meet specific criteria for frequency and severity such as dietary restriction, compensatory behavior, binge eating, significantly low body weight, body image disturbance, and weight/shape overvaluation.6 The majority of individuals who have symptoms of eating disorders do not meet the full-syndrome eating disorder criteria and are considered to have subthreshold eating disorder symptoms.7,8 Research suggests those with subthreshold eating disorder symptoms are at an elevated risk for adverse physical9 and mental10 health outcomes as well as a greater risk for full-syndrome eating disorders.11
Eating disorders are often under-detected, which poses a serious threat to the health of individuals with eating disorder symptoms. Individuals who are not diagnosed are unlikely to be referred to evidence-based treatments known to improve prognosis and mitigate consequences associated with eating disorders. Staggeringly, only one-third of individuals with an eating disorder ever receive the treatment they need.12 There are certain subgroups that are particularly vulnerable for under-detection and undertreatment.13 Some of these underdiagnosed subgroups include men, racial/ethnic minorities, and those with subthreshold eating disorder symptoms that do not meet the criteria for full-syndrome eating disorder classification.13 Although it has yet to be studied, there is related evidence to suggest that the military may represent another demographic that could be at elevated risk for underdiagnosis and undertreatment.
There is some research to suggest that the prevalence of eating disorders among military populations is typically higher than the U.S. population.14–17 Although individuals with eating disorders may self-select into the military,18 there are also several hypothesized aspects of the military environment that may contribute to eating disorder risk. Some of the potential reasons the military represents an environment at elevated eating disorder risk is there are high rates of trauma,19–21 strict physical fitness requirements,20 frequent weigh-ins,20,22 weight regulations,20,22 deployments, and high rates of other psychiatric illness. Despite the military being an environment with potential elevated eating disorder risk, eating disorders may be underreported in military environments. Individuals may fear reporting because they feel pressure to uphold the dominant image that they represent,23 it could impair their likelihood of promotion,24,25 it could disrupt unit cohesion,25 they would be seen as weak,25 or because they had a psychological referral rather than a medical referral.26
Although only a few studies have reported the prevalence of eating disorders in the military, these studies find that the prevalence is at least equal to or higher than population estimates seen in the general public. One study found eating disorder prevalence in the military to be 6% for females and 4% for males and within 2.7 more years of continued service another 3% of females and 3% more males had developed an eating disorder.27 Other studies report prevalence estimates ranging from 1% to 13% depending on the type of eating disorder.14–16 Prevalence of subthreshold eating disorder symptoms are even more common with estimates ranging from 20% to 63% of military members engaging in disordered eating habits.14–16,28 The aforementioned studies rely on anonymous, self-report assessment techniques for their data collection. Conversely, studies that have utilized diagnosis codes from medical records or hospital data to assess the prevalence of eating disorders report much lower prevalence than studies relying on self-reported symptoms. Depending on the type of eating disorder, prevalence of full-syndrome eating disorders according to medical record data ranged from 0% to 5%.29,30 According to a 9-year review using medical records, the incidence rate per year of diagnosed eating disorders within the military was 0.3%.30
There appears to be a discrepancy in prevalence of eating disorders within the military between studies that rely on self-report of eating disorder symptoms and those that rely on medical record diagnosis data leading to the hypothesis that those with eating disorder symptoms may be underdiagnosed. However, the discrepancy between self-reported eating disorder symptoms and diagnosis within the same military-involved sample has never been studied, nor has underreporting of eating disorder symptoms in military-involved students versus civilian students. Thus, the presented study aimed to fill these research gaps by exploring underdiagnosis using a large, diverse college-aged sample of both military-involved and civilian students. The purpose of this study was to examine the association between military affiliation and eating disorder symptoms among college students and the likelihood of eating disorder diagnosis among those with eating disorder symptoms.
METHODS
Study Sample
The Healthy Minds Study (HMS) is a large cross-sectional cohort study of college students that seeks to explore the prevalence and association of different correlates on disordered eating.31 HMS is a voluntary, web-based survey conducted annually and is available to 2- and 4-year colleges and universities in Canada and the United States. Both graduate and undergraduate students partake in the survey among the participating schools. The only inclusion or exclusion criterion is that all participants must be at least 18 years of age. After the university or college elects to participate in the study, the survey is electronically distributed to a list of 4,000 randomly selected students provided by the school. HMS was approved by the institutional review boards on all campuses participating.
The present study utilizes HMS data from three academic years: 2015-2016, 2016-2017, and 2017-2018. The 2015-2016 HMS includes data from 23 different schools with n = 34,299. The 2016-2017 HMS includes data from 54 different schools with n = 53,760. The 2017-2018 HMS includes data from 60 different schools with n = 68,427. During these years, the response rate of the students that the survey was randomly electronically distributed to was 27% in 2015-2016, 23% in 2016-2017, and 23% in 2017-2018. To adjust for potential differences between survey responders and nonresponders and to ensure that all estimates are representative of the full population in terms of basic demographic and other characteristics, we applied sample probability weights, as is typically done in studies using HMS data.31
Of the 156,486 students who participated in HMS from 2015 to 2018, 149,660 (95.6%) provided a response to the question about military involvement. Of these 149,660 individuals, 25,432 did not answer the question about eating disorder symptoms, and an additional 45,867 did not answer the question about eating disorder diagnosis. Of the remaining 78,361 participants, 77 had missing data on gender, 140 had missing data on race/ethnicity, and 951 did not identify as a male or female and were excluded from the sample, leaving a final analytic sample size of 77,193.
Race/ethnicity were reported by participants in the survey. This was assessed using the following question: “Do you consider yourself to be of Hispanic, Latina/o, or Spanish origin?”. In addition, participants were asked: “How do you usually describe your race? (Select all that apply)”. The following were the possible response options: White or Caucasian, Black or African American, Asian or Asian American, American Indian or Native American or Alaskan Native, Middle Eastern or Arab or Arab American, Pacific Islander, Hawaiian Native, or Other (please specify). These options were defined by the investigator. For the purpose of this research, we condensed race/ethnicity into the following subcategories: Hispanic, White, Black, Asian, and Other. If an individual selected Hispanic, then they were classified as being Hispanic, regardless of other additional races reported. The new “Other” subcategory was classified as anyone who checked multiple race categories, other than Hispanic, or anyone who classified themselves as “other race”. Race/ethnicity was collected and adjusted for in the multivariable logistic regression because it is a known correlate of eating disorder risk.32,33
Measures
The HMS survey questions include assessment of demographic information, military status, self-reported eating disorder symptoms, and self-reported eating disorder diagnosis (see Supplementary Table SI for questions and response options). Military status was assessed using the following question: “Have you ever served in the United States Armed Forces, military Reserves, or National Guard?”. The response possibilities included ROTC (Reserve Officers' Training Corps), Reserves, National Guard, and Active Duty and all “yes” responses to this item were collapsed into one “military-involved” category for the remainder of the analyses.
Eating disorder symptoms were measured using the 5-item SCOFF questionnaire.34 Research has found the SCOFF to be a validated, reliable, and effective eating disorder screening tool.34,35 To encourage participants to be forthcoming, the instructions for the SCOFF items in the HMS were as follows: “please answer the following questions as honestly as possible”. Individuals who scored ≥2 on the SCOFF were classified as having symptoms of an eating disorder as is typically done in studies using this eating disorder screening tool.34,35 Setting the threshold at two or more yes answers to the SCOFF questions maximizes the screening tool’s specificity and sensitivity.34
To assess whether someone had been diagnosed with an eating disorder, the following question was used: “Have you ever been diagnosed with any of the following conditions by a health professional (e.g., primary care doctor, psychiatrist, psychologist, etc.)? (Select all that apply)”, with “eating disorder (e.g., anorexia nervosa, bulimia nervosa)” listed as a possible response option.
Analysis
Univariate analyses were conducted to calculate the prevalence of eating disorder symptoms among civilian and military-involved college students in the HMS sample. To examine whether the prevalence of eating disorders in military-involved college students was significantly different than the prevalence of eating disorders in civilian college students, a chi-square test for independence was calculated. Logistic regression was performed to estimate the odds of having eating disorder symptoms according to military status. Both an unadjusted model and a multivariate model adjusted for age,2,36 gender,2,32,33,37 and race/ethnicity,32,33 known correlates of eating disorder behaviors, were performed to estimate the odds of eating disorder symptoms according to military involvement.
Among those with eating disorder symptoms (SCOFF ≥ 2), univariate analysis was conducted to estimate the prevalence of eating disorder diagnosis in both the military-involved and civilian students. To examine whether the likelihood of any lifetime eating disorder diagnosis differs between military-involved and civilian students with eating disorder symptoms, a chi-square test for independence was used. Finally, both unadjusted and adjusted logistic regression models were run to estimate the odds of eating disorder diagnosis among those with eating disorder symptoms according to military involvement.
For all logistic regression models, odds ratios (OR) and 95% confidence intervals (CI) are reported. All analyses were conducted using SPSS version 25.
RESULTS
Of the 77,193 students in the sample, there were 15,671 students (20.3%) with a SCOFF score of ≥2. Symptoms of an eating disorder were seen in 11.9% of males and 24.7% of females. There were 1,549 students (2.0%) who were military-involved in the analytic sample (Table I).
TABLE I.
Sample Characteristics, n (Weighted Percentage)
| Full sample (n = 77,193) | Civilian (n = 75,644) | Military (n = 1,549) | |
|---|---|---|---|
| Age, mean (SD) | 23.06 (6.316) | 22.92 (6.095) | 30.18 (11.045) |
| Parental education (n = 76,642) | |||
| Graduate degree | 30,674 (40.0) | 30,219 (40.2) | 455 (29.6) |
| Bachelor’s degree | 23,421 (30.6) | 23,012 (30.6) | 409 (26.6) |
| Associate’s degree | 5,943 (7.8) | 5,794 (7.7) | 149 (9.7) |
| Some college, no degree | 7,698 (10.0) | 7,496 (10.0) | 202 (13.1) |
| HS degree | 6,636 (8.7) | 6,379 (8.5) | 257 (16.7) |
| Some HS, no degree | 1,232 (1.6) | 1,191 (1.6) | 41 (2.7) |
| Eighth grade | 1,038 (1.4) | 1,014 (1.4) | 24 (1.6) |
| Race/ethnicity | |||
| White | 47,909 (62.1) | 46,847 (62.0) | 1,042 (67.3) |
| Asian | 11,343 (14.7) | 11,247 (14.9) | 96 (6.2) |
| Hispanic | 7,474 (9.7) | 7,323 (9.7) | 151 (9.7) |
| Black | 4,214 (5.5) | 4,109 (5.4) | 105 (6.8) |
| Other race/ethnicity | 6,253 (8.1) | 6,098 (8.1) | 155 (10.0) |
| Gender identity | |||
| Female | 50,843 (65.9) | 50,354 (66.6) | 489 (31.6) |
| Male | 26,350 (34.1) | 25,290 (33.4) | 1,060 (68.4) |
| Sexual orientation (n = 76,880) | |||
| Heterosexual | 68,309 (88.9) | 66,852 (88.7) | 1,457 (94.5) |
| Other sexual orientation | 8,571 (11.1) | 8,486 (11.3) | 85 (5.5) |
| BMI, mean (SD) (n = 73,662) | 24.395 (5.169) | 24.353 (5.163) | 26.510 (5.025) |
| Military involvement | |||
| Military involved | 1,549 (2.0) | – | – |
| No, never served | 75,644 (98.0) | – | – |
| Eating disorder symptoms | |||
| SCOFF positive (≥2) | 15,671 (20.3) | 15,448 (20.4) | 223 (14.4) |
| Eating disorder diagnosis | |||
| Diagnosis among those with ED symptoms | 2,559 (16.8) | 2,535 (16.4) | 24 (10.8) |
SD, Standard Deviation; HS, High School; BMI, Body Mass Index; ED, Eating Disorder.
Of the 15,671 students who scored ≥2 on the SCOFF, 223 of them were military-involved and 15,448 were civilians. The prevalence of having eating disorder symptoms in the military was 14.4%, and the prevalence of having eating disorder symptoms among civilians was 20.4%; this difference was not statistically significant (P = .235). Among males, the prevalence of having eating disorder symptoms between those in the military (11.2%) and civilians (11.9%) was not significantly different (P = .444). Among females, the prevalence of eating disorder symptoms among those in the military (21.3%) was significantly lower than the prevalence among civilians (24.7%, P = .041).
In the unadjusted logistic regression model, military-involved students were less likely to have eating disorder symptoms (OR = 0.781, 95% CI = 0.519, 1.175) when compared to civilians, although this relationship was not statistically significant. In the adjusted model, which controlled for age, gender, and race/ethnicity, those who were in the military were more likely to have eating disorder symptoms compared to civilian students (OR = 1.274, 95% CI = 0.847, 1.918); these results were not statistically significant. See Table II for the unadjusted and adjusted logistic regression results for the overall sample and by gender at birth.
TABLE II.
Eating Disorder Symptoms Among Military-involved and Civilian Students
| Full sample (n = 77,183) | Males (n = 26,350) | Females (n = 50,843) | ||||
|---|---|---|---|---|---|---|
| Unadjusted model | Covariate adjusted model | Unadjusted model | Covariate adjusted model | Unadjusted model | Covariate adjusted model | |
| Military involvement | ||||||
| Civilian (ref) | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
| Military-involved | 0.781 (0.519–1.175) | 1.274 (0.847–1.918) | 1.251 (0.705–2.219) | 1.241 (0.789–1.951) | 0.753 (0.573–0.990)a | 0.943 (0.715–1.245) |
| Age | – | 0.982 (0.973–0.990)a | – | 1.006 (0.990–1.022) | – | 0.968 (0.963–0.974)a |
| Gender | ||||||
| Female (ref) | – | 1.0 | – | – | – | – |
| Male | – | 0.407 (0.383–0.432)a | – | – | – | – |
| Race/ethnicity | ||||||
| White (ref) | – | 1.0 | – | 1.0 | – | 1.0 |
| Asian | – | 1.443 (1.348–1.544)a | – | 1.670 (1.459–1.912)a | – | 1.337 (1.240–1.441)a |
| Hispanic | – | 1.226 (1.125–1.336)a | – | 1.439 (1.213–1.707)a | – | 1.142 (1.036–1.259)a |
| Black | – | 0.750 (0.654–0.860)a | – | 0.730 (0.551–0.968)a | – | 0.768 (0.659–0.895)a |
| Other | – | 1.182 (1.080–1.293)a | – | 1.277 (1.066–1.531)a | – | 1.137 (1.030–1.255)a |
Table values are odds ratios with 95% confidence intervals in parentheses. Adjusted logistic regression controls for age, gender, and race/ethnicity.
Statistically significant confidence interval (P <.05). Ref = reference category.
Among those with eating disorder symptoms, the likelihood of eating disorder diagnosis was higher among civilian students than military-involved students. Of the 15,448 civilian students who had eating disorder symptoms, 16.4% had an eating disorder diagnosis. Of the 223 military-involved students who had eating disorder symptoms, only 10.8% had an eating disorder diagnosis. The chi-square analysis results indicated that the difference between military involvement and eating disorder diagnosis among those with eating disorder symptoms is statistically significant (P = .032). No significant differences were seen in sex-stratified results, but revealed the prevalence diagnosis was substantially lower among males (civilian = 4.3%; military = 3.4%) when compared to females (civilian = 19.4%; military = 19.2%). See Figure 1 for results.
FIGURE 1.
Univariate/bivariate analysis of eating disorder diagnosis among military-involved and civilian students with eating disorder symptoms. *P < .05.
In the unadjusted logistic model examining the odds of having eating disorder diagnosis among those with eating disorder symptoms according to military status, military-involved students with eating disorder symptoms were significantly less likely to have been diagnosed when compared to civilian students with eating disorder symptoms (OR = 0.491, 95% CI = 0.253, 0.952). This relationship was attenuated and no longer statistically significant in the logistic regression model controlling for age, gender, and race/ethnicity (OR = 0.772, 95% CI = 0.451, 1.437). Males with eating disorder symptoms were significantly less likely to be diagnosed with an eating disorder compared to females with eating disorder symptoms in the overall sample (OR = 0.216, 95% CI = 0.156-0.299). Racial/ethnic minorities with eating disorder symptoms were also significantly less likely to be diagnosed with an eating disorder compared to white students with eating disorder symptoms in the overall sample (Asian OR = 0.330, 95% CI = 0.265-0.409; Hispanic OR = 0.542, 95% CI = 0.445-0.661; Black OR = 0.184, 95% CI = 0.120-0.281; other race/ethnicity OR = 0.714, 95% CI = 0.577-0.883). The differences in likelihood of diagnosis among males and racial/ethnic minorities appear to explain why the overall model was no longer statistically significant after covariate adjustment. See Table III for the unadjusted and adjusted logistic regression results for the overall sample, for males, and for females.
TABLE III.
Eating Disorder Diagnosis Among Military-involved and Civilian Students With Eating Disorder Symptoms
| Full sample (n = 15,671) |
Males (n = 3,127) |
Females (n = 12,544) |
||||
|---|---|---|---|---|---|---|
| Unadjusted model | Covariate adjusted model | Unadjusted model | Covariate adjusted model | Unadjusted model | Covariate adjusted model | |
| Military involvement | ||||||
| Civilian (ref) | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 | 1.0 |
| Military-involved | 0.491 (0.253–0.952)a | 0.772 (0.451–1.437) | 0.383 (0.106–1.393) | 0.230 (0.032–1.635) | 1.244 (0.676–2.287) | 1.163 (0.631–2.142) |
| Age | – | 1.005 (0.992–1.019) | – | 1.023 (0.969–1.081) | – | 1.004 (0.993–1.015) |
| Gender | ||||||
| Female (ref) | – | 1.0 | – | – | – | – |
| Male | – | 0.216 (0.156–0.299)a | – | – | – | – |
| Race/ethnicity | ||||||
| White (ref) | – | 1.0 | – | 1.0 | – | 1.0 |
| Asian | – | 0.330 (0.265–0.409)a | – | 0.428 (0.211–0.870)a | – | 0.318 (0.253–0.399)a |
| Hispanic | – | 0.542 (0.445–0.661)a | – | 0.557 (0.264–1.176 a | – | 0.541 (0.441–0.664)a |
| Black | – | 0.184 (0.120–0.281)a | – | 0.111 (0.015–0.837)a | – | 0.189 (0.123–0.292)a |
| Other | – | 0.714 (0.577–0.883)a | – | 0.609 (0.236–1.573) | – | 0.722 (0.583–0.895)a |
Table values are odds ratios with 95% confidence intervals in parentheses. Adjusted logistic regression controls for age, gender, and race/ethnicity.
Statistically significant confidence interval (P < .05). Ref = reference category.
DISCUSSION
In the large and diverse HMS sample, we found high levels of eating disorder symptoms among both military and civilian undergraduate and graduate students. Although the prevalence of symptoms was somewhat higher among civilian students, the difference was not statistically significant. Among females, however, we found significantly higher prevalence among civilian college students (24.7%) when compared to military-involved students (21.3%). Among those with eating disorder symptoms, we found that likelihood of diagnosis was low in both military and civilian college students. Prevalence of eating disorder diagnosis was significantly lower among military-involved college students (10.8%) compared to civilian college students (16.4%). Differences in sociodemographic characteristics (e.g., gender and race/ethnicity) among military-involved and civilian college students appear to explain this association.
Our findings align with previous studies that find high prevalence (e.g., 15-20%) of eating disorder symptoms in the general population7,8 and studies that identify college students2,32 and military-involved individuals14–16,22,27,30 as at-risk groups. Findings from this study support previous research that has demonstrated the ubiquity of underdiagnosis of eating disorders. Previous studies have found high levels of underdiagnosis, with one study showing that only one-third of individuals with a full-syndrome eating disorder will get treatment.12 An even lower likelihood of perceived need for treatment, diagnosis, and treatment was seen in another study of civilian college students with both full-syndrome and subthreshold eating disorders.13 To our knowledge, this is the first study of its kind to examine the prevalence of eating disorders in military-involved college students and to compare self-reported symptoms versus diagnosis in this population.
Given trade-offs between sensitivity and specify inherent in the use of screening tools, cases identified by the SCOFF questionnaire include a mix of individuals: those with DSM defined eating disorders, those with subthreshold conditions, and those with false positives. The high specificity of the SCOFF shown in a recent meta-analysis (0.93) suggests that occurrence of false positives is relatively low.38 Despite the presence of false positives across both civilian and military students, the differences in likelihood of diagnosis among these groups suggest that underdiagnosis may be more common among military students versus civilian students.
Response rate was a key limitation to this study. The response rate for 2015-2016 was 27%, for 2016-2017 was 23%, and for 2017-2018 was 23%. Although we used sample probability weights to adjust for potential differences between nonresponders and responders to the voluntary web-based HMS survey, it is possible that potential differences between these two groups remain (e.g., students with eating disorder symptoms or diagnosis may have been more or less likely to participate). Although the validated eating disorder screening tool SCOFF is widely used to assess eating disorder symptoms, concerns with sensitivity and specificity, depending on the sample, of the measure have been raised.34,35 However, the psychometric properties of the SCOFF are unlikely to differ according to military involvement, and, therefore, it was an appropriate tool to measure eating disorder symptoms for this study.
Another limitation to this study is that it only included college-aged participants, and therefore, there were few active duty military members. The minority number of active duty military personnel was not powered enough to examine the potential differences between different levels of military involvement (i.e., active duty, National Guard/reserves, ROTC, and veterans). We did not know the branch of military service each member was associated with, so we were unable to examine potential differences. Although this study captures a military-involved population, there are concerns of its representativeness to active duty military populations. Because this military demographic captured in this study are individuals who returned to college, there may be differences among the military personnel who choose to return to college and those who spend their careers on active duty. Despite this limitation, we believe this study makes an important contribution to the literature examining disparities in eating disorder treatment seeking and raises important questions that should be explored in other military samples.
Another main limitation to this study is that it relies on self-report data for both eating disorder symptoms and diagnosis. Given our hypothesis that the military represents a demographic that may fear reporting eating disorder symptoms, self-report data may have resulted in false negatives in our assessment of both symptoms and diagnosis. Although capturing accurate estimates of eating disorder symptoms and diagnosis in a military demographic is challenging given the nature of this population, this study was designed to maximize accuracy of responses. Participants are asked to answer the questions as honestly as possible, and the fact that the survey is anonymous may increase the likelihood of honesty.
The final limitation to this study was the phrasing of the question about eating disorder diagnosis. The question asks, “Have you ever been diagnosed with any of the following conditions by a health professional (e.g., primary care doctor, psychiatrist, psychologist, etc.)?” Among those who answered the military involvement and eating disorder symptoms questions, 45,867 participants did not answer the question about eating disorder diagnosis. In addition to several diagnoses listed, the last possible response option for this question was “no, none of these”. It is likely that the high non-response rate of this question may have been due to an oversight of the survey participants based on the question's wording. They may have thought that by not answering the question entirely meant they had not been diagnosed with one of the possible listed disorders thereby missing the negative response option. These individuals were excluded from our analyses, and thus, we predict that underdiagnosis may be more common than we estimate; however, these exclusions would non-differentially apply to military-involved and civilian students.
Despite the limitations, this study provides unprecedented research in an already limited field. It provides key information about the high prevalence of eating disorders and low prevalence of diagnosis among civilian and military populations. This study uses a large, diverse college-aged, sample which coincides for the median age of onset for eating disorders being 18-21 years of age.5 The findings of this study are more generalizable than other population-based studies because it used data from campuses across the United States and Canada. Another strength of this study is that it uses a validated eating disorder screening tool to estimate the prevalence of eating disorder symptoms.34 This study also used sample probability weights to adjust for potential differences between responders and nonresponders to reduce to risk of biases. Finally, the logistic regression used covariate adjustment to adjust for known confounders of eating disorders.
Additional research is needed to ascertain comprehensive prevalence estimates of eating disorders within active duty military populations; these studies should examine both full-syndrome eating disorders and subthreshold eating disorder symptoms. Furthermore, the research questions examined for this study should be replicated in active duty military populations to get a more comprehensive understanding of eating disorder symptoms and diagnosis in the military. An important finding of this study is that underdiagnosis among males and racial/ethnic minorities seen in the general population may be amplified in military populations. As such, future studies should examine eating disorder outcomes among individuals in the military who may feel excluded from the prevailing discourse about eating disorders and marginalized within eating disorder treatment. Other future studies should look at capturing prevalence of disordered eating behaviors around the time of fitness test periods in the military. Research has found the postnatal period to also be a risk period for eating disorders.39 We hypothesize there may be added pressure for active duty, postpartum military women approaching the 1-year deadline to take their fitness test. Therefore, future studies should aim to look at the postnatal period to capture any disordered eating habits among military women. Most importantly, more attention should be paid to getting individuals currently struggling with eating disorder symptoms to give them the resources they need, while also implementing programs and policies within the military designed to decrease eating disorder risk and increase eating disorder diagnosis for those with eating disorder symptoms.
CONCLUSION
Eating disorders pose a significant health risk to those who struggle with them, particularly among those who fail to get diagnosed and treated. Early detection/diagnosis and intervention is beneficial for improving outcomes and likelihood of recovery.40 Unfortunately, underdiagnosis of eating disorders, possibly due to underreporting of symptoms, is far too common, and this represents a threat to the health of military and civilian populations alike.
Supplementary Material
ACKNOWLEDGMENT
None declared.
Contributor Information
1st Lt Sarah E Falvey, Department of Nutritional Sciences, University of Michigan School of Public Health, Ann Arbor, MI 48109-2029, USA.
Samantha L Hahn, Department of Nutritional Sciences, University of Michigan School of Public Health, Ann Arbor, MI 48109-2029, USA; Division of Epidemiology and Community Health, University of Minnesota School of Public Health, Minneapolis, MN 55454, USA; Department of Psychiatry and Behavioral Sciences, University of Minnesota Medical School, Minneapolis, MN 55454, USA.
Olivia S Anderson, Department of Nutritional Sciences, University of Michigan School of Public Health, Ann Arbor, MI 48109-2029, USA.
Sarah K Lipson, Department of Health Law Policy and Management, Boston University School of Public Health, Boston, MA 02118, USA.
Kendrin R Sonneville, Department of Nutritional Sciences, University of Michigan School of Public Health, Ann Arbor, MI 48109-2029, USA.
SUPPLEMENTARY MATERIAL
Supplementary material is available at Military Medicine online.
FUNDING
SLH's time was partially funded by the National Institute of Mental Health (Grant Number; T32MH082761, 1p1i1; S. Crow).
CONFLICT OF INTEREST STATEMENT
None declared.
REFERENCES
- 1.Burnette CB, Simpson CC, Mazzeo SE: Exploring gender differences in the link between weight suppression and eating pathology. Eat Behav 2017; 27: 12–22. [DOI] [PubMed] [Google Scholar]
- 2.Lipson SK, Sonneville KR: Eating disorder symptoms among undergraduate and graduate students at 12 U.S. colleges and universities. Eat Behav 2017; 24: 81–8. [DOI] [PubMed] [Google Scholar]
- 3.Assari S, DeFreitas MR: Ethnic variations in psychosocial and health correlates of eating disorders. Healthcare 2018; 6(2): piiE38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Stice E, Desjardins CD: Interactions between risk factors in the prediction of onset eating disorders: exploratory hypothesis generating analyses. Behav Res Ther 2018; 105: 52–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hudson JI, Hiripi E, Pope HG Jr., Kessler RC: The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biol Psychiatry 2007; 61(3): 348–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.American Psychiatry Association : The diagnostic and statistical manual of mental disorders: DSM-5 (9780890425596 0890425590 9780890425572 0890425574). Washington, DC, American Psychiatric Association Publishing, 2013. Available at https://dsm.psychiatryonline.org/doi/book/10.1176/appi.books.9780890425596; accessed May 3, 2019. [Google Scholar]
- 7.Field AE, Sonneville KR, Micali N, et al. : Prospective association of common eating disorders and adverse outcomes. Pediatrics 2012; 130(2): e289–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Field AE, Sonneville KR, Crosby RD, et al. : High shape concerns predicts becoming obese, binge drinking, and drug use among adolescent and young adult males. JAMA Pediatr 2014; 168(1): 34–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Thein-Nissenbaum J, Rauh M, Carr K, et al. : Association between disordered eating, menstrual dysfunction, and musculoskeletal injury among high school athletes. J Orthop Sports Phys Ther 2011; 41(2): 60–9. [DOI] [PubMed] [Google Scholar]
- 10.Lipson SK, Sonneville KR: Understanding suicide risk and eating disorders in college student populations: results form a National Study. Int J Eat Disord 2020; 53(2): 229–38. [DOI] [PubMed] [Google Scholar]
- 11.Striegel-Moore R, Bulik C: Risk factors for eating disorders. Am Psychol 2007; 62(3): 181–98. [DOI] [PubMed] [Google Scholar]
- 12.Hart LM, Granillo MT, Jorm AF, Paxton SJ: Unmet need for treatment in the eating disorders: a systematic review of eating disorder specific treatment seeking among community cases. Clin Psychol Rev 2011; 31(5): 727–35. [DOI] [PubMed] [Google Scholar]
- 13.Sonneville KR, Lipson SK: Disparities in eating disorder diagnosis and treatment according to weight status, race/ethnicity, socioeconomic background, and sex among college students. Int J Eat Disord 2017; 51(6): 518–26. [DOI] [PubMed] [Google Scholar]
- 14.McNulty PA: Prevalence and contributing factors of eating disorder behaviors in active duty Navy men. Mil Med 1997; 162(11): 753–8. [PubMed] [Google Scholar]
- 15.McNulty PA: Prevalence and contributing factors of eating disorder behaviors in a population of female Navy nurses. Mil Med 1997; 162(10): 703–6. [PubMed] [Google Scholar]
- 16.McNulty PA: Prevalence and contributing factors of eating disorders in active duty service women in the Army, Navy, Air Force, and Marines. Mil Med 2001; 166(1): 53–8. [PubMed] [Google Scholar]
- 17.Forman-Hoffman VL, Mengeling M, Booth BM, Torner J, Sadler AG: Eating disorders, post-traumatic stress, and sexual trauma in women veterans. Mil Med 2012; 177(10): 1161–8. [DOI] [PubMed] [Google Scholar]
- 18.Bodell L, Forney K, Keel P, Gutierrez P, Joiner T: Consequences of making weight: a review of eating disorder symptoms and diagnoses in the United States Military. Clin Psychol (New York) 2014; 21(4): 398–409. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.O’Brien BS, Sher L: Military sexual trauma as a determinant in the development of mental and physical illness in male and female veterans. Int J Adolesc Med Health 2013; 25(3): 269–74. [DOI] [PubMed] [Google Scholar]
- 20.Arditte Hall KA, Bartlett BA, Iverson KM, Mitchell KS: Military-related trauma is associated with eating disorder symptoms in male veterans. Int J Eat Disord 2017; 50(11): 1328–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Arditte Hall KA, Bartlett BA, Iverson KM, Mitchell KS: Eating disorder symptoms in female veterans: the role of childhood, adult, and military trauma exposure. Psychol Trauma 2018; 10(3): 345–51.doi: 10.1037/tra0000301 [DOI] [PubMed] [Google Scholar]
- 22.Garber AK, Boyer CB, Pollack LM, Chang YJ, Shafer MA: Body mass index and disordered eating behaviors are associated with weight dissatisfaction in adolescent and young adult female military recruits. Mil Med 2008; 173(2): 138–45. [DOI] [PubMed] [Google Scholar]
- 23.Acosta JD, Becker A, Cerully JL, et al. : Mental health stigma in the military. RAND Corporation, 2014. Available at https://www.rand.org/content/dam/rand/pubs/research_reports/RR400/RR426/RAND_RR426.pdf; Accessed May 3, 2019. [Google Scholar]
- 24.Coleman SJ, Stevelink SAM, Hatch SL, Denny JA: Stigma-related barriers and facilitators to help seeking for mental health issues in the armed forces: a systematic review and thematic synthesis of qualitative literature. Psychol Med 2017; 47(11): 1880–92. [DOI] [PubMed] [Google Scholar]
- 25.Sharp ML, Fear NT, Rona RJ, et al. : Stigma as a barrier to seeking health care among military personnel with mental health problems. Epidemiol Rev 2015; 37(1): 144–62. [DOI] [PubMed] [Google Scholar]
- 26.Britt TW: The stigma of psychological problems in a work environment: evidence from the screening of service members returning from Bosnia. J Appl Soc Psychol 2006; 30(8): 1599–618. [Google Scholar]
- 27.Jacobson IG, Smith TC, Smith B, et al. : Disordered eating and weight changes after deployment: longitudinal assessment of a large US military cohort. Am J Epidemiol 2009; 169(4): 415–27. [DOI] [PubMed] [Google Scholar]
- 28.Lauder TD, Campbell CS: Abnormal eating behaviors in female reserve officer training corps cadets. Mil Med 2001; 166(3): 264–8. [PubMed] [Google Scholar]
- 29.Litwack SD, Mitchell KS, Sloan DM, Reardon AF, Miller MW: Eating disorder symptoms and comorbid psychopathology among male and female veterans. Gen Hosp Psychiatry 2014; 34(4): 406–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Antczak AJ, Brininger TL: Diagnosed eating disorders in the U.S. Military: a nine year review. Eat Disord 2008; 16(5): 363–77. [DOI] [PubMed] [Google Scholar]
- 31.Lipson SK, Eisenberg D: Mental health and academic attitudes and expectations in university populations: results from the healthy minds study. J Ment Health 2018; 27(3): 205–13. doi: 10.1080/09638237.2017.1417567 [DOI] [PubMed] [Google Scholar]
- 32.Eisenberg D, Nicklett EJ, Roeder K, Kirz NE: Eating disorder symptoms among college students: prevalence, persistence, correlates, and treatment-seeking. J Am Coll Health 2011; 59(8): 700–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Austin S, Nelson L, Birkett M, Calzo J, Everett B: Eating disorder symptoms and obesity at the intersections of gender, ethnicity, and sexual orientation in US high school students. Am J Public Health 2013; 103(2): e16–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Morgan JF, Reid F, Lacey JH: The SCOFF questionnaire. West J Med 2000; 172(3): 164–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hill LS, Reid F, Morgan JF, Lacey JH: SCOFF, the development of an eating disorder screening questionnaire. Int J Eat Disord 2010; 43(4): 344–51. [DOI] [PubMed] [Google Scholar]
- 36.Vohs KD, Heatherton TF, Herrin M: Disordered eating and the transition to college: a prospective study. Int J Eat Disord 2001; 29(3): 280–8. [DOI] [PubMed] [Google Scholar]
- 37.Lavender JM, De Young KP, Anderson DA: Eating Disorder Examination Questionnaire (EDE-Q): norms for undergraduate men. Eat Behav 2010; 11(2): 119–21. [DOI] [PubMed] [Google Scholar]
- 38.Botella J, Sepúlveda A, Huang H, Gambara H: A meta-analysis of the diagnostic accuracy of the SCOFF. Span J Psychol 2013; 16: E92. doi: 10.1017/sjp.2013.92 [DOI] [PubMed] [Google Scholar]
- 39.Bye A, Shawe J, Bick D, Easter A, Kash-Macdonald M, Micali N: Barriers to identifying eating disorders in pregnancy and in the postnatal period: a qualitative approach. BMC Pregnancy Childbirth 2018; 18(1): 114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Treasure J, Stein D, Maguire S: Has the time come for a staging model to map the course of eating disorders from high risk to severe enduring illness? An examination of the evidence. Early Interv Psychiatry 2015; 9(3): 173–84. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.