Abstract
Objective:
To characterize disordered eating behaviors, eating disorder (ED) risk and diagnosis, and treatment seeking behaviors in active-duty military personnel/veterans compared with civilians.
Method:
Self-selecting participants (n=113,388; 1,744 were military personnel/veterans) 18+ years old completed the National Eating Disorders Association’s online screen. Engagement in and frequencies of disordered eating behaviors were compared across military/veteran and civilian groups and were stratified by gender. ED risk and diagnosis and treatment seeking behaviors were also compared.
Results:
Individuals in the military/veteran group were more likely to engage in diuretic/laxative use and excessive exercise compared with civilians. Compared with civilians, the military/veteran group had a lower percentage who screened “at risk for an ED” and a higher percentage who screened for “no risk”. Females in the military/veteran group were more likely to engage in diuretic/laxative use, excessive exercise, and fasting compared with female civilians; males in the military/veteran group were more likely to engage in excessive exercise and less likely to engage in vomiting than male civilians. Of the self-identified military personnel/veterans who screened positive for any ED, 86% had never received treatment, which did not differ significantly from civilians. Notably, 56.7% of those (54.1% of military/veteran group; 56.7% of civilians) who completed an optional item on intention to seek treatment (n=5,312) indicated they would not seek treatment.
Conclusions:
Disordered eating and ED profiles, but not treatment seeking, may differ between military personnel/veterans and civilians who complete an online ED screen. Future work should emphasize treatment options and connecting respondents directly to tailored resources.
Keywords: eating disorders, disordered eating behaviors, treatment seeking, military, veterans, online screening
1. Introduction
Eating disorders (EDs) affect 1-5% of the general population, and the prevalence among military and veteran samples is comparable to or higher than civilian samples (Bartlett & Mitchell, 2015; Beekley et al., 2009; Hudson et al., 2007; Smink, van Hoeken, & Hoek, 2012; Tanofsky-Kraff et al., 2013b). Disordered eating behaviors are also common in the military (reported in up to 7% of men and 30% in women; no available estimates in non-binary/other genders) (Warner et al., 2007). Binge eating (Bartlett & Mitchell, 2015) and compensatory behaviors (vomiting, fasting, excessive exercise, and misuse of diuretics/laxatives) (Tanofsky-Kraff et al., 2013b) are more commonly reported among military personnel and veterans than the general U.S. population.
Most studies assessing EDs in the military report higher ED prevalence in women than men (e.g., Bartlett & Mitchell, 2015; McNulty, 2001; Williams, Stahlman, and Taubman, 2018). Military-specific factors that may contribute to increased risk, especially in women, include regimented lifestyles, strict physical fitness and weight requirements (Bartlett & Mitchell, 2015), mandatory calorically dense meals (MREs: Meals, Ready to Eat), and rigorous physical demands (Beekley et al., 2009). Military personnel must undergo routine body measurements and intensive training and fitness exercises, which are associated with elevated risk of disordered eating behaviors and EDs (Beekley et al., 2009; McNulty, 2001). Despite no longer having to meet fitness standards, female veterans report the influence of their military experiences on disordered eating behaviors (e.g., promoting binge-like eating on irregular meal schedules), rapidly meeting weight requirements post-pregnancy, and socialization to poor eating habits (Breland et al., 2017). Additionally, traumatic experiences related to military service (e.g., sexual and combat trauma and PTSD) are associated with higher risk of disordered eating behaviors and EDs (Blais et al., 2017; Breland et al., 2018; Forman-Hoffman et al., 2012; Arditte Hall et al., 2017; Jacobson et al., 2009; Buchholz, King, & Wray, 2018). Considering that over half of military personnel report at least one stigmatizing experience within the military based on weight and/or shape (Schvey et al., 2017), early ED identification and treatment in military personnel and veterans are imperative, both from military performance and general wellness perspectives.
Even in civilian populations, individuals with EDs rarely seek treatment (Gratwick-Sarll, Mond, & Hay, 2013; Mond, Hay, Rodgers, & Owen, 2006). The issues are particularly complex in the military: ED diagnoses may disqualify individuals from entering military service (Armed Forces Health Surveillance Center, 2014) and may result in a medical discharge in active-duty personnel (Tanofsky-Kraff et al., 2013a), both of which could encourage symptom concealment. Additionally, receiving timely care through the U.S. Veteran’s Affairs Healthcare system may be challenging since EDs are not typically assessed or treated (Huston, Iverson, & Mitchell, 2018). Accordingly, treatment seeking for EDs remains understudied in military personnel and veterans.
The anonymity of online ED screening tools presents an opportunity to fill research gaps on military and veteran samples. In 2017, the National Eating Disorders Association (NEDA) launched a free and confidential online screening tool designed to anonymously screen individuals for probable ED diagnosis or risk and immediately connect them to personalized feedback and resources. In 2018, additional demographic questions were added to better characterize respondents including the designation of active-duty military or veteran.
We used data from the NEDA screen to compare military/veteran versus civilian respondents on engagement in (i.e., the percentage of individuals reporting a behavior) and frequencies of (i.e., in those who reported the behavior, the frequency with which they occurred) disordered eating behaviors in the past three months (binge eating, fasting, vomiting, excessive exercise, and diuretic/laxative use); 2) compare engagement in and frequencies of disordered eating behaviors between the military/veteran and civilian groups within genders; 3) compare the percentage of respondents who self-identified as active-duty military personnel/veterans with those who did not (i.e., civilians) on meeting self-report diagnostic criteria for EDs, at risk for EDs, or no risk; and 4) compare the military/veteran and civilian groups across current treatment status and intent to seek treatment. Based on existing literature, we hypothesized that the military group would report greater engagement in and frequencies of some disordered eating behaviors; female military respondents would report more ED behaviors than female civilians, but this would not be observed in males; the prevalence of individuals meeting self-report ED diagnoses would be at least comparable in the military/veteran and civilian groups; and the military/veteran group would be less likely to seek treatment.
2. Materials and Method
2.1. Procedure
Data were collected from February 2018-January 2019 from NEDA’s online ED screen (see Fitzsimmons-Craft et al., 2019; Fitzsimmons-Craft et al., 2020). The screen was publicized on NEDA’s website and through social media, email listservs, NEDA’s helpline, and print media. Individuals self-selected to participate—active recruitment of military personnel and veterans was not undertaken. Respondents completed demographic questions and the Stanford-Washington Eating Disorders screen (SWED; Graham et al., 2019), which evaluated ED symptoms and probable diagnostic status. Upon completion, participants were presented with tailored feedback and resources and optional questions on intention to seek treatment. The study received approval by the Institutional Review Board at the University of North Carolina at Chapel Hill.
2.2. Participants
Individuals of all ages were eligible to complete the screen (total n=226,059). For this study, respondents were excluded from data analysis if they 1) reported citizenship outside the U.S. (n=29,625); 2) were younger than age 18 (the minimum age to enlist in the U.S. military without parental consent) (n=81,271); 3) did not include current height or weight, or reported unlikely current height (< 48 inches or > 84 inches) or weight (< 50 pounds or > 650 pounds), as these were necessary to determine probable diagnosis (n=942); 4) had BMI values <10 kg/m2 or >100 kg/m2 (n=37); 5) reported compensatory behavior frequencies >500 episodes in 12 weeks (n=72; Fitzsimmons-Craft et al., 2019); 6) did not report gender (n=206); or 7) did not report race (n=518). Those with missing data (n=1,775 from criteria 3-7 above) were more likely to be older, Hispanic, have lower income, and identify as LGBTQ+ and as a racial minority (all p’s<.01). The resulting study sample comprised 113,388 individuals.
2.3. Measures
2.3.1. Demographics:
Participants completed questions on gender identification, age, partnership status, race, ethnicity, LGBTQ+ identification, and disability status (e.g., physical, psychological, or other). Gender was categorized as female, male, or a combined non-binary/other/prefer to self-describe (see footnote1). Participants indicated if they identified as a member of a combined active military duty/veteran group assessed in a single response option; all others were assigned to the civilian referent group. Suicidal ideation over the past two weeks was assessed; we report frequencies of individuals who denied any suicidal ideation and those who reported any suicidal thoughts. Those who endorsed any suicidal thoughts were directed to call 911 or the Crisis Call Center in the feedback portion of the screen.
2.3.2. Disordered Eating Behaviors and Probable Eating Disorder Diagnosis:
The SWED is an 18-item measure that assesses disordered eating behavior frequency and probable ED risk and diagnosis. Participants self-reported the number of episodes of disordered eating behaviors over the past three months. Based on SWED responses and BMI, individuals were categorized hierarchically into one of the following ED self-report diagnostic categories: (1) AN; (2) bulimia nervosa (BN); (3) binge-eating disorder (BED); (4) sub-threshold bulimia nervosa (subBN); (5) sub-threshold binge-eating disorder (subBED); (6) purging disorder (PD); (7) unspecified feeding or eating disorder (UFED); (8) at risk for an ED (endorsed high weight/shape concerns); (9) avoidant/restrictive food intake disorder (ARFID); or (10) no risk. All future mentions of screening positive for ED diagnosis indicate a probable diagnosis. The SWED demonstrates good sensitivity (0.68 to 0.90) and specificity (0.79 to 0.99; Graham et al., 2019) for the current DSM-5 defined ED diagnostic categories (American Psychiatric Association, 2013). It has been used in both men and women and in older samples (Fitzsimmons-Craft et al., 2019).
2.3.3. Treatment Seeking:
To assess treatment history, respondents indicated if they were currently in treatment, were not currently in treatment but had been in treatment previously, or had never received treatment for an ED. After completing the screen and receiving feedback, optional items were presented regarding intention to seek treatment; responses were “Definitely not,” “Probably not,” “Probably,” or “Definitely.”
2.4. Data Analysis
Analyses were conducted using Statistical Analysis Software version 9.4 (2014). For all models below, respondents with missing data for an analysis were only excluded for that specific analysis. To correct for multiple testing, all p-values were adjusted using False Discovery Rate (FDR; Benjamini et al, 2001), and corrected values (q-values) <.05 were considered significant.
We calculated the percentage (n) for all demographic items (including suicidal ideation), disordered eating behaviors, ED diagnosis, treatment history, and intention to seek treatment for the military/veteran and civilian groups as well as by gender within these groups. Chi-square tests were used to evaluate demographic differences between the military/veteran and civilian groups.
To evaluate the association between military/veteran versus civilian groups and likelihood of engaging in disordered eating behaviors, we scored each behavior as 0 if there were no reported episodes of the behavior and 1 if there was at least one reported episode of the behavior over the past 3 months. We applied logistic regression models predicting engagement in each behavior from military/veteran or civilian group. Gender, age (with the 75-84 and the 85+ groups combined because of small sample sizes), race, and ethnicity were included as covariates in these models. We then stratified the logistic regression analyses by gender (female, male, or non-binary/other gender). Age was included as a covariate in the stratified analyses (except for the evaluation of vomiting in males due to small cell sizes). Race and ethnicity were not included as covariates in the stratified models because of small cell sizes.
To evaluate associations between military/veteran versus civilian status and frequencies of disordered eating behaviors, we limited our sample to respondents who reported at least one episode of the respective behavior over the past 3 months. We Winsorized the skewed frequency data for each behavior to reduce effects of outliers (Salkind, 2010). Specifically, observations in the top 0.05% of each behavior were replaced with the maximum threshold value. Means (standard deviations) were calculated using these Winsorized variables for the number of times participants engaged in each disordered eating behavior for the military/veteran and civilian groups and by gender over the past three months. Negative binomial regression models were applied to evaluate differences between the military/veteran and civilian groups. Gender, age, race, and ethnicity were entered in all models as covariates. These same analyses were conducted stratified by gender, with age as a covariate.
We then regrouped the ED diagnostic categories into 3 risk categories: “any ED” (all individuals meeting self-report diagnostic criteria for AN, BN, BED, subBN, subBED, PD, UFED, and ARFID), “at risk,” and “no risk” given that all diagnoses were based on self-report. Differences between the military/veteran and civilian groups were assessed with a chi-square test. To compare differences in ED treatment history (current and past with no treatment) between the military/veteran and civilian groups for those with any ED, we conducted a logistic regression analysis with gender, age, race, and ethnicity as covariates. Finally, to evaluate differences between the military/veteran and civilian groups on intent to seek treatment, an exploratory chi-square test was performed using data from those who responded to this item.
3. Results
Table 1 presents demographic characteristics for the military/veteran and civilian groups and by gender within groups. Of the 113,388 respondents, 1,744 (1.5%) identified as active military personnel/veterans. Significant differences emerged between military and civilian groups on age, race, ethnicity, income, LBTQ+, partnership status, disability, and suicidal ideation (all qs<.02). The mean BMI for the military/veteran group was 26.7 kg/m2 (SD=6.9) and civilian group was 26.1 kg/m2 (SD=7.8).
Table 1.
The percentage (number) of participants endorsing each category of the demographic characteristics for the military/veteran and civilian groups and stratified by gender within group.
| Military/Veteran n=1,744 |
Civilian n=111,644 |
||||||||
|---|---|---|---|---|---|---|---|---|---|
| Total %(n) |
Gender | Total %(n) |
Gender | ||||||
| Female %(n) n=1,267 |
Male %(n) n=438 |
Non-binary / Other %(n) n=39 |
Female %(n) n=101,059 |
Male %(n) n=6,740 |
Non-binary / Other %(n) n=3,845 |
||||
| Age (years) | 18-24 | 46.6 (812) | 48.5 (615) | 40.9 (179) | 46.2 (18) | 61.0 (68,082) | 61.0 (61,653) | 52.9 (3,564) | 74.5 (2,865) |
| 25-34 | 22.6 (395) | 24.6 (311) | 17.1 (75) | 23.1 (9) | 20.7 (23,068) | 20.5 (20,717) | 24.3 (1,637) | 18.6 (714) | |
| 35-44 | 13.4 (233) | 13.6 (172) | 12.8 (56) | 12.8 (5) | 8.4 (9,325) | 8.4 (8,483) | 10.3 (696) | 3.8 (146) | |
| 45-54 | 9.1 (159) | 8.7 (110) | 11.0 (48) | 2.6 (1) | 5.3 (5,922) | 5.4 (5,450) | 6.3 (424) | 1.2 (48) | |
| 55-64 | 5.3 (92) | 3.6 (46) | 10.0 (44) | 5.1 (2) | 3.5 (3,895) | 3.5 (3,556) | 4.4 (300) | 1.0 (39) | |
| 65-74 | 2.3 (40) | 0.8 (10) | 6.4 (28) | 5.1 (2) | 1.1 (1,195) | 1.0 (1,066) | 1.6 (108) | 0.6 (21) | |
| 75-84 | 0.6 (11) | 0.2 (3) | 1.8 (8) | 0.0 (0) | 0.1 (137) | 0.1 (124) | 0.2 (11) | <0.1 (2) | |
| 85+ | 0.1 (2) | 0.0 (0) | 0.0 (0) | 5.1 (2) | <0.1 (20) | <0.1 (10) | 0.0 (0) | 0.3 (10) | |
| Partnership status | Married / domestic partnership | 37.9 (660) | 38.4 (486) | 37.8 (165) | 23.7 (9) | 20.7 (23,049) | 21.1 (21,279) | 21.5 (1,445) | 8.5 (325) |
| Living with partner | 7.0 (122) | 7.3 (92) | 6.2 (27) | 7.9 (3) | 10.3 (11,442) | 10.3 (10,370) | 8.0 (537) | 14.0 (535) | |
| Partnered living separately | 13.6 (237) | 14.5 (184) | 10.5 (46) | 18.4 (7) | 15.7 (17,490) | 15.7 (15,877) | 12.1 (814) | 20.9 (799) | |
| Single | 34.0 (591) | 32.5 (411) | 39.6 (173) | 18.4 (7) | 50.1 (55,878) | 49.6 (50,037) | 55.9 (3,767) | 54.2 (2,074) | |
| Divorced / separated | 6.7 (117) | 6.8 (86) | 5.0 (22) | 23.7 (9) | 2.8 (3,079) | 2.8 (2,867) | 2.2 (149) | 1.6 (63) | |
| Widowed | 0.8 (14) | 0.6 (7) | 0.9 (4) | 7.9 (3) | 0.5 (546) | 0.5 (490) | 0.3 (23) | 0.9 (33) | |
| Race | American Indian / Alaska Native | 1.6 (27) | 1.7 (22) | 0.9 (4) | 2.6 (1) | 0.8 (875) | 0.7 (748) | 0.9 (63) | 1.7 (64) |
| Asian | 3.1 (54) | 3.1 (39) | 3.4 (15) | 0.0 (0) | 3.6 (4,076) | 3.6 (3,619) | 4.9 (330) | 3.3 (127) | |
| Black / African American | 4.3 (75) | 4.4 (56) | 4.3 (19) | 0.0 (0) | 3.0 (3,295) | 2.9 (2,948) | 3.4 (228) | 3.1 (119) | |
| Native Hawaiian / Pacific Islander | 0.6 (10) | 0.5 (6) | 0.7 (3) | 2.6 (1) | 0.2 (241) | 0.2 (203) | 0.2 (17) | 0.6 (21) | |
| White | 79.1 (1,379) | 78.8 (999) | 80.8 (354) | 66.7 (26) | 83.3 (93,025) | 83.9 (84,813) | 80.5 (5,426) | 72.5 (2,786) | |
| More than one race | 6.7 (117) | 7.1 (90) | 5.0 (22) | 12.8 (5) | 4.8 (5,351) | 4.6 (4,634) | 4.8 (326) | 10.2 (391) | |
| Other | 4.7 (82) | 4.3 (55) | 4.8 (21) | 15.4 (6) | 4.3 (4,781) | 4.0 (4,094) | 5.2 (350) | 8.8 (337) | |
| Hispanic | 12.0 (209) | 12.4 (157) | 11.0 (48) | 10.3 (4) | 10.0 (11,181) | 9.8 (9,917) | 11.9 (803) | 12.0 (461) | |
| Income | Less than $20,000 | 17.6 (303) | 16.7 (210) | 18.5 (80) | 33.3 (13) | 20.7 (22,513) | 20.1 (19,831) | 19.5 (1,292) | 37.7 (1,390) |
| $20,000-$39,999 | 19.1 (329) | 19.9 (250) | 16.4 (71) | 20.5 (8) | 16.6 (18,072) | 16.4 (16,195) | 16.6 (1,102) | 21.0 (775) | |
| $40,000-$59,999 | 18.0 (310) | 19.5 (244) | 13.4 (58) | 20.5 (8) | 14.7 (15,969) | 14.7 (14,530) | 14.6 (965) | 12.8 (474) | |
| $60,000-$79,999 | 13.4 (231) | 14.0 (175) | 12.7 (55) | 2.6 (1) | 12.2 (13,323) | 12.4 (12,207) | 11.6 (766) | 9.5 (350) | |
| $80,000-$99,999 | 10.0 (173) | 9.6 (121) | 11.6 (50) | 5.1 (2) | 9.6 (10,520) | 9.8 (9,709) | 8.9 (589) | 6.0 (222) | |
| $100,000-$149,999 | 10.8 (187) | 10.7 (134) | 12.1 (52) | 2.6 (1) | 13.7 (14,943) | 13.9 (13,688) | 14.7 (971) | 7.7 (284) | |
| $150,000+ | 11.1 (192) | 9.6 (120) | 15.3 (66) | 15.4 (6) | 12.5 (13,614) | 12.7 (12,486) | 14.1 (932) | 5.3 (196) | |
| LGBTQ+ | 11.9 (207) | 11.7 (148) | 6.2 (27) | 82.0 (32) | 19.0 (21,181) | 15.8 (15,976) | 29.4 (1,984) | 83.8 (3,221) | |
| Disability | 6.0 (105) | 5.8 (73) | 5.5 (24) | 20.5 (8) | 4.3 (4,826) | 3.9 (3,933) | 4.3 (291) | 15.7 (602) | |
| Suicidal Ideation | 33.2 (579) | 35.9 (455) | 23.3 (102) | 56.4 (22) | 43.5 (48,569) | 47.2 (43,218) | 33.6 (2,537) | 73.2 (2,814) | |
3.1. Disordered Eating Behaviors
Table 2 lists the percentage (n) of individuals engaging in each disordered eating behavior (indicating at least one episode in the past three months) for the military/veteran and civilian groups and stratified by gender within groups. Results from the logistic regression models evaluating group differences in the presence or absence of the disordered eating behaviors are also presented. For the total sample, the military/veteran group had a 33% greater odds of diuretic/laxative use and 46% greater odds of engaging in excessive exercise than civilians. There were no significant differences in engagement in binge eating, vomiting, or fasting between the military/veteran and civilian groups. Stratifying by gender revealed that females in the military/veteran group were more likely to engage in diuretic/laxative use, excessive exercise, and fasting than civilians. No other significant differences were found for females. Males had a different pattern of results: although the military/veteran group was more likely to engage in excessive exercise than civilians, they were less likely to engage in vomiting. For males, no differences were found for binge eating, diuretic/laxative use, and fasting. No significant differences between groups were found for non-binary/other gender, but the wide confidence intervals indicate a lack of power.
Table 2.
Percentage (number) of participants reporting at least one episode of the disordered eating behavior over the previous 3 months for the military/veteran and civilian groups and stratified by gender within group. Odds ratios (OR) and 95% confidence intervals (CI) from logistic regression models evaluating group differences. Significance after correcting for multiple testing using False Discovery Rate (q-values) is indicated.
| Disordered Eating Behavior | Military/Veteran % (n) |
Civilian % (n) |
OR (95% CI) | q-value |
|---|---|---|---|---|
| Total Sample | ||||
| Binge eating | 72.1 (1,256) | 75.6 (80,911) | 1.03 (0.93, 1.15) | .695 |
| Vomiting | 22.2 (387) | 24.6 (27,435) | 1.06 (0.94, 1.19) | .466 |
| Diuretic / laxative use | 22.5 (391) | 19.8 (22,090) | 1.33 (1.19, 1.50) | <.0001 |
| Excessive exercise | 45.0 (783) | 40.2 (44,788) | 1.46 (1.32, 1.61) | <.0001 |
| Fasting | 41.6 (722) | 43.2 (48,021) | 1.12 (1.01, 1.24) | .083 |
| Female | ||||
| Binge eating | 73.9 (936) | 72.8 (73,461) | 1.08 (0.96, 1.23) | .385 |
| Vomiting | 25.8 (326) | 25.0 (25,274) | 1.12 (0.98, 1.27) | .215 |
| Diuretic / laxative use | 26.5 (336) | 20.4 (20,568) | 1.42 (1.26, 1.62) | <.0001 |
| Excessive exercise | 47.4 (600) | 40.6 (40,993) | 1.46 (1.30, 1.64) | <.0001 |
| Fasting | 44.9 (567) | 43.0 (43,293) | 1.21 (1.08, 1.36) | .006 |
| Male | ||||
| Binge eating | 67.5 (295) | 69.0 (4,646) | 0.91 (0.74, 1.13) | .520 |
| Vomiting | 10.8 (47) | 15.9 (1,069) | 0.64 (0.47, 0.87) | .022 |
| Diuretic / laxative use | 10.1 (44) | 11.4 (769) | 0.89 (0.65, 1.23) | .624 |
| Excessive exercise | 37.6 (164) | 34.9 (2,347) | 1.31 (1.06, 1.61) | .047 |
| Fasting | 31.1 (135) | 37.0 (2,488) | 0.90 (0.72, 1.11) | .443 |
| Non-binary/Other Gender | ||||
| Binge eating | 65.8 (25) | 73.0 (2,804) | 0.71 (0.36, 1.41) | .443 |
| Vomiting | 36.8 (14) | 28.4 (1,092) | 1.82 (0.92, 3.63) | .208 |
| Diuretic / laxative use | 29.0 (11) | 19.6 (753) | 1.69 (0.83, 3.45) | .290 |
| Excessive exercise | 50.0 (19) | 37.7 (1,448) | 1.95 (1.01, 3.76) | .138 |
| Fasting | 52.6 (20) | 58.4 (2,240) | 1.08 (0.55, 2.13) | .858 |
Table 3 presents the means (standard deviations) for each disordered eating behavior frequency for those who reported at least one episode of the respective behavior in the past 3 months for the military/veteran and civilian groups and stratified by gender within group. Incidence rate ratios (95% CI) from the negative binomial regression models are listed. The military/veteran group reported 1.2 times the frequency of excessive exercise compared with the civilian group, but no other differences between these groups were observed. The gender stratified analyses were similar: both females and males in the military/veteran group reported significantly higher frequencies of excessive exercise than civilians. No other significant differences were observed.
Table 3.
Means (standard deviations [SD]) for those who reported at least one episode of a disordered eating behavior in the past 3 months for the military/veteran and civilian groups and stratified by gender within group. Results (incidence rate ratios [IRR] and 95% confidence intervals [CI] from negative binomial regression analyses. Significance after correcting for multiple testing using False Discovery Rate (q-values) is indicated.
| Disordered Eating Behavior | Military/Veteran Mean (SD) |
Civilian Mean (SD) |
IRR (95% CI) | q-value |
|---|---|---|---|---|
| Total Sample | ||||
| Binge eating | 16.9 (24.7) | 16.0 (23.9) | 1.00 (0.94, 1.06) | .941 |
| Vomiting | 18.6 (36.7) | 16.1 (31.1) | 1.10 (0.97, 1.25) | .290 |
| Diuretic / laxative use | 12.6 (21.1) | 12.7 (22.1) | 0.94 (0.84, 1.06) | .458 |
| Excessive exercise | 18.9 (26.5) | 15.2 (22.7) | 1.19 (1.10, 1.29) | <.0001 |
| Fasting | 11.9 (20.4) | 11.1 (17.0) | 1.04 (0.96, 1.13) | .458 |
| Female | ||||
| Binge eating | 16.5 (25.0) | 16.0 (23.8) | 0.99 (0.92, 1.06) | .858 |
| Vomiting | 18.9 (37.5) | 16.2 (31.5) | 1.12 (0.97, 1.29) | .277 |
| Diuretic / laxative use | 11.8 (19.5) | 12.7 (22.1) | 0.89 (0.78, 1.01) | .207 |
| Excessive exercise | 17.8 (25.8) | 15.2 (22.7) | 1.14 (1.04, 1.26) | .022 |
| Fasting | 12.2 (20.8) | 11.1 (17.0) | 1.06 (0.96, 1.16) | .402 |
| Male | ||||
| Binge eating | 18.2 (22.9) | 16.3 (24.4) | 1.02 (0.90, 1.16) | .824 |
| Vomiting | 12.5 (20.4) | 14.2 (27.3) | 0.75 (0.51, 1.11) | .288 |
| Diuretic / laxative use | 19.4 (29.4) | 12.7 (20.9) | 1.25 (0.86, 1.83) | .402 |
| Excessive exercise | 22.3 (28.5) | 16.8 (24.1) | 1.31 (1.09, 1.59) | .022 |
| Fasting | 10.2 (17.8) | 10.7 (16.9) | 0.98 (0.80, 1.21) | .890 |
| Non-binary/Other Gender | ||||
| Binge eating | 18.4 (30.1) | 15.6 (24.9) | 1.25 (0.73, 1.74) | .695 |
| Vomiting | 30.5 (55.1) | 14.8 (26.5) | 2.31 (1.14, 4.71) | .069 |
| Diuretic / laxative use | 11.4 (26.2) | 13.9 (23.8) | 0.35 (0.16, 0.81) | .051 |
| Excessive exercise | 25.2 (27.6) | 14.1 (22.4) | 1.34 (0.77, 2.33) | .443 |
| Fasting | 16.4 (24.4) | 12.2 (17.6) | 1.12 (0.68, 1.82) | .749 |
3.2. Self-report Diagnostic and Risk Status
Table 4 provides the percent (n) for the military/veteran and civilian groups across all self-report ED diagnoses. The percentage of individuals screening positive for any ED was 85.4% for both the military/veteran and civilian groups. However, the chi-square test was significant (χ2[2, 113,388] = 21.80, q<.001) as a lower percentage of the military/veteran group than civilians screened as at risk for an ED (9.3% vs 11.2%).
Table 4.
Percent (n) of the military/veteran and civilian groups in each diagnostic and risk category.
| Probable Diagnosis / Risk Category |
Military/Veteran % (n) |
Civilian % (n) |
|---|---|---|
| Any ED | 85.4 (1,490) | 85.4 (95,315) |
| AN | 2.5 (43) | 4.8 (5,396) |
| BN | 11.0 (191) | 9.8 (10,959) |
| BED | 5.6 (98) | 5.7 (6,393) |
| SubBN | 22.9 (400) | 23.9 (26,711) |
| SubBED | 4.0 (70) | 5.3 (5,878) |
| PD | 1.7 (30) | 1.7 (1,856) |
| UFED | 32.9 (573) | 29.8 (33,284) |
| ARFID | 4.9 (85) | 4.3 (4,838) |
| At risk | 9.3 (162) | 11.2 (12,472) |
| No risk | 5.3 (92) | 3.4 (3,857) |
AN = anorexia nervosa; BN = bulimia nervosa; BED = binge-eating disorder; subBN = sub-threshold bulimia nervosa; subBED = sub-threshold binge-eating disorder; PD = purging disorder; UFED = unspecified feeding or eating disorder; ARFID = avoidant/restrictive food intake disorder. Diagnostic and risk categories are listed hierarchically, such that respondents could only receive one diagnosis.
3.3. Treatment Seeking
Regarding treatment history for those with any ED, 88.7% (n=1,321) of the military/veteran group and 86.0% (n=81,927) of the civilian group reported never having received ED treatment; only 2.4% (n=36) of military personnel/veterans and 2.8% of civilians (n=2,699) were currently in ED treatment and the remainder indicated prior treatment. The groups did not differ significantly in regard to receiving any treatment (OR = 0.87, 95% CI = [0.74, 1.02], q=.208).
Of those who screened positive for any ED, 5.5% (5,312) completed the optional question on intention to seek treatment, of whom 109 were military personnel/veterans. Over half (56.7%) indicated they would “probably not” or “definitely not” seek treatment, with no significant difference (q=.695) between the military/veteran (54.1%; n=59) and civilian groups (56.7%; n=2,948).
4. Discussion
Results of this community-based study of military/veteran and civilian groups (completed by over 100,000 respondents, 1,744 of whom self-identified as active-duty military personnel or veterans) yielded mixed support for our hypotheses. Importantly, individuals were likely drawn to the NEDA screen based on personal interest in EDs given that over 85% of both the military and civilian groups screened self-report positive for any ED. Our conclusions must therefore be considered within the context of the self-selecting sample.
4.1. Disordered Eating Behaviors
Our hypotheses that, overall, individuals in the military would show increased engagement in some disordered behaviors than civilians and that females in the military might show greater frequencies of engagement than civilian women were partially supported. After accounting for gender, age, and race, those in the military/veteran group were significantly more likely to engage in and report higher frequencies of excessive exercise episodes than those in the civilian group. This finding should be interpreted within the confines of physical fitness and weight standards required for service by the U.S. military (Bartlett & Mitchell, 2015). Excessive exercise was captured by the question, “pushed yourself very hard; had to stick to a specific exercise schedule no matter what – for example, even when you were sick/injured or if it meant missing a class or other important obligation; felt compelled to exercise.” It is possible that those in the military/veteran group were more inclined to endorse the item either due to the fitness demands of the military or social desirability. Using this definition of excessive exercise presents a similar problem for active military personnel as it does for athletes, given that they often miss important events due to training and compete despite illness or injury; future work necessitates emphasizing the rigid, compulsive, or addictive properties of excessive exercise beyond rigorous training regimens (Adkins & Keel, 2005; Flatt et al., 2020; Scharmer et al., 2020).
The higher likelihood of using diuretics/laxatives and fasting in females in the military/veteran group compared with civilians is consistent with prior research suggesting more severe disordered eating pathology in military personnel than civilians (Tanofsky-Kraff et al., 2013b). Active military personnel must undergo routine body measurements to ensure readiness, which may encourage risky weight loss methods to meet standards (McNulty, 2001). Several researchers and clinicians have raised concerns about unhealthy military culture endorsing “situational disordered eating” to meet fitness standards (e.g., McNulty, 2001; Tanofsky-Kraff et al., 2013a; Warner et al., 2007). It is counterproductive if the tests that evaluate military readiness themselves lead to unhealthy weight loss behaviors that can impair readiness.
The extent to which the observed significant differences between military and non-military respondents are clinically meaningful warrants further consideration. However, the overall engagement in and frequencies of disordered eating behaviors in the military/veteran group were higher than civilians, suggesting an increased severity of overall ED psychopathology in a community military/veteran sample. Future efforts to replicate and confirm our findings should consider assessing the longitudinal course of disordered eating behaviors, ED pathology, and comorbid conditions along with the associated personal and healthcare costs. Our results provide direction for future comprehensive pre-planned investigations in military personnel, including how differences in disordered eating behavior engagement and frequency may be explained by unique military stressors, risk factors, and triggers. The absence of significant findings for individuals who identified as non-binary or preferred to self-describe may reflect the small samples sizes, particularly in the military/veteran group (n=39). Future research should ensure that this group is adequately represented.
4.2. Eating Disorders and Risk
Our first hypothesis that ED prevalence would be comparable between the military/veteran and civilian groups was supported with over 85% of each group meeting criteria for a self-reported threshold or subthreshold ED diagnosis. Within the any ED group, the overall distribution of individuals screening positive for diagnostic categories for the military/veteran and civilian groups was similar. We observed small variations between the groups (e.g., differences were up to ~3 percentage points). If these finding are generalizable across the military, this could reflect substantial need for medical care and financial support for military personnel and veterans. In addition, a smaller percentage of the military/veteran group screened positive for at risk for an ED, but the absolute differences were small (<2%) and may not be clinically meaningful in this sample considering that the overwhelming majority of respondents screened positive for any ED. Prior research suggests military personnel may be less likely to self-report mental health concerns due to stigma (Clement et al., 2015) and potential impact on military career (e.g., possible disqualification from entering military service and medical discharge; Armed Forces Health Surveillance Center, 2014; Bartlett & Mitchell, 2015), which may disproportionately impact women and non-binary individuals given higher ED prevalence in the general population (Coelho et al., 2019; Smink, van Hoeken, & Hoek, 2012). However, since the survey was anonymous, military personnel/veterans may have been more willing to report symptoms. It should be noted that the issues of self-report of EDs and disordered eating behaviors in veterans may be quite different than for those active in the military, for which future research is necessary. Thus, a confidential online screening tool may be an excellent avenue for safe disclosure for military personnel and veterans concerned with their behaviors and mental health.
4.3. Treatment Seeking
Of those screening positive for any ED, approximately 1 in 50 was currently in treatment, highlighting the importance of connecting individuals directly to treatment and encouraging treatment seeking behaviors. For individuals having served in the military and who are concerned about anonymity due to concerns about medical discharge or are seeking services outside of the Veteran’s Affairs care system, a confidential online screening tool is an easily accessible option that directly pairs individuals with tailored treatment recommendations based on their screening results, including options for online programs and peer support, to help reduce treatment seeking barriers. That is, such a tool can, outside of the military system, not only provide individuals with the knowledge they may need help but can also link them with outside services, including digital options that provide a high level of anonymity and confidentiality. Those in the military/veteran group were as likely to have received treatment, prior or current, as civilians; our hypothesis that the military/veteran group would be less likely to seek treatment was not supported. Factors inhibiting treatment seeking, including barriers to ED treatment, exist within both the civilian and military/veteran groups as over half of individuals who screened positive for any ED indicated that they would either “probably not” or “definitely not” seek treatment in the future. Thus, it is important to simultaneously reduce logistical and perceived barriers to treatment and increase access to providers and education, particularly through online programs and resources given the ease of reach and accessibility. Future research should focus on uptake of recommended treatment options and resources after completion of the screen and how to overcome barriers, particularly for those who did not want to seek treatment.
4.4. Strengths and Limitations
Strengths of the current study include the ability to reach a large number of individuals (both military/veteran and civilian) who may not typically be ascertained through traditional research recruitment methods given NEDA’s recognition as one of the most prominent ED advocacy organizations in the U.S. The military/veteran sample was predominantly female between the ages of 18 and 34, on whom research is sparse. This sample does not reflect active-duty military personnel, who are predominantly male (84% in 2017; Barroso, 2019), and the general veteran population (89% in 2021; Pew Research Center, 2021). Our military/veteran sample did reflect active military personnel in terms of socioeconomic status as reported in 2017 (Barroso, 2019), but a larger proportion of our sample identified as white (82% vs 57%) and not-Hispanic (89% vs 64%) than active-duty servicemembers. An additional limitation was that we could not distinguish between active-duty servicemembers and veterans. Therefore, results cannot be considered to reflect the actual prevalence of EDs or disordered eating in active-duty military, reservists, or veterans. Second, some of the analytic tests may have been overpowered given the large sample, even after correcting for skewed distributions and multiple testing, yielding significant results despite marginal differences. Third, our diagnoses were based on algorithms derived from self-report of symptoms. Verification via clinical interview would be desirable, although limiting, especially for a population unlikely to disclose. Fourth, as indicated above, the exercise question may not have accurately discriminated between the physical demands of the military and excessive exercise as a disordered eating behavior. Finally, participants could respond more than once due to the anonymous nature of the publicly available survey, which could be addressed through response authenticators in future iterations.
5. Conclusions
Using an anonymous self-selecting online ED screening tool, the majority of military/veteran and civilian groups screened positive for ED diagnoses (85%), <3% of whom were currently in treatment. The military/veteran group reported more excessive exercise and diuretic/laxative use than civilians. Females in the military/veteran group also engaged in more disordered eating behaviors than their civilian counterparts. Results were mixed for males. Despite high frequencies of ED diagnoses and behaviors, more than half reported no intention to seek treatment. Although our sample was skewed toward those with current or past pathology, it nonetheless underscores that EDs and disordered eating are concerning in the military and that future research is essential to better understand barriers to treatment-seeking, the efficacy of tailored digital interventions paired directly with anonymous screening, and the importance of education about EDs given the impact on military readiness and overall well-being.
Highlights.
Over 85% of the self-selecting sample screened positive for an eating disorder
Of those who screened positive for an eating disorder, <3% were in treatment
Military/veteran group engaged in more disordered eating behaviors than civilians
Female military/veteran group reported more disordered eating than female civilians
No differences between military/veteran and civilians across treatment status
Acknowledgements:
The authors would especially like to thank the National Eating Disorders Association and their staff for their generosity, support, and contributions to the screening tool. We would also like to thank all respondents and the research team for their time and dedication to this study.
Role of Funding Sources:
RF is supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. (DGE-1650116). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. CB acknowledges funding from the Swedish Research Council (Vetenskapsrådet, award: 538-2013-8864) and from the National Institutes of Health (R01 MH120170 and R01 MH119084). This work was also supported by R01 MH100455, K08 MH120341, and K01 DK120778. None of the listed funding sources were directly involved in study design, collection, analysis, or interpretation of the data, writing the manuscript, or the decision to submit the manuscript for publication.
Footnotes
Declaration of Interest:
CM Bulik reports: Shire (grant recipient, Scientific Advisory Board member); Idorsia (consultant); Pearson (author, royalty recipient); Lundbeckfonden (grant recipient)
We recognize the terminology used to describe gender and sex were often conflated in research. Although the terms “female” and “male” refer to sex, the screening tool described these terms as genders, so we have retained this language to stay true to how the original data were captured.
References
- Adkins EC, & Keel PK (2005). Does “excessive” or “compulsive” best describe exercise as a symptom of bulimia nervosa? International Journal of Eating Disorders, 38(1), 24–29. [DOI] [PubMed] [Google Scholar]
- American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA. [Google Scholar]
- Arditte Hall KA, Bartlett BA, Iverson KM, & Mitchell KS (2017). Military-related trauma is associated with eating disorder symptoms in male veterans. International Journal of Eating Disorders, 50(11), 1328–1331. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Armed Forces Health Surveillance Center (AFHSC) (2014). Diagnoses of eating disorders among active component service members, US Armed Forces, 2004-2013. Medical Surveillance Monthly Report, 21(9), 8. [PubMed] [Google Scholar]
- Barroso A (2019). The changing profile of the U.S. Military: Smaller in size, more diverse, more women in leadership. Retrieved February 17, 2021, from https://www.pewresearch.org/fact-tank/2019/09/10/the-changing-profile-of-the-u-s-military/
- Bartlett BA, & Mitchell KS (2015). Eating disorders in military and veteran men and women: A systematic review. International Journal of Eating Disorders, 48(8), 1057–1069. [DOI] [PubMed] [Google Scholar]
- Beekley MD, Byrne R, Yavorek T, Kidd K, Wolff J, & Johnson M (2009). Incidence, prevalence, and risk of eating disorder behaviors in military academy cadets. Military Medicine, 174(6), 637–641. [DOI] [PubMed] [Google Scholar]
- Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I (2001). Controlling the false discovery rate in behavior genetics research. Behavioral Brain Research, 125(1-2), 279–84. [DOI] [PubMed] [Google Scholar]
- Blais RK, Brignone E, Maguen S, Carter ME, Fargo JD, & Gundlapalli AV (2017). Military sexual trauma is associated with post-deployment eating disorders among Afghanistan and Iraq veterans. International Journal of Eating Disorders, 50(7), 808–816. [DOI] [PubMed] [Google Scholar]
- Breland JY, Donalson R, Li Y, Hebenstreit CL, Goldstein LA, & Maguen S (2018). Military sexual trauma is associated with eating disorders, while combat exposure is not. Psychological Trauma: Theory, Research, Practice, and Policy, 10(3), 276–281. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Breland JY, Donalson R, Nevedal A, Dinh JV, & Maguen S (2017). Military experience can influence women’s eating habits. Appetite, 118, 161–167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Buchholz LJ, King PR, & Wray LO (2018). Rates and correlates of disordered eating among women veterans in primary care. Eating Behaviors, 30, 28–34. [DOI] [PubMed] [Google Scholar]
- Clement S, Schauman O, Graham T, Maggioni F, Evans-Lacko S, Bezborodovs N, Morgan C, Rüsch N, Brown JS, & Thornicroft G (2015). What is the impact of mental health-related stigma on help-seeking? A systematic review of quantitative and qualitative studies. Psychological medicine, 45(1), 11–27. [DOI] [PubMed] [Google Scholar]
- Coelho JS, Suen J, Clark BA, Marshall SK, Geller J, & Lam PY (2019). Eating Disorder Diagnoses and Symptom Presentation in Transgender Youth: a Scoping Review. Current psychiatry reports, 21(11), 107. [DOI] [PubMed] [Google Scholar]
- Fitzsimmons-Craft EE, Balantekin KN, Graham AK, Smolar L, Park D, Mysko C, … Wilfley DE (2019). Results of disseminating an online screen for eating disorders across the U.S.: Reach, respondent characteristics, and unmet treatment need. International Journal of Eating Disorders, 52(6), 721–729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Fitzsimmons-Craft EE, Balantekin KN, Graham AK, DePietro B, Laing O, Firebaugh M, … & Wilfley DE (2020). Preliminary data on help-seeking intentions and behaviors of individuals completing a widely available online screen for eating disorders in the U.S. International Journal of Eating Disorders. Advance online publication. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Flatt RE, Thornton LM, Fitzsimmons-Craft EE, Balantekin KN, Smolar L, Mysko C, Wilfley DE, Taylor CB, DeFreese JD, Bardone-Cone AM, Bulik CM (2020). Comparing Eating Disorder Characteristics and Treatment in Self-Identified Competitive Athletes and Non-Athletes from the National Eating Disorders Association Online Screening Tool. International Journal of Eating Disorders. Advance online publication. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Forman-Hoffman VL, Mengeling M, Booth BM, Torner J, & Sadler AG (2012). Eating disorders, post-traumatic stress, and sexual trauma in women veterans. Military Medicine, 177(10), 1161–1168. [DOI] [PubMed] [Google Scholar]
- Graham AK, Trockel M, Weisman H, Fitzsimmons-Craft EE, Balantekin KN, Wilfley DE, & Taylor CB (2019). A screening tool for detecting eating disorder risk and diagnostic symptoms among college-age women. Journal of American College Health, 67(4), 357–366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gratwick-Sarll K, Mond J, & Hay P (2013). Self-recognition of eating-disordered behavior in college women: further evidence of poor eating disorders “mental health literacy”? Eating Disorders, 21(4), 310–327. [DOI] [PubMed] [Google Scholar]
- Hudson JI, Hiripi E, Pope HG, & Kessler RC (2007). The prevalence and correlates of eating disorders in the national comorbidity survey replication. Biological Psychiatry, 61(3), 348–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Jacobson IG, Smith TC, Smith B, Keel PK, Amoroso PJ, Wells TS, … Team, for the M. C. S. (2009). Disordered eating and weight changes after deployment: longitudinal assessment of a large us military cohort. American Journal of Epidemiology, 169(4), 415–427. [DOI] [PubMed] [Google Scholar]
- McNulty AF (2001). Prevalence and contributing factors of eating disorder behaviors in active duty service women in the Army, Navy, Air Force, and Marines. Military Medicine, 166(1), 53–58. [PubMed] [Google Scholar]
- Mond J, Hay P, Rodgers B, & Owen C (2006). Self-recognition of disordered eating among women with bulimic-type eating disorders: A community-based study. International Journal of Eating Disorders, 39(8), 747–753. [DOI] [PubMed] [Google Scholar]
- Pew Research Center. (2021). Looking forward at the changing profile of U.S. veterans. Retrieved August 11, 2021, from https://www.pewresearch.org/fact-tank/2021/04/05/the-changing-face-of-americas-veteran-population/ft_21-04-06_veterans_stacked_bar/.
- SAS [computer program]. (2014). Version 9.4. Cary, NC: SAS Institute Inc. [Google Scholar]
- Salkind NJ (2010). Encyclopedia of Research Design (Vols. 1-0, pp. 181–186). Thousand Oaks, CA: SAGE Publications, Inc. [Google Scholar]
- Scharmer C, Gorrell S, Schaumberg K, & Anderson D (2020). Compulsive exercise or exercise dependence? Clarifying conceptualizations of exercise in the context of eating disorder pathology. Psychology of Sport and Exercise, 46, 101586. [PMC free article] [PubMed] [Google Scholar]
- Schvey NA, Barmine M, Bates D, Oldham K, Bakalar JL, Spieker E, … Sbrocco T (2017). Weight stigma among active duty U.S. military personnel with overweight and obesity. Stigma and Health, 2(4), 281–291. [Google Scholar]
- Smink FRE, van Hoeken D, & Hoek HW (2012). Epidemiology of eating disorders: incidence, prevalence and mortality rates. Current Psychiatry Reports, 14(4), 406–414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tanofsky-Kraff M, Bulik CM, Marcus MD, Striegel RH, Wilfley DE, Wonderlich SA, & Hudson JI (2013a). Binge eating disorder: the next generation of research. The International journal of eating disorders, 46(3), 193–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Tanofsky-Kraff M, Sbrocco T, Theim KR, Cohen LA, Mackey ER, Stice E, Henderson JL, McCreight SJ, Bryant EJ, & Stephens MB (2013b). Obesity and the US military family. Obesity (Silver Spring, Md.), 21(11), 2205–2220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Warner C, Warner C, Matuszak T, Rachal J, Flynn J, & Grieger TA (2007). Disordered eating in entry-level military personnel. Military medicine, 172(2), 147–151. [DOI] [PubMed] [Google Scholar]
- Williams VF, Stahlman S, & Taubman SB (2018). Diagnoses of eating disorders, active component service members, US armed forces, 2013–2017. MSMR, 25(6), 18–25. [PubMed] [Google Scholar]