Skip to main content
  • Research article
  • Open access
  • Published:

Initial validation of a screening tool for disordered eating in adolescent athletes



Disordered eating (DE) is a growing problem among all athletes, particularly adolescents. To help prevent the progression of DE to a clinical eating disorder (ED), a brief screening tool could offer an efficient method for early identification of DE in athletes and facilitate treatment. The aim of this study is to validate a screening tool for DE that will identify male and female adolescent athletes of all sports and levels of competition who are at risk for DE. The Disordered Eating Screen for Athletes (DESA-6) consists of only 6 items and was designed for use in both male and female athlete populations.


Validation involved two phases: Phase I consisted of screening high school athletes using the Eating Attitudes Test (EAT-26) and the DESA-6; and Phase II included inviting all high school athletes categorized as “at risk” after screening, plus age- and self-reported gender- matched athletes categorized as not “at risk”, to complete the same surveys a second time along with clinical interview. Validity and regression analyses were used to compare the DESA-6 to the EAT-26 and EDE 17.0D.


When comparing to clinical interview, the DESA-6 had a total sensitivity of 92% and specificity of 85.96%, respectively. Upon comparison of concurrent validity, Phase II DESA-6 had a strong significant positive correlation for both males and females when compared to Phase II EDE 17.0D.


A brief, easy to administer screening tool for recognizing DE that can be used by physicians, psychologists, athletic trainers, registered dietitians, and other sport/healthcare staff is of utmost importance for early intervention, which can lead to improved treatment outcomes. The DESA-6 is a promising tool for risk assessment of DE in athletes.

Plain English

Disordered eating, a subclinical spectrum of eating disorders, can potentially progress into numerous physical health complications. Disordered eating may progress to a clinical eating disorder if left untreated. If individuals with disordered eating are identified quickly, treatment outcomes may improve. Athletes are particularly susceptible to disordered eating. This manuscript attempts to introduce the Disordered Eating Screen for Athletes (DESA-6) screening tool, which aims to quickly identify both males and females who may be at risk for disordered eating. Current screening tools are generally only focused on validation in females, but the DESA-6 also aims to be validated in males. Additionally, the DESA-6 is much shorter than previous screening tools and significantly shorter than the current gold standard clinical interview. This study showed the DESA-6 have promised as a tool that athletic trainers, primary care physicians, and sport/healthcare staff may use to quickly screen a student athlete for disordered eating behaviors.


The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) defines eating disorders (EDs) as “the persistent disturbance of eating or eating-related behavior, which results in altered consumption of food and a significant impairment of physical health and/or psychosocial functioning [1].” There are two primary presentations of EDs, anorexia nervosa (AN) and bulimia nervosa (BN), both of which are characterized by an undue influence of body weight and/or shape on self-evaluation [1]. Additional EDs include binge eating disorder (BED) and other specified feeding and eating disorder (OSFED). The prevalence of EDs in the general population is estimated to be less than 1% for AN and 1–2% for BN [2].

Previous researchers have indicated that eating-related issues exist on a spectrum, with the clinically diagnosable DSM-5 EDs at the more extreme end of the spectrum [2]. Within the spectrum of eating-related issues is a condition referred to as “disordered eating” (DE), a term which has been used in research to generally describe the abnormal and potentially harmful eating behaviors that do not meet diagnostic criteria for clinical EDs [3]. Although it is believed that DE lies on the same spectrum of eating-related issues as ED, there is currently no formally accepted definition for DE. Furthermore, previously suggested definitions may not encompass the full spectrum of potential behaviors. An editorial on the continuum of DE defined DE as a variety of abnormal eating behaviors (e.g. restrictive eating, fasting, excessive eating, diet pills, laxatives/enemas, and purging [4]. This article provided a comprehensive list of potential abnormal eating behaviors; however, it did not include potential abnormal time parameter restrictions. In athletes, DE has been defined as a general term that describes the spectrum of abnormal eating behaviors with the goal of achieving or maintaining an unhealthy body weight [3].” However, not all DE behaviors in athletes are intended to achieve a lower weight. Athletes, especially male athletes, may engage in DE behaviors with the intent to increase muscle mass to improve performance [5, 6].

With the previous definitions and associated drawbacks in mind, we propose the following definition for disordered eating: intentional chronic abnormal, unhealthy eating/drinking behaviors that can lead to clinically relevant problems and do not necessarily meet DSM-5 criteria for eating disorders. This can include a number of behaviors including (but not limited to): restrictive eating (including excessive fasting or frequently skipping meals); overeating or binge eating; purging; and/or the use of weight loss supplements (e.g., diet pills, laxatives, diuretics). In an athlete population, DE is often done with the intent to alter body weight, shape, and/or athletic performance. Although eating patterns similar to those with DE can arise from a lack of nutrition knowledge by the athlete, it is important to distinguish those athletes who engage in these behaviors intentionally as having DE. Adolescent athletes in particular are susceptible to unintentional DE behavior, as they likely have little knowledge of optimal food intake and may have little access to sports dietitians or nutritional knowledge.

DE is a growing problem among all athletes, particularly adolescent athletes [3, 4]. Adolescents with a pattern of DE often results in them having low energy availability, electrolyte abnormalities and dehydration, which all negatively impact physical and mental health, as well as athletic performance [7]. DE can lead to injuries that result in more participation days lost and longer recovery periods from injuries [8]. If untreated, DE could progress to a clinical ED, putting these adolescents at an increased risk of ED-related mood, anxiety, impulse-control and substance use disorders [9]. Additionally, EDs have one of the highest mortality rates of all mental disorders [7] and are associated with many potentially life-threatening complications, including cardiac arrhythmias, bone loss, colon hypofunction and kidney failure [10]. In order to potentially prevent the progression of DE to a clinical ED, a brief screening tool could offer an efficient method for identifying DE in athletes, thus facilitating early identification and necessary treatment [11,12,13].

Prevalence estimates of DE/EDs among adolescent athletes vary widely due to the differing criteria and definitions between research studies; however, a number of studies have indicated the prevalence of DE and EDs is higher among adolescent athletes than in the general population [14, 15]. The National Athletic Trainers’ Association (NATA), the American College of Sports Medicine (ACSM) and the International Olympic Committee (IOC) have released statements supporting regular screening for DE in athletes using a screening tool designed specifically for athletes [7, 16, 17]. However, the NATA Position Statement noted that widely used questionnaires, such as the Eating Attitudes Test (EAT-26), the Eating Disorder Inventory (EDI) and the Eating Disorder Examination Questionnaire (EDE-Q), should be used cautiously as these screening tools have not been validated in athletes [7]. The EDE-Q is a self-report questionnaire based on the Eating Disorder Examination (EDE), which is the gold standard for clinically diagnosing EDs and is currently on version 17, referred to as EDE 17.0D. For athletes, specifically, there are six ED-specific screening tools in existence, with the four most common being the Athletic Milieu Direct Questionnaire (AMDQ), Female Athlete Screening Tool (FAST), Physiologic Screening Test (PST), and the Brief Eating Disorder in Athletes Questionnaire (BEDA-Q) [11, 13, 18, 19].

A review by Knapp et al. discusses reasons the four athlete-related screeners should be used with caution [20]. Although the AMDQ was validated utilizing the EDE-16, the tool was validated only in female collegiate athletes, therefore it is unknown if the AMDQ could be generalized to non-collegiate or recreational athletes. Additionally, the AMDQ evaluates presence of menstruation, thus is cannot be used to screen male athletes [13]. The FAST exhibits a similar issue as the wording is specific to female athletes and it was validated in a population of female collegiate athletes ages 18 to 23 years [18]. Neither of these screeners would be appropriate to use in an adolescent athlete population of mixed sexes. The PST was developed specifically for DE screening [11]; however, it requires physiologic measurements (e.g., body fat percentage, parotid gland evaluation), which require specialized equipment and/or training, thus making this screener difficult to use in the field. Additionally, the PST includes a question about the regularity of menstrual cycles, so it cannot be used in male athletes. Lastly, although the BEDA-Q was validated in female adolescent athletes, it is unknown if it would be valid for all levels of competition or male adolescent athletes [19]. The BEDA-Q was developed using DSM-IV criteria, which included amenorrhea as a diagnostic criterion, so it is unclear if this screening tool could be generalized to include male athletes.

A standardized, brief, efficient screening tool for DE that could be given to a large population of both male and female athletes is necessary [10, 19]. The aim of this study is to validate a screening tool for DE that will identify male and female adolescent athletes of all sports and levels of competition who are at risk for DE. The Disordered Eating Screen for Athletes (DESA-6) consists of only 6 items and was designed specifically for use in both male and female athlete populations. We hypothesize that the DESA-6 will be an efficient screening tool to identify DE in adolescent athletes, and may help decrease the potential for progression to clinical ED.


All methods were approved by the university’s Institutional Review Board, where the full protocol can be accessed. After initial generation of the DESA-6 questions, we conducted pilot testing in a population of adult athletes using the EAT-26 as a comparison to the DESA-6. After a final version of the DESA-6 was decided, the validation involved two phases: Phase I, which consisted of screening high school athletes using the EAT-26 and the DESA-6; and Phase II in which we invited all high school athletes categorized as “at risk” after screening, along with age- and self-reported gender- matched athletes categorized as not “at risk”, to complete the surveys a second time with clinical interview in which we used the gold-standard EDE 17.0D for comparison. Participants were recruited from Midwestern United States high schools during the months of January through June of 2019.

Eating attitudes test (EAT-26)

The EAT-26 is a 26-item self-report instrument that utilizes a 6-point Likert scale ranging from 1 (never) to 6 (always). The EAT-26 is an abbreviated version of a 40-item scale developed by Garner et al. and is a validated psychometric measure that identifies risk for EDs in the general population [21]. The criterion validity of the EAT-26 was evaluated and found to have an overall accuracy of 0.90 [22]. Additionally, the EAT-26 has a high degree of internal reliability with an alpha coefficient of 0.79 in patients with AN and an alpha coefficient of 0.94 comparing patients with AN and normal controls [21].

Although the EAT-26 was validated in the general population and not athletes [21], it was chosen as the external validity comparison because it is a well-accepted tool for detecting EDs and is a relatively brief measure for use in the field. A score of 20 or greater on the EAT-26 indicates risk for ED [22, 23]; however, the purpose of the DESA-6 is to detect DE and not ED, so a different scoring system was needed. A study of female athletes found a mean EAT-26 score of 12.66 was associated with DE [24]; therefore, a score of 12 was chosen for this study to indicate someone being “at risk” for DE. was agreed upon with expert opinion from two sports psychiatrists.

Eating disorder examination (EDE 17.0D)

The Eating Disorder Examination is the current gold standard for diagnosing EDs [25]. It is a psychometrically validated semi-structured interview. It has been revised numerous times with the aim of maximizing reliability [26]. The EDE 17.0D is based on the diagnostic criteria outlined in the DSM-5 to evaluate for AN, BN, binge eating disorder and unspecified EDs. It consists of 62 items categorized into four subscales (Restraint, Eating Concern, Weight Concern, Shape Concern) and takes up to an hour to complete. Previous work on validity of the EDE has demonstrated the majority of the individual items are able to discriminate between patients with and without EDs [25, 26]. Additionally, inter-rater reliability of all items has been found to be high [25]. The EDE 17.0D was chosen as the Concurrent validity test because it is the current gold standard for diagnosing EDs [25] and no gold standard for DE exists.

DESA-6 question development

The DESA-6 was designed for this study by a former professional triathlete based on a previous study that surveyed over 1000 triathletes in the United States [27]. A review of literature was conducted on DE/EDs in athletes and a pool of 18 items was created. The initial pool included triathlon-specific questions and “competitive athletes” questions, which applied to all sports. Interviews were then conducted with randomly selected triathletes to assess the items. The questions, along with the EAT-26 for comparison purposes, were distributed randomly to triathletes via an online survey, which was completed by 1033 adult triathletes. The chi square test was used in that previous study to assess the statistical significance of relationships between EAT-26 scores and the “triathlon-specific” and “competitive athlete factors.” Following statistical analysis, the items were narrowed down to six, which all assessed “competitive athlete factors,” [27] meaning no questions were specific to triathletes, thus applicable to all sports.

Each of the final six questions were purposely designed to assess a specific facet of DE unique to athletes (see Supplemental Digital Content 1, DESA-6). As DE/EDs are associated with an increased injury rate, more frequent injuries or more severe injuries may indicate higher risk for DE [8]; therefore, question 1 was intended to assess the frequency and severity of injuries due to the increased risk for injury and increased length of recovery following injury [10]. Question 2 was intended to assess fear of weight gain but was specifically tailored to athletes by asking about periods of time when they cannot train or are not training as intensely, so fear of weight gain may be more pronounced. This reflects the DSM-5 Criteria B of AN, which is an intense fear of gaining weight and/or of becoming fat [1]. Many athletes believe that losing weight results in improved performance [28], and with DE may be unhappy with their weight even if they are normal weight or underweight; therefore, Question 3 evaluates happiness with weight and Question 4 assesses the intensity of dissatisfaction with current weight. These questions are related to the DSM-5 Criteria C of AN, which is a disturbance in the way in which one’s body shape or weight is experienced [1]. Dieting is a risk factor for development of EDs in female elite athletes [16, 29], thus, Question 5 is intended to assess the presence of dieting. Because being told to lose weight by a coach/authority figure was found to be a risk factor for development of EDs [14], Question 6 is intended to evaluate the presence of this pressure. To verify that the wording was appropriate and sufficient for all athletes, the six questions and answer choices were reviewed for face validity and adjusted following interviews with athletes of different sports and experts in the field, including sports psychiatrists, sports psychologists, and registered dietitians. A positive score, which is a score of 3 or greater, on the DESA-6 is intended to indicate risk for DE.

Pilot testing

A total of 22 randomly selected adult athletes aged 18–19 years old were recruited at a local sports medicine clinic. Individuals who were recruited were selected at random from the patient list of two sports medicine providers. Inclusion criteria was self-identification as an athlete participating in a “sport”, defined as an activity requiring a higher level of physical exertion and/or skill than walking. This convenience sample was used for pilot testing because it was an available, responsive population that could prove valuable for future validation of the DESA-6.

Participants completed both the DESA-6 and EAT-26. None were excluded. There were 11 female athletes and 11 male athletes. Twelve different sports were represented in this sample. Following informed consent, athletes were provided a paper survey that included demographic questions (age, a free-text gender question, sport, number of training hours per week), anthropometric questions (self-reported height and weight), the DESA-6, and the EAT-26. Identifying information of the participants was not collected in order to assure anonymity. Athletes completed the survey in a private room with a trained researcher available outside the room.

DESA-6 validation

Phase I: initial screening in adolescent athletes

Power was calculated a priori analysis using the standard alpha of 0.05, power 0.95, and effect size of 0.3, which gave a suggested sample size of 111 for phase I of research. High school student athletes (ages 12 to 19 years, defined as those participating in a high school sponsored sport) were recruited from 12 different Midwest high schools via emails to administrators and athletic trainers. Student athletes were provided with consent forms by school faculty approximately one week prior to survey administration by a member of lab personnel in order to obtain parental consent. Student athletes aged 18 to 19 years were provided adult consent forms. A convenience sample of 308 student athletes provided consent, and none were excluded. A total of 17 sports were represented, including basketball, baseball, football, etc. (see Table 3 for full list of sports). Student athletes were provided a paper survey that included demographic questions (age, a free-text gender question, sport, number of training hours per week), anthropometric questions (self-reported height and weight), the DESA-6, and the EAT-26.

Phase II: clinical interviews

Based on Phase I results, student athletes classified as “at risk” for DE based on EAT-26 scores (score ≥ 12) were invited to attend the clinical interview to determine presence of DE/EDs. A total of 77 student athletes were classified as “at risk” and all were invited to participate in Phase II. 41 of these “at risk” students consented to participate. We then invited age- and self-reported gender- matched athletes classified “not at risk” to serve as a control group. A total of 82 student athletes (n = 41 at risk) completed Phase II and none were excluded. A total 15 sports were represented (gymnastics and rugby were not represented versus Phase I).

Student athletes under the age of 18 years came to the interview with a parent/guardian to provide written consent and assent. Adult student athletes completed a written consent form. Prior to the clinical interview, participants completed the DESA-6 and EAT-26 again. Due to scheduling conflicts, as many athletes were actively competing, time between initial completion of the DESA-6 and the EAT-26 ranged from 3 weeks to 12 weeks. Height was verified with a stadiometer (Seca 216, Germany) and weight was verified with a scale (Detecto 758C, Welch City, MO). The private clinical interview utilized the EDE 17.0D, which is the gold standard for diagnosing EDs (2). Interviewers (n = 5) were trained to conduct the interviews by a board-certified child and adolescent psychiatrist with extensive experience treating DE/EDs and athletes. This board-certified child and adolescent psychiatrist then tabulated the results of the EDE 17.0D and interpreted the results of the self-administered DESA-6. The interviewers verified that each survey was completely answered before moving on with the next phase of the study, which resulted in no missing data on the DESA-6 or EDE 17.0D tests. No observed adverse effects from performing the DESA-6, EAT-26, or ED 17.0D were noted.

Statistical analyses

An EDE 17.0D with a positive score on two out of four subscales was considered positive for DE. This accommodates for the subclinical nature of DE in comparison to ED, which is measured with a score of 3 or more subscales [26]. A DESA-6 score of three or more was considered a positive screening due to its ability to achieve strong sensitivity while maintaining an acceptable specificity. As a score of three was used as a cutoff value, there were no indeterminate scores. Validity and regression analyses were used to compare the DESA-6 to the EAT-26 and EDE 17.0D. All regressions were interpreted via guidelines from “Statistics without maths for Psychology”. Phase I data of the DESA-6 were compared to Phase I EAT-26 scores to show concurrent validity. Phase I data of the DESA-6 were compared to Phase II EDE 17.0D scores to show the predictive validity. Phase II data of the DESA-6 were compared to Phase II EDE 17.0D scores to show construct validity. Phases I and II data of the DESA-6 scores were compared using regression analysis to determine test-retest reliability. Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) analysis was then explored to further assess correlation between Phase II EDE 17.0D and DESA-6 scores [30]. Statistics were performed with Excel Data ToolPak Version 2016, Software, SPSS version 25, and [31].


Pilot testing showed that the DESA-6 was written at an appropriate reading level as no participants reported confusing regarding the wording of the questions or answer choices. There were no changes made to the DESA-6 following pilot testing.

Table 1 describes the demographics of athletes recruited in Phases I and II. There were more females than males in Phase II due to more females scoring positively in Phase I. Figure 1 demonstrates the flow of recruitment for Phase I and Phase II of the study.

Table 1 Participant Demographics
Fig. 1
figure 1

Flow of Participants

Table 2 describes the prevalence found in the adolescent athlete population during Phase I. More female athletes than male athletes were reported due to recruiting variance. Females (Median = 1) were shown to have a significantly higher rate of DE than males (Median = 0) with a Mann-Whitney U test for medians showing U = 7735, p < 0.001. This is consistent with populations of previous studies showing higher rates of DE amongst female athletes than male athletes. Total population rates were within the range of previous studies [15]. As participants were being screened for DE behaviors, a clinical spectrum of eating pathologies, other diagnoses are unlikely.

Table 2 Prevalence of DE

Table 3 describes the number of participants who screened positively in Phase I on either the DESA-6 or the Phase I EAT-26. Because athletes were selected at random from a larger population of athletes for participation, this resulted in slightly more females than males and a large discrepancy between the rate of response between each sport. We categorized by sport type, with 5 participants representing aesthetic sports, 234 participants playing ball, 56 from endurance sports, and 13 from weight-dependent sports. The highest number of positive results on either survey were from basketball athletes; however, there was a larger number of basketball athletes surveyed than any other sport.

Table 3 Number of Athletes who Screened Positive for DE & ED

Additionally, defining “at risk” student athletes as those with a score of 12 or greater on the EAT-26 (versus the typical positive score of 20 or greater), allowed the inclusion of athletes typically considered borderline cases. In Phase II, 41 athletes classified as “at risk” in Phase I consented to complete the EDE 17.0D, with 19 of the “at risk” group screening positive. Of the total 82 Phase II athletes, the DESA-6 screened positive for a total of 31 athletes while the EDE-17D found only 25 to have DE. Table 4 details the comparison of these scores. The DESA-6 favors a positive screen in borderline cases, screening positive for 7 (77.8%) of the borderline positive cases and 7 (75.0%) of the borderline negative cases.

Table 4 Overview of DESA-6 Scores in comparison to EDE-17 Subscales

Diagnostic accuracy of the DESA-6 to the EDE-17 and EAT-26

Table 5 shows the validity analyses for the DESA-6 versus the EAT-26 (Phase I) and versus the EDE 17.0D (Phase II) separated into male and female participants. Compared to the EAT-26 in the total sample analyses, the DESA-6 had a total sample sensitivity of 62.5% and a specificity of 87.7%. When comparing to the EDE 17.0D, the DESA-6 had a total sensitivity of 92% and specificity of 85.96%, respectively. Amongst males there was a sensitivity of 100% and a specificity of 85.0%. Amongst females there was a sensitivity of 90.9% and a specificity of 86.5%. The DESA-6 showed an overall accuracy of 87.80% when compared to the EDE 17.0D. Positive predictive value and negative predictive value are used to compare the rates in which positive and negative screens of the DESA-6 match the EAT-26 or EDE-17.

Table 5 DESA-6 Diagnostic Accuracy Measures of the DESA-6 to the EDE-17

Figure 2 further compares Phase II DESA-6 Scores and EDE 17.0D scores using ROC Curve analysis. This is a calculation of the overall ability of a parameter can distinguish between an experimental and control group. ROC-AUC showed a moderate approaching high correlation of 0.89 between the two measures [32].

Fig. 2
figure 2

ROC Curve Analysis of the DESA-6 in comparison to the EDE-17.0D

Construct validity

Construct validity of the DESA-6 was performed via regression analysis. Upon comparison of concurrent validity, Phase II DESA-6 had a strong significant positive correlation [r (23) = .80, p < 0.001] for males and [r (59) = .80, p < 0.001] for females when compared to Phase II EDE 17.0D. Comparison of the DESA-6 to the EAT-26 showed a low to moderate significant positive correlation with scores of [r (146) = .49, p < 0.001] in males and a moderate to strong positive correlation of [r (160) = .63, p < 0.001] in females.


Test-retest reliability comparing DESA-6 scores in Phase I to scores in Phase II showed a strong significant positive correlation [r (23) = 0.83, p < 0.001] in males and a significant positive correlation in females [r (59) = 0.76, p < 0.001].


The aim of this study was to validate a brief screening tool to detect risk for DE in adolescent athletes. The DESA-6 showed acceptability in detecting athletes both at risk for DE and athletes without risk for DE with sensitivity of 92% and specificity of 85.96%, respectively, when compared to the gold-standard EDE 17.0D. In comparison, the AMDQ had a sensitivity of 80% and a specificity of 77% [13, 20], the PST had a sensitivity of 86.5% and specificity of 77.7% [11], and the BEDA-Q had a sensitivity of 82.1% and specificity of 84.6% (20). These three screening tools were each created specifically for female collegiate athletes (AMDQ, PST) or female elite athletes (BEDA-Q), so it is worth noting that the DESA-6 had comparable sensitivity and specificity, but in a broader athletic population.

The proportion of adolescent athletes within this study population who were at risk for DE, as determined by the DESA-6, was 30.86%. Two studies conducted in collegiate female athlete populations found 35% of the female athletes had DE [11, 13] and a study by Torstveit et al. in 2008 on adolescent and adult elite female athletes found a prevalence of DE of 32.8% [33]. These prevalence rates are slightly higher than the results of this study, which is likely because our study included both male and female athletes. Although limited research exists on the individual prevalence of DE in female versus male athletes, a study by Sundgot-Borgen and Torstveit using the EDE 16.0 to assess EDs in adolescent and adult female and male elite athletes reported EDs in 20% of female athletes and 8% of male athletes [34]. This study also found a higher prevalence of DE in female athletes with 34.4% of female athletes scoring positive and 12.8% of male athletes scoring positive. The percent of male athletes scoring positive is also similar to a study by Rosendahl et al. in 2009 which found a DE prevalence rate of 10.4% in male high school athletes using the EAT-26 [35].

Although the results of the current study align similarly with previous research, it is important to note that each used different instruments to assess the presence of DE in their population, which contributes to the wide variation in reported rates. Additionally, none of these tools are specific to athletes, which also contributes to the variation.

Previous research has indicated athletes of aesthetic sports (e.g., wrestling, gymnastics) are at higher risk for DE/EDs [34, 36, 37]. In the present study, wrestlers showed the highest prevalence of DE (46%, n = 6 of 13, scored positive). A study of DE in Division I college athletes by Engel et al. found elevated levels of drive for thinness, food restriction and purging behaviors in wrestling compared to other sports [38]. Additionally, a study by Chapman and Woodman in 2016 found wrestlers reported a greater incidence of DE [39]. The number of gymnasts in our population was too low (n = 1) for further analysis. This is due to the high schools involved in this research, as only one offered gymnastics as a school sport.

Research has also indicated endurance athletes (e.g., cross-country, swimming) are at higher risk. In the present study in Phase II, 40% (n = 6 of 15) of cross-country athletes and 40% (n = 6 of 15) of swimmers scored positive on the DESA-6. Higher-than-average rates of DE among distance runners and swimmers was expected, as previous research has indicated endurance athletes are at higher risk for DE/EDs [34]. Additionally, a study by Schtscherbyna et al. reported 44.9% of adolescent female elite swimmers had DE [36].

Research has also indicated that risk for DE/EDs is not isolated to endurance or aesthetic sports [2, 11, 13, 18]. A study by Black et al. found DE/EDs among athletes representing 10 of 12 sports studied, including swimming, tennis, track & field, volleyball, cross country, golf, gymnastics, cheer, dance and modern dance [11]. A study by McNulty et al. found female athletes with EDs in 5 different sports, including crew, cross country/track & field, lacrosse, soccer and softball [18]. In the present study, DE was detected in athletes of 14 out of 17 sports represented. The highest prevalence of DE among ball sports was lacrosse (41%, n = 7 of 17 scored positive). This above-average prevalence may be due to all lacrosse athletes assessed being female. The only sports in this study without positive DESA-6 scores were cheer, ice hockey and rugby. This is likely due to the fact that the number of athletes assessed in each of these sports was small (n = 4, 1 and 2, respectively). Thus, this result does not mean athletes in these sports are not at risk for DE as the sample size was too small for analysis.

Strengths, limitations, and future directions

Strengths of the study include the fact that the sample size includes male athletes and multiple sport types. The DESA-6 is unique because of the inclusion of male athletes in our populations, which are typically not included in samples for statistical validation of screening tools regarding DE behavior risk. Additionally, the brevity and lack of required physiological measurements makes the DESA-6 easy to administer to athletes of all types. Also, this study was able to assess validity across multiple diagnostic tools. However, this study is not without limitations. First, the DESA-6 was pilot tested in an athlete population at the higher end (ages 18 to 19 years) of the age range of the validation population (ages 12–19 years) due to convenience, thus limiting our ability to interpret results from pilot testing for applicability in the lower end of the validation population age range. Second, as data were collected through self-report surveys, the results may be susceptible to response bias. We attempted to limit response bias as much as possible by having the athletes complete the EAT-26 and DESA-6 in private and assuring the participants that results would not be shared with coaches or their school. Additional response bias could be due to only 41 of 77 “at risk” athletes consenting to complete Phase II. Multiple attempts were made to contact all of the “at risk” athletes. Third, although the DESA-6 was developed based on the DSM-5 criteria for ED and expert opinion, the lack of a universally accepted definition and diagnostic criteria for DE gives room for question of the content validity of the DESA-6. However, the similarity in results between the DESA-6 and other screeners, as well as high sensitivity and specificity based on the gold-standard EDE 17.0D give support for validity of this screener. A further limitation of this study is that current results are specific to adolescent athletes of one geographical area (Midwest USA); therefore, generalizability to athletes of other ages or geographic regions cannot be done at this time. Additionally, participants in phase II of the study were not matched to a control by sport, though they were matched by age and gender. A further limitation may be the variations in length of time between completion of Phase I and Phase II, which unfortunately was due to the complicated practice and competition schedule of the student athletes. Further research aimed at addressing these concerns is necessary, including testing validity of the DESA-6 in adult athletes (ages > 19 years) and younger athletes of different levels of competition, inclusion of genders other than male or female, as well as different regions of the United States.


The DESA-6 is currently the shortest screening tool for flagging risk of DE in athletes, as well as the only known screening tool that can be used for adolescent athletes aged 13 to 19 years of both genders and all sports. A screening tool which assesses DE in athletes is essential due to the high prevalence of DE/EDs in athletes versus the general population [2, 19]. As all athletes face a unique variety of struggles pertaining to eating and nutritional needs, a tool that successfully accounts for male and female athletes of all sports and all levels of competition is essential. Additionally, the use of screening tools for the early detection of DE in athletes has been recommended to help prevent long term consequences, as well as to provide benefits associated with early intervention. This necessity has been recognized by multiple previous studies, as well as the NATA, IOC, and ACSM [7, 16, 17, 20]. A shorter, more efficient screening tool for DE that could be given to a large athlete population without extensive investment of time and resources is necessary, as no current tools can quickly evaluate a large group of adolescent athletes [10, 19]. With only 6 questions, the DESA-6 can be used by physicians, athletic trainers, registered dietitians, or other sport/healthcare personnel to screen an athlete or an entire team quickly and efficiently (brief administration and scoring duration). The DESA-6 does not require psychometric expertise, an extensive time commitment, nor a large number of resources, thus making it a promising tool for risk assessment of DE in adolescent athletes. Future research investigating the validity of the DESA-6 in additional athlete populations, such as adult athletes and sports that were under-represented in this study, could further expand the generalizability of the DESA-6. This could potentially result in the ability to use the DESA-6 to assess all athletes, no matter age, gender, sport, or level of competition.

Availability of data and materials

The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. Additionally, the results of the present study do not constitute endorsement by ACSM. The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.



Disordered eating


Eating disorder


Anorexia nervosa


Bulimia nervosa


Eating Attitudes Test


Other specified feeding and eating disorder


Disordered Eating Screen for Athletes


Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition


Binge Eating Disorder


National Athletic Trainers’ Association


American College of Sports Medicine


International Olympic Committee


Eating Disorder Examination Questionnaire


Eating Disorder Examination


Athletic Milieu Direct Questionnaire


Female Athlete Screening Tool


Physiologic Screening Test


Brief Eating Disorder in Athletes Questionnaire


  1. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 5th ed. Arlington (VA): American Psychiatric Association; 2013.

    Book  Google Scholar 

  2. Bryne S, McLean N. Eating disorders in athletes: a review of the literature. J Sci Med Sport. 2001;4(2):145–59.

    Article  Google Scholar 

  3. Beals K. Disordered eating among athletes: a comprehensive guide for health professionals. Hum Kinet. 1st edition (April 12, 2004). 2004:4–19.

  4. Sundgot-Borgen J, Torstveit MK. Aspects of disordered eating continuum in elite high-intensity sports. Scand J Med Sci Spor. 2010;20(S2):112–21.

    Article  Google Scholar 

  5. Baum A. Eating disordered in the male athlete. Sports Med. 2006;36(1):1–6.

    Article  Google Scholar 

  6. Mountjoy M, Sundgot-Borgen J, Burke L, et al. The IOC consensus statement: beyond the Female Athlete Triad – Relative Energy Deficiency in Sport (RED-S). Br J Sport Med. 2014;48:491–7.

    Article  Google Scholar 

  7. Bonci CM, Bonci LJ, Granger LR, et al. National Athletic Trainers’ Association position statement: preventing, detecting, and managing disordered eating in athletes. J Athl Train. 2008;43(1):80–108.

    Article  Google Scholar 

  8. Rauh MJ, Nichols JF, Barrack MT. Relationships among injury and disordered eating, menstrual dysfunction, and low bone mineral density in high school athletes: a prospective study. J Athl Train. 2010;45(3):243–252.B.

    Article  Google Scholar 

  9. Hudson JI, Hiripi E, Pope HG, Kesslet RC. The prevalence and correlates of eating disorders in the national comorbidity survey replication. Biol Psychiatry. 2007;61:348–58.

    Article  Google Scholar 

  10. DePalma MT, Koszewski WM, Romani W, Chase JG, Suiderhof NJ, McCoy PM. Identifying college athletes at risk for pathogenic eating. Br J Sports Med. 2002;36:45–50.

    Article  Google Scholar 

  11. Black DR, Larkin DC, Coster DC, Leverenz LJ, Abood DA. Physiologic screening test for eating disorders/disordered eating among female collegiate athletes. J Athl Train. 2003;38(4):286–97.

    PubMed  PubMed Central  Google Scholar 

  12. Cotton M, Ball C, Robinson P. Four simple questions can help screen for eating disorders. J Gen Intern Med. 2003;18:53–6.

    Article  Google Scholar 

  13. Nagel DL, Black DR, Leverenze LJ, Coster DC. Evaluation of a screening test for female college athletes with eating disorders and disordered eating. J Athl Train. 2000;35(4):431–40.

    PubMed  PubMed Central  Google Scholar 

  14. Martinsen M, Sherman RT, Thompson RA, Sundgot-Borgen J. Coaches’ knowledge and management of eating disorders: a randomized controlled trial. Med Sci Sports Exerc. 2015;47(5):1070–8.

    Article  Google Scholar 

  15. Martinsen M, Sundgot-Borgen J. Higher prevalence of eating disorders among adolescent elite athletes than controls. Med Sci Sports Exerc. 2013;45(6):1188–97.

    Article  Google Scholar 

  16. Nattiv A, Loucks AB, Manore MM, Sanborn CF, Sundgot-Borgen J, Warren MP. American College of Sports Medicine Position Statement: the female athlete triad. Med Sci Sports Exerc. 2007;39(10):1867–82.

    Article  Google Scholar 

  17. Ljungqvist A, Jenoure P, Engbretsen L, et al. The International Olympic Committee (IOC) Consensus Statement on periodic health evaluation of elite athletes. Br J Sports Med. 2009;43(9):631–43.

    Article  Google Scholar 

  18. McNulty KY, Adams CH, Anderson JH, Affenito SG. Development and validation of a screening tool to identify eating disorders in female athletes. J Am Diet Assoc. 2001;101(8):886–92.

    Article  Google Scholar 

  19. Martinsen M, Holme I, Pensgaard AM, Torstveit MK, Sundgot-Borgen J. The development of the Brief Eating Disorder in Athletes questionnaire. Med Sci Sports Exerc. 2014;46(8):1666–75.

    Article  Google Scholar 

  20. Knapp J, Aernia G, Anderson A. Eating disorders in female athletes: use of screening tools. Curr Sports Med Rep. 2014;13(4):214–8.

    Article  Google Scholar 

  21. Garner DM, Garfinkel PE. The Eating Attitudes Test: an index of symptoms of anorexia nervosa. Psychol Med. 1979;9:273–9.

    Article  Google Scholar 

  22. Mintz LB, O’Halloran MS. The Eating Attitudes Test: validation with DSM-IV eating disorder criteria. J Pers Assess. 2000;74(3):489–503.

    Article  Google Scholar 

  23. Garfinkel PE, Newman A. The Eating Attitudes Test: twenty-five years later. Eat Weight Disord. 2001;6:1–24.

    Article  Google Scholar 

  24. Haase AM. Weight perception in female athletes: associations with disordered eating correlates and behavior. Eat Behav. 2011;12:64–7.

    Article  Google Scholar 

  25. Cooper Z, Fairburn CG. The Eating Disorder Examination: A semi-structured interview for the assessment of the specific psychopathology of eating disorders. Int J Eat Disord. 987;6(1):1–8.;2-9.

  26. Berg KC, Peterson CB, Frazier P, Crow SJ. Psychometric evaluation of the Eating Disorder Examination and Eating Disorder Examination-Questionnaire: a systematic review of the literature. Int J Eat Disorder. 2012;45:428–38.

    Article  Google Scholar 

  27. Kennedy SF, Ley AF, Fugate BN, Hayek A. Factors associated with disordered eating behaviors among triathletes. Spartan Med Res J. 2017;2(1):1–14.

    Article  Google Scholar 

  28. Ghoch ME, Soave F, Calugi S, Grave RD. Eating disorders, physical fitness and sport performance: a systematic review. Nutrients. 2013;5:5140–60.

    Article  Google Scholar 

  29. Hulley AJ, Hill AJ. Eating disorders and health in elite women distance runners. Int J Eat Disord. 2001;30(3):312–7.

    Article  Google Scholar 

  30. Fawcett T. An introduction to ROC analysis. Pattern Recog Letters. 2006;27:861–74.

    Article  Google Scholar 

  31. Schoonjans, F. (2020). MedCalc's Free statistical calculators. Retrieved Febuary 21, 2020, from

    Google Scholar 

  32. Steiner DL, Cairney J. What’s under the ROC? AN introduction to receiver operating characteristics curves. Can J Psychiatry. 2007;52(2):121–8.

    Article  Google Scholar 

  33. Torstveit MK, Rosenvinge JH, Sundgot-Borgen J. Prevalence of eating disorders and the predictive power of risk models in female elite athletes: a controlled study. Scan J Med Sci Spor. 2008;18(1):108–18.

    Article  Google Scholar 

  34. Sundgot-Borgen J, Torstveit MK. Prevalence of eating disorders in elite athletes is higher than in the general population. Clin J Sport Med. 2004;14:25–32.

    Article  Google Scholar 

  35. Rosendahl J, Bormann B, Aschenbrenner K, Aschenbrenner F, Strauss B. Dieting and disordered eating in German high school athletes and non-athletes. Scand J Med Sci Sports. 2009;19(5):731–9.

    Article  Google Scholar 

  36. Schtscherbyna A, Soares EA, de Oliveria FP, Ribeiro BG. Female athlete triad in elite swimmers of the city of Rio de Jeneiro, Brazil. Nutrition. 2009;25(6):634–9.

    Article  Google Scholar 

  37. Martinsen M, Bratland-Sanda S, Eriksson AK, Sundgot-Borgen J. Dieting to win or to be thin? A study of dieting and disordered eating among adolescent elite athletes and non-athlete controls. Br J Sports Med. 2010;44(1):70–6.

    Article  Google Scholar 

  38. Engel SG, Johnson C, Powers PS, Crosby RD, Wonderluch SA, Wittrock DA, Mitchell JE. Predictors of disordered eating in a sample of elite division I college athletes. Eat Behav. 2003;4:333–43.

    Article  Google Scholar 

  39. Chapman J, Woodman T. Disordered eating in male athletes: a meta-analysis. J Sports Sci. 2016;34(2):101–10.

    Article  Google Scholar 

Download references


The authors would like to acknowledge Dr. Robert Norris for his assistance; Jacob Babb, Zach Bauer, Tonye Burutolu, Jackie Davis, Cailyn D’Haem, Dr. Alexander Hayek, Ali Moshrefi, Tom Porter, Moshiur Rahman, Jack Sims, Melanie Smalls, and Peter Stephan for their assistance with data collection.


This research was made possible through the American Academy of Child and Adolescent Psychiatry (AACAP) Pilot Research Award for Junior Faculty and Child and Adolescent Psychiatry Fellows, supported by AACAP; its contents are the responsibility of the authors and do not necessarily reflect the official views of AACAP. The role of the AACAP Pilot Research Award was entirely participant compensation.

Author information

Authors and Affiliations



SK, JK, EM, designed the study protocol. RM, DG, HK recruited/interviewed participants according to the protocol. HK helped contact and recruit local high schools and obtained administrative permission to run the study. DG formulated the Tables. SK, JK, EW, RM, DG, wrote the first draft of the manuscript. SK supervised all other authors and analyzed/revised the article critically. All authors contributed to the interpretation of data. All authors contributed to and have approved the final manuscript.

Authors’ information

RM, DG, and SK have recently been accepted into JOED with the article, “Prevalence of Disordered Eating in Athletes Categorized by Emphasis on Leanness and Activity Type – A Systematic Review”. RM and DG have been invited to speak regarded disordered eating and eating disorders in athletes from several organizations, including the Michigan Osteopathic Neuropsychiatric Society, for continuing medical education credit. RM and DG were awarded second place at the Michigan Osteopathic Association Scientific Research Exhibit competition, where they presented on the various types of athletes that exist in literature. RM was awarded first place at the research competition hosted by the Michigan Association of Osteopathic Family Practitioners for a similar topic. SK is a child and adolescent psychiatrist who is primarily focused on the treatment of disordered eating and eating disorders, especially those affecting athletes.

Corresponding author

Correspondence to Samantha F. Kennedy.

Ethics declarations

Ethics approval and consent to participate

Informed consent was obtained at all stages of participation within this study and no attempt at deception was made. This study was approved by the Michigan State University Institutional Review Board, under study protocol number LEGACY17–1078.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing financial interests or other conflicts of interest.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

DESCRIPTION: This manuscript introduces and validates the Disordered Eating Screen for Athletes (DESA-6) screening tool. This tool is currently designed to identify risk for disordered eating in adolescent athletes. Identification of disordered eating risk is very important because early intervention has been linked with improved treatment. Creation of this tool would allow athletic trainers, physicians, and concerned school nurses or administrators to quickly identify student athletes who may be at risk for disordered eating. Future research will attempt to validate the DESA-6 in adult, college, and professional athletes.

Supplementary Information

Additional file 1.

Disordered Eating Screening for Athletes: 6 Question Screening Tool.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Kennedy, S.F., Kovan, J., Werner, E. et al. Initial validation of a screening tool for disordered eating in adolescent athletes. J Eat Disord 9, 21 (2021).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: