Longer-Term Follow-Up of College Students Screening Positive for Anorexia Nervosa: Psychopathology, Help Seeking, and Barriers to Treatment

Abstract

The purpose of this study was to conduct a longer-term (i.e., 9-month) follow-up of students identified with possible anorexia nervosa (AN) as part of the Healthy Body Image Program, an online platform for screening and delivering tailored feedback and interventions, offered at 36 U.S. universities. Participants were 61 individuals who screened positive for AN and who completed the follow-up. Regarding results, some indices of ED pathology and psychiatric comorbidity decreased over time, while others did not. Participants most commonly endorsed feeling ashamed, nervous, validated, and sad in response to receiving the referral. One-third (33%) reported already being in treatment at the time they received the referral, 26% initiated treatment since that time, and 41% did not initiate treatment. The most common reasons for seeking treatment were emotional distress, concern with eating, and health concerns. The strongest treatment barriers were believing one should be able to help themselves, believing the problem was not serious enough to warrant treatment, and not having time. Findings highlight the high level of pathology in students identified with possible AN, even nine months after they were first identified and provided resources, and the relatively low rates of treatment utilization given the seriousness of these illnesses.

Eating disorders (EDs) are serious mental illnesses associated with high morbidity and mortality, impairment, and comorbid psychopathology, and anorexia nervosa (AN) is associated with the highest mortality rate of any psychiatric disorder (Klump, Bulik, Kaye, Treasure, & Tyson, 2009). EDs are especially common in college-age individuals; however, less than 20% of students with EDs receive treatment (Eisenberg, Nicklett, Roeder, & Kirz, 2011). Treatment delays result in prolonged illness and poorer prognosis. Along with issues in treatment access, ED screening on college campuses is lacking, with only 22% of colleges offering year-round screening opportunities (National EDs Association, 2013). Thus, opportunities for improved screening of and intervention for EDs on college campuses are indicated.

Internet-based technologies have the potential to improve care for EDs on college campuses and in other settings by overcoming barriers to treatment, offsetting in-person clinical demands, increasing access, and reducing costs. As such, we developed the Healthy Body Image (HBI) Program—an online platform for screening and delivering tailored screen feedback and interventions (e.g., Fitzsimmons-Craft et al., 2018; Wilfley, Agras, & Taylor, 2013). HBI identifies individuals at low risk for, high risk for, or with a possible clinical/subclinical ED (other than AN) and offers tailored evidence-based online/mobile interventions. HBI also identifies individuals with possible AN and offers a referral to in-person care at their respective institution. HBI thus directly links from screening to online/mobile interventions for the majority of participants and demonstrates program uptake of 44–51% for individuals at low risk for, high risk for, or with a possible clinical/subclinical ED other than AN (Fitzsimmons-Craft et al., 2018). Given the more intensive medical monitoring that is warranted for individuals with possible AN, HBI provides a referral to in-person treatment only. Although the efficacy of the online/mobile interventions utilized has been established (Beintner, Jacobi, & Taylor, 2012; Saekow et al., 2015), to date, the rate of treatment uptake for individuals with possible AN identified through HBI is unknown. Given that illness duration can last for 20+ years for more than half of those with AN (Fichter, Quadflieg, Crosby, & Koch, 2017), there is a pressing need for strategies that promote early identification of symptoms and treatment linkage. It is thus of the utmost importance to understand the impact of the strategy utilized for individuals identified with possible AN in HBI.

A systematic review of 13 studies of perceived barriers and facilitators toward help-seeking for EDs identified stigma and shame, denial of illness severity, practical barriers (e.g., cost), low motivation, negative attitudes toward help-seeking, lack of encouragement, and lack of knowledge about resources as the most prominent barriers (Ali et al., 2017). Facilitators of help-seeking included the presence of other mental health problems and concerns about health. The included studies mainly utilized transdiagnostic adult female samples and very few focused on college students. To our knowledge, no research has specifically examined reasons for or barriers to seeking treatment in a sample of college students identified with possible AN.

The purpose of this study was to conduct a longer-term (i.e., 9-month, which was chosen so as to examine change over the time period of an academic year) follow-up of students identified with possible AN as part of the HBI Program offered at 36 U.S. universities. We aimed to report, in a primarily descriptive fashion, on: 1) change in ED pathology and psychiatric comorbidity over time; 2) reactions to receiving a referral for further evaluation and/or treatment for eating and body image concerns, including examination of any differences between those who were and were not in treatment at the time; 3) reasons for seeking treatment; 4) suggestions for increasing acceptance of the referral; 5) insurance coverage; 6) health care utilization and differences in change in symptoms between those who were already in treatment for eating related problems at the time of receiving the referral, those who initiated treatment since that time, and those who had not initiated treatment; and 7) barriers to treatment. These data will shed light on the outcomes and health care utilization of college students identified with possible AN, as well as reasons for seeking and barriers to treatment, all of which will inform best practices for engaging college students with possible AN with treatment.

Method

Participants and Procedure

Participants for this study were identified via two separate HBI initiatives: a statewide initiative in Missouri (N = 8 participating universities; Fitzsimmons-Craft et al., 2018) and a national randomized controlled trial (N = 28 participating universities). All students on the participating college campuses ages 18+ years were eligible for participation. Recruitment efforts primarily targeted undergraduate students, but graduate students and postdoctoral fellows were also eligible. Both projects received institutional review board approval at the coordinating university, and all participants provided informed consent.

All participating universities were required to identify an on-campus clinical representative, typically a counselor in the counseling or health center, to maximize student uptake and safety. In collaboration with this individual, we collaboratively developed campus-specific recruitment strategies, including use of: electronic media; social media; printed materials; presentations; and counselors offering HBI to individuals in need. We note that individuals thus voluntarily elected to participate in this screening initiative and that HBI was not administered as a population-based screen.

As part of the HBI screen, students completed the Stanford-Washington University ED Screen (SWED), and individuals with possible AN were identified. The SWED is a brief screening tool that assesses ED behaviors, pathology, and impairment to detect individuals at high risk for or with an ED (Graham et al., 2018). The SWED includes questions from the Weight Concerns Scale (WCS; Killen et al., 1994) and items adapted from the ED Examination-Questionnaire (Fairburn & Beglin, 1994), the ED Diagnostic Scale (Stice, Telch, & Rizvi, 2000), and the Clinical Impairment Assessment (Bohn & Fairburn, 2008). Based on available data, responses are used to sort individuals into one of four categories: 1) possible AN, based on body mass index (BMI) and elevated weight and shape concerns; 2) possible clinical/subclinical ED other than AN, based on binge eating and/or purging behaviors in the past 3 months; 3) high risk for an ED, based on elevated weight and shape concerns; and 4) low risk for an ED, based on not screening into one of the above categories. The SWED screening algorithm has been validated and used in past research, with sensitivity and specificity being high for cases of DSM-5 EDs compared to diagnostic interview (≥0.78 for AN, bulimia nervosa, and binge eating disorder; see Graham et al., 2018). Most relevant to the current study, individuals were identified with possible AN based on a) lowest past year BMI (based on self-reported height and weight) OR current BMI < 18.5 kg/m²; and b) WCS ≥ 59, indicating very high weight and shape concerns.

Across these two initiatives, the HBI screen was completed 7,348 times by unique respondents,¹ with 234 of those (3.2%) screening positive for possible AN. We note that this prevalence rate is similar, albeit slightly elevated, compared to other work screening college students. For instance, the American College Health Association’s National College Health Assessment (2018) found that 1.9% of females and 0.7% of males reported diagnosis or treatment for AN in the past 12 months. Elevated prevalence in HBI may reflect the use of a diagnostic screening tool (rather than self-reported diagnosis/treatment) or that individuals with these concerns voluntarily elected to take the HBI screen at rates elevated compared to individuals without these concerns.

Upon completing the screen, individuals screening positive for possible AN saw the following: “Your results suggest that you are experiencing concerns about your weight and shape and are engaging in behaviors that are likely interfering with your health and happiness. These symptoms are comparable to individuals with EDs or who are at a great risk of developing one. We recommend that you seek further evaluation and/or treatment, as it may be beneficial for you to be evaluated by a mental health professional and/or medical doctor. Resources available on your campus are listed below” [with campus-specific information on how to make an appointment at the counseling/health center provided]. We note that the baseline information on the individuals screening positive for possible AN across the two HBI initiatives has been previously presented by Fitzsimmons-Craft et al. (2018, 2019). Of these 234 individuals, 123 completed a consent form that stated that their contact information would be included in a registry for future follow-up research. That is, we did not have consent to contact the remaining 111 individuals who screened positive for AN for participation in follow-up research. These individuals did not opt out of participation in follow-up research; they were simply not asked to provide their consent—this aspect was added to the consent in a later version. The 123 individuals for whom we had permission to contact for follow-up research were contacted by email and phone to complete a follow-up survey nine months after they originally completed the HBI screen in exchange for a $5 gift card. After one week of non-response to the invitation, participants were sent a reminder email. After an additional two weeks of non-response, a final reminder email was sent and a final reminder phone call was made.

Measures

Demographics.

Participants self-reported on their age, gender, race, ethnicity, height, current weight, and lowest past year weight. Height, current weight, and lowest past year weight were collected as part of both the original HBI screen and the 9-month follow-up and were used to calculate current and past year lowest BMI at both timepoints.

Stanford-Washington University ED Screen (SWED).

As mentioned, the SWED is a brief screening tool that assesses ED behaviors, pathology, and impairment (Graham et al., 2018). The SWED was completed as part of both the original HBI screen and the 9-month follow-up.

Psychiatric Comorbidity.

To assess psychiatric comorbidity, three questions were used: 1) “On average, during the past two weeks have you felt depressed or hopeless?”; 2) “On average, during the past two weeks have you felt anxious or tense?”; and 3) “On average, during the past two weeks, has it been difficult for you to fall asleep or stay asleep at night?” These questions were responded to on a 5-point scale ranging from not at all to extremely and were completed as part of both the original HBI screen and the 9-month follow-up.

Questions about Receiving the Referral.

Participants were asked a series of questions developed for the current study regarding their reaction to receiving the referral at the 9-month follow-up: 1) “Do you remember receiving the referral for further evaluation and/or treatment for eating and body image concerns” (yes/no); 2) “What was your reaction to receiving this referral?”, with the following options provided: surprised; happy; proud; validated; annoyed; angry; sad; nervous; ashamed; other (select all that apply); if a response option was selected, participants were asked to explain what made them feel that way; and 3) “Were you already receiving treatment for eating or body image concerns at the time you received the referral?” (yes/no).

Reasons for Seeking Treatment.

The following questions about reasons for seeking treatment were modified from Cachelin and Striegel-Moore (2006) and asked at the 9-month follow-up: 1) “At the time you received a referral for treatment from the HBI Program, did you believe you had a significant eating problem” (yes/no); 2) “Since that time, do you believe you have had a significant eating problem?” (yes/no); 3) “In the past nine months, have you wanted help for an eating problem?” (yes/no); 4) “In the past nine months, have you sought treatment from a physician, counselor, or other health care provider for an eating or associated weight problem?” (yes/no); and 5) “What made you decide to seek treatment?”, with the following options provided: concern with weight; encouragement of friends; concern with eating; initiated by parents; emotional distress; health concerns; or other (select all that apply). If participants chose other, they were asked to describe.

Ways to Increase Referral Acceptance.

At the 9-month follow-up, participants were asked, “Is there anything we could have done differently that might have increased your initial acceptance of the referral?”, with the following options provided: provided more specific information on why exactly the referral was being made; provided more information about what eating and body image concerns entail; provided more information about the problems associated with eating and body image concerns; provided you with a brief online program designed to help you explore what (if anything) you wished to change, the function the eating and weight concerns served in your life, and your values and what (if anything) would have needed to change in order for you to live in accordance with them; or other (select all that apply). If participants chose other, they were asked to describe.

Insurance Coverage.

The following questions about insurance coverage were modified from Cachelin and Striegel-Moore (2006): 1) “During the past nine months, were you covered by any health insurance?” (yes/no); if yes: “For how many months of the past nine months were you covered by health insurance?”; 2) “During the past nine months, were you unable to obtain certain medical or psychological treatments because you had no insurance for these services?” (yes/no); and 3) “During the past nine months, were you unable to obtain certain medical or psychological treatments because you had exceeded the amount of coverage your insurance policy allows for these services?” (yes/no).

Health Care Utilization.

Participants were asked questions about their health care utilization in the past nine months including: receipt of treatment for eating related problems (e.g., outpatient, inpatient); receipt of treatment for any other mental health problem (e.g., outpatient, inpatient); changes to physical health or receipt of medical treatment for any reason (e.g., outpatient medical care, emergency room); use of prescription psychiatric medication, including for an eating related problem; and utilization of self-help programs for eating related problems. Participants were also asked whether a health care provider had specifically diagnosed them with an ED in the past nine months. These questions were presented in yes/no format.

Barriers to Seeking Treatment.

If participants reported that they did not receive any treatment for eating problems in the past nine months, they responded to 18 items modified from Cachelin and Striegel-Moore (2006) regarding barriers to seeking treatment. These items were responded to on a 5-point scale ranging from strongly disagree to strongly agree. Alpha for the current study was .84.

Analytic Strategy

We used paired-samples t-tests to examine change in ED pathology and psychiatric comorbidity from the time of the initial screen to the 9-month follow-up. Questions about receiving the referral, reasons for seeking treatment, ways to increase acceptance of the referral, insurance coverage, health care utilization, and barriers to seeking treatment were examined descriptively. Finally, we used analyses of variance (ANOVAs) and chi-square tests to compare change in ED pathology and psychiatric comorbidity, as well as student status, between those who were already in treatment for eating related problems at the time of receiving the referral, those who initiated treatment since that time, and those who had not initiated treatment.

Results

Descriptives

Of the 123 individuals who screened positive for possible AN and were eligible for participation in the 9-month follow-up study, 61 completed (49.6%). At the time of the initial screen, these 61 completers ranged in age from 18–34 years, with a mean age of 20.95 years (SD = 3.34). Participants were primarily female (95.1%) with fewer identifying as male (3.3%) or gender non-conforming (1.6%). Most participants identified as White (57.4%), 24.6% as Asian or South Asian, 9.8% as multiracial, and 6.6% as other races (1.6% missing). Regarding ethnicity, 13.1% identified as Hispanic. The 61 completers attended 18 different universities, including public and private universities and small and large schools. There were no differences between those who did (n = 61) and did not (n = 62) complete the follow-up on demographics, ED pathology, or psychiatric comorbidity (ps > .099). We note that at the time of completing the follow-up, 70.5% of participants reported being undergraduate students, 19.7% graduate students, and 9.8% not currently being a student.

Comparison of ED Pathology and Psychiatric Comorbidity Over Time

Table 1 depicts analyses examining change in ED pathology and psychiatric comorbidity from the time of the initial screen to the 9-month follow-up. Past year lowest BMI, weight concerns, and excessive exercise frequency significantly improved (ps < .009), while other indices of ED pathology did not (ps > .059). In terms of ED-related clinical impairment, impairment in school work and self-image significantly decreased over time (ps = .005), while impairment in relationships did not (p = .160). In terms of psychiatric comorbidity, depression and anxiety significantly decreased over time (ps < .018), while sleep problems did not (p = .096). Levels of all ED pathology and psychiatric comorbidity indices remained high at the 9-month follow-up.

Table 1.

Comparison of Eating Disorder Pathology and Psychiatric Comorbidity Over Time (N = 61)

	Initial Screen	9-Month Follow-Up	Significance
Eating disorder symptoms
Current body mass index	18.97 (0.44)	19.92 (0.45)	t(59) = −1.92; p=.060 d = −0.25
Lowest past year body mass index	17.22 (1.80)	18.35 (1.97)	t(58) = −3.02; p=.004 d = −0.39
Weight Concerns Scale score	77.19 (13.04)	65.22 (17.96)	t(60) = 5.37; p<.001 d = 0.69
Binge eating frequency (past 3 months)	12.54 (23.46)	9.07 (19.66)	t(58) = 1.56; p=.125 d = 0.20
Vomit frequency (past 3 months)	10.93 (32.66)	8.74 (30.34)	t(60) = 0.74; p=.460 d = 0.10
Diuretics/laxatives frequency (past 3 months)	2.46 (7.06)	2.33 (8.66)	t(60) = 0.15; p=.885 d = 0.02
Excessive exercise frequency (past 3 months)	16.66 (25.43)	9.15 (20.91)	t(60) = 2.74; p=.008 d = 0.35
Fasting frequency (past 3 months)	10.41 (16.99)	7.89 (15.32)	t(60) = 1.08; p=.286 d = 0.14
Eating disorder-related clinical impairment
School work	3.38 (1.11)	2.92 (1.07)	t(60) = 2.91; p=.005 d = 0.37
Relationships	3.46 (1.10)	3.25 (1.09)	t(60) = 1.42; p=.160 d = 0.18
Self-image	4.26 (0.75)	3.89 (0.99)	t(60) = 2.89; p=.005 d = 0.37
Psychiatric comorbidity
Depression	3.20 (1.28)	2.82 (1.27)	t(60) = 2.46; p=.017 d = 0.31
Anxiety	3.97 (1.03)	3.54 (1.18)	t(60) = 3.48; p=.001 d = 0.45
Sleep problems	3.05 (1.37)	2.72 (1.38)	t(60) = 1.69; p=.096 d = 0.22

Questions about Receiving the Referral

The majority of participants (93.4%) remembered receiving a referral for further evaluation and/or treatment for eating and body image concerns. In terms of reaction to receiving the referral, 16.4% endorsed feeling surprised (e.g., “I didn’t think I had disordered eating habits,” “I didn’t think I needed it”), 8.2% happy (e.g., “I was happy that what I was doing was considered disordered so I didn’t feel as bad if I didn’t do the habits,” “Of making progress”), 4.9% proud (e.g., “Felt some sense of self-control,” “I hate to say it, but I felt proud that my ability to limit my food intake had reached a concerning level”), 26.2% validated (e.g., “I didn’t feel I really deserved help until someone else told me I had a problem,” “I know I need help”), 16.4% annoyed (e.g., “At the time, I didn’t think I needed any help,” “I thought I had overcome my problems”), 3.3% angry (e.g., “I am angry at myself,” “I do not want my weight to affect me as much as it does”), 24.6% sad (e.g., “Sad that I’m struggling with such a difficult issue,” “Sad that this was my life”), 36.1% nervous (e.g., “About change,” “Nervous as I realized I may have to enter treatment”), 37.7% ashamed (e.g., “I hate that I feel that way about my body,” “I don’t like admitting that I’m insecure about my body”), and 24.6% other (e.g., “Apathetic—I’m not too concerned what a survey thinks of me,” “Expected it”). About a third (32.8%) of participants reported that they were already receiving treatment for eating or body image concerns at the time they received the referral. When examining reactions to receiving the referral between those were (n = 20) and were not (n = 41) in treatment at the time of the referral, there were no differences (ps > .067).

Reasons for Seeking Treatment

Nearly two-thirds (62.3%) of participants believed they had a significant eating problem at the time they received a referral from HBI, with 57.4% believing they had a significant eating problem since that time. Over 60% (60.7%) reported wanting help for an eating problem in the past nine months, with 59.0% actually seeking treatment for an eating or associated weight problem in the past nine months. In terms of what made participants decide to seek treatment, 19.7% endorsed concern with weight, 11.5% encouragement of friends, 36.1% concern with eating, 14.8% initiated by parents, 45.9% emotional distress, 31.1% health concerns, and 11.5% other (e.g., “Couldn’t focus on anything: it was controlling my life,” “Recommended by MD”).

Ways to Increase Acceptance of Referral

Regarding things that could have been done to increase acceptance of the referral, 32.8% endorsed providing more specific information on why the referral was being made, 23.0% endorsed providing more information about what eating and body image concerns entail, 19.7% endorsed providing more information about the problems associated with eating and body image concerns, 50.8% endorsed providing a brief online program, and 19.7% endorsed other (e.g., “Nothing else. The problem lies with me, not with other people trying to persuade me”).

Insurance Coverage

Regarding insurance coverage, 95.1% of participants endorsed being covered by some type of health insurance in the past nine months, with the majority of those (94.8%) being covered for all nine months. Some participants endorsed being unable to obtain certain treatments because they had no insurance for these services (13.1%) or because they had exceeded the amount of coverage their insurance policy allowed (15.5%).

Health Care Utilization

Health care utilization data are presented in Table 2. In the last nine months, 52.5% of the sample reported receiving treatment for eating related problems, 44.3% reported receiving treatment for any other mental health problem, 36.1% reported experiencing changes to their physical health or receiving medical treatment for any reason, 44.3% had taken prescribed psychiatric medication, and 23.0% had utilized a self-help program for eating related problems.

Table 2.

Health Care Utilization (N = 61)

	n (% of total N of 61)
1. Any treatment for eating related problems in the last nine months	32 (52.5%)
1.a. Any outpatient treatment	22 (36.1%)
1.a.1. Individual therapy	21 (34.4%)
1.a.2. Group therapy	9 (14.8%)
1.a.3. Weight loss program	1 (1.6%)
1.a.4. Nutritional counseling	15 (24.6%)
1.a.5. Other outpatient treatment not mentioned above	5 (8.2%)
1.b. Any inpatient treatment, partial hospitalization/day treatment, or residential treatment	6 (9.8%)
1.b.1. Inpatient treatment	6 (9.8%)
1.b.2. Partial hospitalization or day treatment	4 (6.6%)
1.b.3. Residential treatment	2 (3.3%)
2. Any treatment for any other mental health problem (i.e., non-eating related problem) in the last nine months	27 (44.3%)
2.a. Any outpatient treatment	22 (36.1%)
2.a.1. Individual therapy	21 (34.4%)
2.a.2. Group therapy	2 (3.3%)
2.a.3. Other outpatient treatment not mentioned above	7 (11.5)
2.b. Any inpatient treatment	1 (1.6%)
3. Changes to physical health or received medical treatment for any reason in the last nine months	22 (36.1%)
3.a. Outpatient medical care	9 (14.8%)
3.b. Emergency room	6 (9.8%)
3.c. Hospitalized	4 (6.6%)
4. Taken prescribed psychiatric medication in the last nine months	27 (44.3%)
4.a. Psychiatric medication was for an eating related problem	10 (16.4%)
5. Utilized a self-help program for eating related problems in the last nine months	14 (23.0%)
5.a. Book	8 (13.1%)
5.b. Online support group	2 (3.3%)
5.c. Mobile app	2 (3.3%)
5.d. Online program	3 (4.9%)
5.e. Online counseling	0 (0.0%)
6. During the past nine months, health care provider specifically diagnosed with:
Anorexia nervosa	17 (27.9%)
Bulimia nervosa	5 (8.2%)
Binge eating disorder	4 (6.6%)
Other specified feeding or eating disorder or eating disorder not otherwise specified	7 (11.5%)
Received diagnosis other than the ones listed above	7 (11.5%)
I don’t know my diagnosis	30 (49.2%)

Overall, 20 individuals (32.8%) reported already being in treatment at the time they received the referral, 16 (26.2%) initiated treatment for eating related problems in the last nine months, and 25 (41.0%) did not initiate treatment since receiving the referral. That is, of those who were not in treatment at the time of the referral, 39.0% initiated treatment. Only 27.9% of the sample was specifically diagnosed by a health care provider with AN in the past nine months. When comparing those who were already in treatment for eating related problems at the time of receiving the referral, those who initiated treatment since that time, and those who had not initiated treatment on change from initial screen to 9-month follow-up on ED pathology and psychiatric comorbidity (i.e., variables listed in Table 1) using ANOVAs, groups differed on change in excessive exercise (p = .004, partial η² = .18), with follow-up Tukey tests indicating that those who had initiated treatment in the last nine months demonstrated greater improvement than those who had not initiated treatment (p = .003). Groups did not differ on change in any other construct (ps > .116, partial η²s < .08). We also compared those who were already in treatment for eating related problems at the time of receiving the referral, those who initiated treatment since that time, and those who had not initiated treatment by student status (i.e., undergraduate student, graduate student, or not currently a student) using a chi-square test. Results indicated no group differences (p = .097).

Treatment Barriers

Participants who reported that they did not receive any treatment for eating problems in the past nine months (n = 29 out of 61 survey completers = 47.5%) were prompted to respond to questions regarding treatment barriers (Table 3). The strongest barriers were believing one should be able to help themselves, believing the problem was not serious enough to warrant treatment, and not having time to seek treatment, followed by treatment being too expensive, feeling shame or embarrassment, and having a lack of trust in providers. Those factors deemed least likely to be a barrier included not believing an ED was a psychological problem, having a fear of being separated from family, and not having transportation.

Table 3.

Barriers to Seeking Treatment Reported by Participants Who Reported Not Seeking or Receiving Any Treatment for Eating Problems in the Past Nine Months (n = 29)

Barrier	Mean	SD
1. I felt shame or embarrassment	3.14	1.38
2. I have not known where to go to find help	2.54	1.17
3. I believed I should be able to help myself	3.93	1.30
4. I worried about being labeled or judged	2.68	1.44
5. I believed that my problem was not serious enough to warrant treatment	3.89	1.40
6. Treatment was too expensive	3.18	1.54
7. I have been unaware of the different treatment options available	2.79	1.26
8. I have had a lack of trust in providers	3.04	1.29
9. I have turned to other sources of support such as family and friends	2.50	1.32
10. I have had a lack of social support for seeking treatment	2.82	1.39
11. I worried that providers would lack expertise or not understand my issues	2.54	1.26
12. I have not believed that an eating disorder is a psychological problem	1.86	1.27
13. I have had a fear of being separated from my family	1.86	1.11
14. I have not had transportation to get to treatment	2.07	1.09
15. I have been concerned about stigma	2.68	1.42
16. I did not have time to seek treatment	3.61	1.45
17. I have been very happy with my weight	2.11	0.96
18. I do not believe that I have any problems with eating or weight	2.39	1.10

Note. All items rated on a 1 (strongly disagree) to 5 (strongly agree) scale.

Discussion

The purpose of this study was to conduct a 9-month follow-up of students identified with possible AN as part of the HBI Program offered at 36 U.S. universities. Results revealed several key findings. First, amongst this sample of follow-up completers, lowest past year BMI, weight concerns, excessive exercise frequency, and some indices of ED-related clinical impairment significantly improved over time, while other aspects of ED pathology did not (i.e., current BMI, binge eating, vomiting, diuretics/laxatives, fasting). Regarding psychiatric comorbidities, depression and anxiety significantly decreased over time, while sleep problems did not. On the whole, these data demonstrate the general intractability of AN. Further, the sample remained quite impaired at the 9-month follow-up; for example, participants reported extremely high weight and shape concerns and engaging in binge eating and vomiting on average nine times each over the past three months.

Second, the vast majority of participants (93.4%) remembered receiving the referral, suggesting value to offering screening for EDs and providing referral information, as at a minimum, participants recall this recommendation being made. The most common reactions to receiving the referral were feeling ashamed, nervous, validated, and sad. Only about one-third of participants reported being in treatment at the time they received the referral, and there were no differences in reactions between those who were and were not in treatment. However, this raises the issue of the impact of providing feedback on ED symptoms for those already engaged in treatment and whether this could be harmful (e.g., participant could feel less motivated to continue in treatment because screening feedback indicated they still have ED symptoms). Future work should consider modifying feedback for those already in treatment. Overall, the fact that only about one-third of individuals were in treatment at the time they received the referral highlights the great need for accessible, ongoing ED screening opportunities that provide participants with tailored feedback and clearly identify next steps for accessing treatment.

Third, among completers of the 9-month follow-up, as noted, about a third reported already being in treatment at the time they received the referral, 26% initiated treatment since that time, and 41% did not initiate treatment. It is possible that those who completed the follow-up may have been more likely to pursue treatment than those who did not—these numbers may thus represent an overestimation of the proportion of individuals originally identified with possible AN who sought treatment. A follow-up study of a 1996 college-based screening program found that of those who received a recommendation for further evaluation of disordered eating, nearly one-half kept at least a first appointment with a provider (Becker, Franko, Nussbaum, & Herzog, 2004). In contrast, Lipson et al. (2017) found much lower rates of help-seeking—in a random sample of university students identified with disordered eating and offered an online program, 18.1% enrolled. Notably, these studies did not exclusively study individuals with AN. In general, individuals who opted in to participation in HBI may have been more motivated to ultimately seek treatment than would have been the case should screening have been mandated or conducted randomly. In terms of what made participants decide to seek treatment, the most common reasons were emotional distress, concern with eating, and health concerns. Results suggest that HBI may have been successful in helping at least some students with possible AN move toward obtaining treatment, which is significant given the numerous negative consequences, including high costs, associated with delayed treatment, and highlight key messages to include in feedback messaging to increase the likelihood of an individual following up with the referral information provided (e.g., emphasizing health concerns). Indeed, when asked about things that could have been done to increase participants’ acceptance of the referral, the most common responses were providing a brief online program, providing more specific information on why the referral was made, and providing more information on what eating and body image concerns entail. This information can be used to create even more compelling referral messages in future efforts, and future research should consider offering online resources to individuals identified with possible AN to explore their motivation for treatment and potentially serve as a gateway to treatment. Some research has explored providing online programming for AN, with programs designed to prevent symptom progression in women at risk for AN (Ohlmer, Jacobi, & Taylor, 2013) or aimed at parents of adolescents with AN (Lock, Darcy, Fitzpatrick, Vierhile, & Sadeh-Sharvit, 2017).

Fourth, it is notable that change in all measures of ED pathology and psychiatric comorbidity, aside from excessive exercise, did not differ between those who were already in treatment for eating related problems at the time of receiving the referral, those who initiated treatment since that time, and those who did not, suggesting the highly impaired nature of this sample, the intractability of AN, and the need for early intervention. We also did not find differences in student status by treatment seeking, suggesting that the different value systems, responsibilities, and time commitments graduate students have may not be further motivating treatment seeking. Amongst those who did not access treatment, the strongest barriers were believing one should be able to help themselves, believing the problem was not serious enough, and not having time, followed by treatment being too expensive, feeling shame, and having a lack of trust in providers. A minority of participants endorsed that they were not able to obtain treatment because they did not have coverage for these services (13.1%) or had exceeded their allowed coverage (15.5%), highlighting the need for improved coverage for ED treatment. However, this was not a primary barrier, and the availability of health insurance does not necessarily promote greater use of mental health services (Olfson, Kessler, Bergland, & Lin, 1998). Future work should address the key treatment barriers for students with possible AN identified in this study, including providing psychoeducation that the problem is serious enough to warrant treatment and offering highly accessible treatment options.

Strengths include the use of a large, national sample comprising individuals from a wide variety of universities, increasing generalizability, and the relatively high percentage of non-Caucasian participants. However, findings should be interpreted in the context of limitations, including screening results being based on self-report questionnaires. Additionally, data on treatment receipt are based on the self-report data of participants who chose to complete the follow-up and may thus represent an overestimation of actual treatment receipt for participants who screened positive for AN in HBI. Further, we did not include validity checks to identify individuals who have been randomly or inconsistently responding. In future work, it would be useful to include such items to detect inattentiveness and enhance the validity of the data.

The ultimate goal of HBI is to reduce the incidence and prevalence of EDs on college campuses through increasing access to screening, prevention, and treatment. As such, it is imperative to understand the outcomes and health care utilization of students identified with possible AN to inform future work and improve treatment engagement and ultimately outcomes. Findings highlight the high pathology in students identified with possible AN, even nine months after they were first identified and provided with resources, and the relatively low rates of treatment utilization given the seriousness of these illnesses. Future work should consider mandating ED screening on college campuses to increase identification of individuals with AN and offering online resources, including programs focused on increasing motivation for care, to individuals identified with possible AN to increase likelihood of treatment follow through.

Clinical Implications.

• We followed up with students with anorexia nervosa 9 months after identification

• On the whole, participants remained quite impaired at follow-up

• Given the seriousness of anorexia nervosa, treatment utilization rates were low

• Future work should consider mandating eating disorder screening on college campuses

• Need to explore options for increasing follow through with treatment from screening