Access to Care and Worsening Eating Disorder Symptomatology during the COVID-19 Pandemic

Shelter-in-place orders and social distancing guidelines, in response to the COVID-19 pandemic, have limited traditional face-to-face interactions and led to many clinical providers transitioning to the use of videoconferencing platforms. The present study aims to assess how the COVID-19 pandemic has impacted adolescents’/young adults’ (AYA) eating disorder (ED)-related care, and how access to, changes in, perceived disruptions to, and quality of care are associated with ED thoughts and behaviors. Methods We examined their access to treatment, in care, quality of care since the pandemic started. Our results demonstrate that patients with established care have maintained treatment and perceive their care as high quality, though the are experiencing worsening ED thoughts and behaviors. Patients who perceive their quality of as than usual might likely have intrusive ED thoughts. Continued access to care could also be protective against increased ED behaviors. Ultimately, our study highlights the for continued of during this challenging time. This study examined the association between markers of care and ED thoughts and behaviors in youth, ve months after Massachusetts enacted stay-at-home orders due to the COVID-19 pandemic. Our participants reported marked increases in both intrusive ED thoughts and behaviors, which they attributed to the COVID-19 pandemic. Reassuringly, the majority of our cohort was able to maintain access to at least one member of their care team via telehealth and some were able to continue to see their providers in person. However, half had to stop some aspect of their treatment because of the pandemic. Despite these changes, patient-reported quality of care did not suffer, and most patients did not perceive a disruption to their care. Those who perceived a treatment disruption to care were more likely to rate their care lower than those who did not perceive a disruption. None of the markers of care described was statistically associated with increased ED thoughts and behaviors, though our results may indicate that patients who perceive their quality of care as worse than usual are more likely to have intrusive ED thoughts.

have transitioned to using videoconferencing platforms. Clinicians who care for patients with EDs worry that these sudden changes in accessing treatment, on top of mental health challenges associated with the pandemic, may contribute to worsening ED symptoms. In this study, we asked adolescent and young adult patients with EDs about their symptoms, access to treatment, changes in care, disruptions in treatment and quality of care since the pandemic started. Our results demonstrate that patients with established care teams have maintained treatment and perceive their care as high quality, though the majority are experiencing worsening ED thoughts and behaviors. Patients who perceive their quality of care as worse than usual might be more likely to have intrusive ED thoughts. Continued access to care could also be protective against increased ED behaviors. Ultimately, our study highlights the need for continued support of patients during this challenging time.

Background
Since the beginning of 2020, the COVID-19 pandemic has had a profound impact on youth around the world. Though these effects have been universal, individuals with pre-existing psychiatric illnesses are especially vulnerable to the consequences of COVID-19. 1 Individuals with eating disorders (ED) are at particular risk, as stress and anxiety may lead to worsening eating disordered cognitions, which may further lead to negative behaviors and detrimental physical effects. 2 Not surprisingly, individuals with EDs have reported increased social isolation, rumination about eating, feelings of anxiety and depression, and decreased feelings of control and social support during the COVID-19 pandemic. [3][4][5] Clinicians working with patients with EDs have also voiced concern that the changes caused by COVID-19 may increase ED symptoms, decrease protective factors, and exacerbate barriers to care. 6 The standard of care for ED treatment involves a multi-disciplinary team to address the medical, psychological, and nutritional components of this serious, life-threatening psychiatric illness. 7 Shelter-inplace orders and social distancing guidelines have limited face-to-face interactions, leading to residential and day program closures and limiting in-person outpatient care from all disciplines. Thus, for most patients there have been reductions in in-person contact with clinical teams. 8 Recognizing the need to limit in-person clinical care, governing bodies responded by changing national and state payment policies and lifting technology restrictions, encouraging clinicians to maintain treatment continuity. 9 In response, providers across disciplines have rapidly transitioned their practices to videoconferencing platforms, commonly referred to as telehealth. 10,11 ED clinicians have published recommendations on how to adapt multi-disciplinary ED care to telehealth, 12-14 however, little is known about how patients perceive this change in their care as a result of the COVID-19 pandemic.
Experts worry that these sudden changes and limited access to in-person care, coupled with increasing mental health challenges associated with the pandemic, may have serious implications for patients with EDs. 3 The absence of in-person care creates speci c challenges for ED treatment, namely, monitoring for changes in weight and vital signs, essential markers of illness. 14,15 Additionally, some believe that the therapeutic alliance, necessary for successful clinical care, could be disrupted by the transition to telehealth. 16 Though many raise concerns that reduction in in-person access to clinical providers may contribute to the worsening of ED symptoms, little research currently exists. 8,12,13,17 Thus, we set out to do the following: 1) examine four markers of care during the time of the COVID-19 pandemic: reported access to (telehealth and/or in-person), changes to, perceived disruption of, and reported quality of EDrelated care; and 2) examine whether these four markers of care are associated with increased ED thoughts and behaviors. We hypothesized that decreased access to, changes in, perceived lower quality of, as well as perceived disruption to care would be associated with increased ED thoughts and behaviors.

Study Sample
We used data from the Registry of Eating Disorders and their Co-Morbidities Over Time in Youth (RECOVERY), a longitudinal registry of adolescent/young adult patients ages 10-27 (average age, 17.1 years at enrollment) seeking ED treatment. RECOVERY study participants were previously recruited from the outpatient ED program at Boston Children's Hospital from June 2017 to February 2020.
Participants provided written consent and assent prior to enrollment. Using web-based surveys sent via Research Electronic Data Capture (REDCap a HIPAA compliant database) patients, their parents, and their clinicians answered questions regarding ED behaviors and treatment every three months for the rst year and every six months thereafter. Participants received remuneration for each completed survey.
In July 2020, in response to the COVID-19 pandemic, we invited all patient participants to complete an added survey with the aim of understanding the impact of the COVID-19 pandemic on ED treatment, behaviors, and general well-being. The COVID-19 survey was separate from the regularly scheduled RECOVERY surveys, and participants were informed that additional remuneration was not available. The RECOVERY study and the additional COVID-19 survey were approved by the Boston Children's Hospital Institutional Review Board.

Survey Measures
Sample demographic characteristics were obtained from the participants' baseline RECOVERY surveys. Information regarding treatment access, perceived disruption of, reported changes to, and perceived quality of care, as well as the effects of the COVID-19 pandemic on ED thoughts and behaviors, were obtained from the COVID-19 Survey. This survey was adapted from one developed at the University of North Carolina Center of Excellence for Eating Disorders. 3 Primary predictor variables Access to Care: Participants were asked about their access to multi-disciplinary ED care since the pandemic via the question, "I have been able to access my providers…". Participants could then check all that applied: "via telehealth," "in person," "neither (I have not been able to access my providers at all)." Responses were then dichotomized to indicate any access to care (via telehealth or in-person) vs. no access.

Changes to Care
Participants were asked about changes in different elements of their care in a series of two questions.
First, respondents were asked to report what care they were receiving prior to the pandemic (i.e., "Before the pandemic, I was involved in…"). Participants could then check all that applied from weight checks, nutrition, or therapy appointments. Participants were then asked to report on the care they have received since the pandemic (i.e., "Since the pandemic, I have been involved in…"), and were asked to check all that applied from the same list of potential appointments. We then compared the pre-and post-pandemic reported care and created a variable indicating any stopped care (if any of the three -weight checks, therapy, or nutrition-changed from pre-to post-pandemic) v. no change to care.
Quality of Care: Patients were asked to rate the overall quality of their care they received during the rst few months of the pandemic, compared to care received prior to the pandemic, using a 4-point Likert scale: "better than usual", "as good as usual", "somewhat worse than usual," and "worse than usual".

Perceived Treatment Disruption
Patients were asked to report whether their overall ED treatment had been disrupted as a result of the COVID-19 pandemic with response options of yes, no, not applicable.

Primary outcomes Intrusive ED thoughts and behaviors
Participants were asked, "how has the COVID-19 pandemic affected intrusive eating disorder thoughts" using a 5-point Likert scale, ranging from "increased signi cantly" to "decreased signi cantly." Responses were dichotomized to indicate no increase vs. any increase in ED thoughts. Respondents also reported in three separate questions whether they had engaged in more restrictive behaviors, compensatory behaviors, or binging behaviors in the past three months "because of COVID-19 related factors." Answers to these three questions were on a 4-point Likert scale ranging from "not at all" to "daily or more." Responses from the three questions were combined and then dichotomized to indicate frequently or daily vs. never or rarely for any of the three ED behaviors.

Additional variables
Age Age at time of COVID-19 survey completion was calculated from the date of survey completion and the date of birth obtained from the RECOVERY study baseline survey.

Sex
Self-reported sex assigned at birth (female, male or another sex) was obtained at the time of RECOVERY study baseline survey completion.
Race/Ethnicity: Patients were asked to select all that applied from the following options on the RECOVERY baseline survey: Hispanic/Non-Hispanic, American Indian or Alaska Native, Asian, Black or African American, Middle Eastern/North African, Native Hawaiian or other Paci c Islander, White/Caucasian or another race. We constructed a mutually exclusive race/ethnicity variable consisting of non-Hispanic white, non-Hispanic Black or African-American, Asian, Multiracial, Other race, and Hispanic.
ED Diagnosis: Patients were asked to select all that applied from the following options on the RECOVERY baseline survey: anorexia nervosa (AN), atypical anorexia nervosa (AAN), Avoidant Restrictive Food Intake Disorder (ARFID), bulimia nervosa (BN), binge-eating disorder (BED), purging disorder, other eating issue(s)/disorder(s), and I don't know/Unsure. AN and AAN were collapsed to create an indicator for restrictive diagnoses vs. other ED diagnoses.

Length of treatment
This variable was calculated from the date of patient's rst ED clinic appointment to the date of COVID-19 survey completion.

Statistical Analysis
We examined frequencies (percent) for categorical variables and means (standard deviation) for continuous variables. We compared responders to the COVID-19 survey to non-responders from the RECOVERY cohort on demographic factors (age, race/ethnicity and sex assigned at birth) and ED diagnosis using t-tests for continuous variables and tests for categorical variables. Multivariable logistic regression was used to examine the associations of access to, changes in, quality of, and perceived disruption to ED care during the COVID-19 pandemic with changes in ED thoughts and behaviors. Regression analyses were adjusted for age, sex at birth, race/ethnicity, ED diagnosis and length in treatment. All analyses were conducted using SAS (v9.4; Cary, NC).

Sample Characteristics
Of the 161 participants enrolled in the RECOVERY study, eighty-nine (55%) participants responded to the COVID-19 survey. Respondents to the survey did not differ from non-respondents on age at enrollment, χ 2 race/ethnicity, sex assigned at birth. Participants with restrictive ED diagnoses (AN and AAN) were more likely to respond to the survey (p = 0.03) compared to those with non-restrictive diagnoses (BED, BN, ARFID, purging disorder, other eating issue(s)/disorder(s), or I don't know/Unsure).
Of the eighty-nine respondents, sixteen (18%) were excluded from further analysis because they were not engaged in treatment prior to the pandemic. Of these remaining 73 participants, the mean age was 19.1 ± 3.0 years ( Table 1). The majority were female (93%), White/non-Hispanic (79%) and living with a restrictive ED diagnosis (85%). Approximately half (53%) had been in treatment for two or more years. Access to care remained high, with 92% of respondents reporting continued access to at least one provider via telehealth or in person ( Table 2). Of those continuing in care, 88% used telehealth to see at least one provider. However, nearly half (47%) of the 73 participants who were actively engaged in ED care prior to the pandemic reported stopping at least one aspect of their ED treatment: sixteen (22%) stopped mental health counseling/therapy, seven (10%) stopped nutrition visits, and twenty-three (32%) stopped weight checks with their medical provider. Approximately one-third (32%) perceived a disruption in treatment. Of those with telehealth access (n = 64), 9% found care to be better than usual, 59% as good as usual, while 30% said somewhat worse than usual, and 2% much worse. There was no association between access to care via telehealth and perceived disruption of care (p = 0.99) or quality of care (p = 0.36), however respondents who perceived a disruption to their treatment were more likely to report lower quality of care (p = 0.004).

ED Thoughts and Behaviors
Many respondents (81%) endorsed increased intrusive ED thoughts as a result of the COVID-19 pandemic. The same percentage of participants (though not necessarily the same patients) reported an increase in ED behaviors: engaging in restrictive or compensatory behaviors or binging on food at least once in the past three months because of COVID-19-related factors. Thirty-three (45%) participants reported engaging in the restrictive/compensatory/or binging behaviors frequently or daily.
The association of access to, changes in, perceived quality of and perceived disruption to care with ED Thoughts and Behaviors Unadjusted and adjusted associations between access to, changes in, perceived quality of and perceived disruption to care, with ED thoughts and behaviors are presented in Table 3 Those who perceived quality of care as worse than usual were more likely to have intrusive ED thoughts (aOR = 8.00; 95% CI: 0.85-75.7), though these associations were approaching signi cance (p = 0.07). Table 3. Unadjusted and adjusted odds of ED thoughts or behaviors associated with markers of care (N=73).
Model adjusts for age, sex at birth, race/ethnicity, restrictive diagnosis and length of treatment. Abbreviations: OR, odds ratio; CI, con dence interval a Separate models predicting outcome from each predictor. b Intrusive ED thoughts increased vs. no change/decreased. c Composite measure for engaging in restrictive or compensatory behaviors or binging on food frequently or daily vs. never or rarely in the past 3 months. d Models adjusting for age, sex at birth, race/ethnicity, restrictive diagnosis and length of treatment Discussion This study examined the association between markers of care and ED thoughts and behaviors in youth, ve months after Massachusetts enacted stay-at-home orders due to the COVID-19 pandemic. Our participants reported marked increases in both intrusive ED thoughts and behaviors, which they attributed to the COVID-19 pandemic. Reassuringly, the majority of our cohort was able to maintain access to at least one member of their care team via telehealth and some were able to continue to see their providers in person. However, half had to stop some aspect of their treatment because of the pandemic. Despite these changes, patient-reported quality of care did not suffer, and most patients did not perceive a disruption to their care. Those who perceived a treatment disruption to care were more likely to rate their care lower than those who did not perceive a disruption. None of the markers of care described was statistically associated with increased ED thoughts and behaviors, though our results may indicate that patients who perceive their quality of care as worse than usual are more likely to have intrusive ED thoughts.
Like many institutions across the globe, 11,18−20 our clinicians transitioned their practices to videoconferencing platforms shortly after stay-at-home advisories were issued in March 2020. 23 Despite providers' concerns about patients' perceptions of adapting evidence-based ED treatment to telehealth, [12][13][14]21 the majority of our patients who accessed their care via telehealth felt their quality of care was as good as, or better, than usual. Previous research has shown telehealth to be effective for use in psychotherapy, 22,23 as well as effective in adolescent/young adult populations, 24-26 however little research exists examining the use of this technology for multi-disciplinary ED treatment. 27,28 Although more research is necessary to determine e cacy, 4,29 our ndings suggest providing ED care via telehealth is well accepted and provides good quality of care for ED patients. Moreover, the expansion and continued use of telehealth for multi-disciplinary ED treatment could mitigate barriers to care and increase access for patients who are less connected or away for college or employment, even after the restrictions of the pandemic are lifted. 30 Similar to other studies, our patients with EDs reported increases in ED thoughts and behaviors related to the COVID-19 pandemic. 3,5,31,32 One study attributed an increase in ED thoughts during the COVID-19 pandemic to the lack of in-person care and absence of distractions. 33 Our ndings suggest that continued access to care may be associated with more intrusive ED thoughts, but also could have a protective effect on actual ED behaviors. Although our multivariate analyses were not statistically signi cant, the trends noted could have clinical implications. Treatment may serve as a protective factor, preventing an increase in ED behaviors. This nding is comparable to another study by Schlegl and colleagues, who found an increase in ED cognitions, but not an increase in behaviors during the initial months of the COVID-19 pandemic. 5 It is critical to understand how to support our patients so that they may remain resilient against their ED behaviors, particularly during these times of stress and isolation, which are known risk factors for patients with EDs. [33][34][35][36][37] Based on our results, future research should continue to explore how access to care may mitigate ED symptomatology.
One limitation of the present study was the total number of respondents, which decreased the power to address our primary question.
Post-hoc power calculations coupled with wide con dence intervals suggest that we were underpowered to detect statistically signi cant associations between our markers of care and ED symptomatology, due to the high prevalence of the symptomatology and moderately small sample size. Given the effect size we observed for the association between access to care and intrusive ED thoughts, we had 66% power to detect a statistically signi cant difference. We observed signi cant associations between our quality of care measure and intrusive ED thoughts in bivariate analysis, though this effect was attenuated after adjustment for other factors. It is likely our ndings may be a conservative estimate of the true association: those who perceive quality of care as worse than usual are more likely to have intrusive ED thoughts. An additional limitation is the response rate; only half of participants in the RECOVERY study responded to the COVID-19 survey. There is a possibility that those who did not respond were affected by COVID-19 differently than those who did respond, which could affect the results of this study. Notably, this survey was sent out of sequence and remuneration was not available for completion of this survey, which may explain some loss of respondents. Overall, there were no differences between survey respondents and non-respondents in age, race/ethnicity, sex, but those with restrictive EDs were more likely to respond to the survey.
An additional limitation is the generalizability of our ndings. The majority of our study participants identi ed as white and were diagnosed with restrictive EDs. Future research should examine the differences in markers of care by ED diagnosis and demographic groups, particularly with respect to race/ethnicity and socioeconomic status. Additionally, our participants may be more representative of individuals who were already engaged in care and connected to their health care team, as most participants were in treatment and the majority had been in treatment for more than two years. All the participants in this cohort had providers with ED expertise at some point, making it an easier transition to COVID-19 modi ed treatment. Published literature has reported increased ED incidence as a result of the pandemic, 38 which is consistent with our clinical experience during this time. Anecdotally, we have noted that it is especially di cult for these new patients to nd providers. Thus, our ndings may not be generalizable to patients with new onset EDs and therefore it is critical that we examine the effect of COVID-19 on the incidence of EDs and access to care for new patients.

Conclusions
The present study demonstrates the profound impact COVID-19 has had on patients' ED symptomatology and access to treatment. Our ndings suggest that patients with established care teams prior to the pandemic were able to maintain some aspect of their care via telehealth and perceived it to be high quality. Our patients with EDs reported marked increases in intrusive ED thoughts and behaviors because of the pandemic. Disruption in, decreased quality of, changes to and decreased access to care were all associated with increased ED symptomatology, though these associations were not signi cant likely due to lack of statistical power. More work is needed to better understand the relationship between care changes and ED symptomatology in larger populations. In the meantime, providers should continue to support patients via telehealth, and in person when possible, who are at risk of worsening EDs during this di cult time. The datasets generated and/or analyzed during the current study are not publicly available due protection of patient privacy but are available de-identi ed from the corresponding author on reasonable request and IRB approval.