The Evolving Needs of Children Hospitalized for Eating Disorders During the COVID-19 Pandemic

Our institution is a part of a large health system in the Northeastern United States. This study was conducted at the tertiary academic medical center which is the referral center for patients with EDs from the surrounding community. At the academic medical center site, patients with EDs who meet admission criteria as outlined in a clinical pathway¹²  are admitted to the pediatric hospitalist service and managed in collaboration with child psychiatrists, dieticians, social workers, and child life services. Patients are medically managed on a strict protocol with slow introduction of calories and monitoring for refeeding syndrome. The child psychiatry team sees patients once a day on weekdays. Weekly multidisciplinary family meetings are conducted to provide updates and discuss disposition planning. Disposition is determined by the child psychiatry team and social workers based on patient presentation, treatment history, availability of local resources, insurance, ability to engage in family-based therapy, and patient and family preference. Patients are regularly referred to 9 programs throughout the Northeast with options for in-person and virtual care. The majority of programs serve patients who are at minimum 12 years old. There are 3 partial hospitalization programs (PHP) in-state. There is only 1 in-state residential program for privately insured patients and no in-state residential programs for Medicaid patients.

Approval for chart review was obtained from the Institutional Human Research Protection Program before study initiation. Our data analytics team provided a list of patients admitted between January 1, 2017 and June 30, 2021 for ED management with at least 1 of the following diagnosis: F50 - eating disorders, F50.00 - anorexia nervosa unspecified, F50.01 - anorexia nervosa restricting type, F50.02 - anorexia nervosa binge eating-purging type, F50.2 - bulimia nervosa, F50.8 - other eating disorders, F50.81 - binge eating disorder, F50.82 - avoidant-restrictive food intake disorder, F50.89 - other specified eating disorder, or F50.9 - eating disorder unspecified. Our child psychiatry team, which was led by the same providers throughout the study period, cross-checked the initial list and added additional patients who were admitted for other ED-related diagnoses including bradycardia and intentional weight loss or were found to have an ED requiring inpatient management during a hospitalization. Two researchers (MS, JL) conducted manual reviews of each chart in the electronic health record. We defined the prepandemic study period as January 1, 2017 to February 28, 2020, and the during-pandemic study period as March 1, 2020 to June 30, 2021.

Patient demographics at first admission, characteristics of hospital course, and disposition at discharge were summarized using descriptive statistics. Differences between the pre- and during-pandemic study periods were calculated using χ² test or Fisher’s exact test, as appropriate for expected cell counts of <5. Average monthly admission rate pre- and during-pandemic was compared using Poisson regression. To examine the effect of repeat admissions, we separated our analyses of LOS to first admissions and all admissions. LOS for first admissions and all admissions was visually summarized using boxplots, stratified by study period. LOS was modeled using Poisson regressions with robust standard errors for all admissions to account for the effect of repeated admissions on the standard errors.¹³  To decompose the cross-sectional and longitudinal effects of patient age, we parametrized time as the month since the start of the pandemic (defined as March 1, 2020) and patient’s age at admission. To examine whether changes in LOS were associated with patient age, we compared LOS pre- and during the pandemic as stratified by age at first admission and all admissions. Age at admission was characterized as <13 years old, 13 to 17 years old, and >18 years. LOS were summarized as means with 95% confidence intervals (CI). Data were analyzed in SPSS v28.¹⁴  Hypothesis tests were conducted at the 2-sided α level of 0.05.

There were 85 patients who required inpatient admission from January 1, 2017, to June 30, 2021. As shown in Table 1, there were no significant differences in sex, race, insurance status, or family composition between the 2 time periods. The most common comorbidities were anxiety and depression with no significant differences in rates of anxiety (33% vs 33%, P = .5) and depression (30% vs 31%, P = .8) pre- versus during the pandemic. Most patients were admitted only once to our institution during the study period.

There were 108 encounters for inpatient admission during the study period. There was an increase in the rate of admissions per month pre- versus during the pandemic (1.4 per month vs 3.6 per month, P < .001). There were no significant differences in admission parameters (ideal body weight, BMI, and orthostatic vital signs) except for a heart rate <45. Electrocardiogram abnormalities, need for electrolyte repletion, and nasogastric tube feeding pre- versus during the pandemic were not significantly different (Table 2). In contrast, psychotropic medications were changed or initiated more often during the pandemic as compared to before the pandemic (11% vs 31%, P = .01).

The mean, median, and range for the LOS for first admissions and for all admissions prepandemic and during the pandemic are shown in Fig 1. There was an increase in the mean LOS for first admissions (11.5 days vs 16.6 days) and for all admissions during the pandemic (12.6 vs 18.0 days). Controlling for age on admission, we found that longer LOS was associated with younger age during the pandemic for first admissions and for all admissions (P < .001). Comparing LOS pre- versus during the pandemic, longer LOS at first admission was observed for children < age 13 years (11.6 days [95% CI 9.5–14.2] vs 16.1 days [95% CI 13.4–19.4]) and for children 13 to 17 years (11.7 days [95% CI 10.5–13.1] vs 17.4 days [95% CI 16.1–18.9]) but not for patients >18 years (9.3 days [95% CI 6.7–12.8] vs 10.8 days [95% CI 7.9–14.5]). This was confirmed for all admissions.

As shown in Fig 2, during the pandemic, fewer patients were able to be discharged from the hospital (43% vs 24%, P = .03). A larger proportion of patients required higher levels of care including partial hospitalization program, residential care, or transfer to an inpatient psychiatric unit.

We found an increase in the number of admissions for eating disorders during the COVID-19 pandemic, longer lengths of hospital stay associated with younger children, an increase in the number of patients requiring psychotropic medication management, and a marked shift in discharge disposition away from outpatient management toward higher levels of care. Although the increase in concerns for disordered eating during the pandemic has been widely discussed^6,15,16  and increases in admissions documented in other children’s hospitals,^7,8  our study highlights the changing needs of children and adolescents during hospitalization for eating disorder treatment and the added strain on inpatient settings during the pandemic.

The increase in the rate of hospitalizations and an average increase in LOS of 5 days may have a considerable impact on already strained health care resources. Longer lengths of stay were often because of lack of bed availability in PHP or residential eating disorder programs. Inpatient care of EDs is time-intensive¹²  and requires ongoing interdisciplinary collaboration among providers in the hospital and the community. Within the hospital setting, this includes pediatric hospitalists, mental health providers, dieticians, social workers, and child life providers all working together daily to provide individualized and high-quality care for the patients.⁵  In our institution, in the absence of inpatient adolescent medicine clinicians, eating disorder specialists, or family-based therapy, once ED patients are medically cleared, patients remain on the inpatient unit on the hospitalist service with minimal treatment of their underlying ED while awaiting disposition. The longer patients remain without ED-specific treatment, the more protracted their presentations may become,^17,18  increasing concerns for worsening manifestations of their ED or delayed recovery after discharge.

We found increased lengths of stay for younger children (<13 years of age) during the pandemic. In our experience, ED programs often accept and provide evidence-based treatments for adolescents that have been based on developmental stage. Younger children and adolescents differ in presentation of EDs, with research suggesting that younger ED patients present at a lower percentage of ideal body weight and lose weight more rapidly.^19,20  Finding an age appropriate eating disorder program for children <13 years of age is an additional challenge. The differences between young children and adolescents have important diagnostic and prognostic implications that should be taken into consideration during disposition planning. The reasons for younger age at hospital admission may represent worsening or earlier manifestations of EDs during the pandemic and merit further study.

We noted an increase in the percent of admissions in which psychotropic medications were initiated or changed during the pandemic period. There is limited evidence to support the use of psychotropic pharmacotherapy (serotonin reuptake inhibitors [SSRIs] or atypical antipsychotics) for children with AN and subthreshold EDs.²¹  SSRIs are often prescribed to address comorbid diagnoses of depression or obsessive-compulsive disorder, particularly if the depression or obsessive-compulsive disorder predates the diagnosis of the ED. Despite the limited literature thus far in the field, clinically, when thought rigidity is so severe that it interferes with treatment progress and hospital care goals, our institution’s child psychiatry team does recommend the initiation of an atypical antipsychotic after discussion with the patient and their family. Several factors may have contributed to the increase in psychotropic medication management seen during the pandemic in our study. Although prescribing practices may vary by provider, our core child psychiatry team, consisting of 2 main clinicians, has remained consistent throughout the study period. Nonetheless, the providers may have gained experience and comfort with prescribing psychotropic medications for EDs over time. Although we did not find a significant increase in rates of anxiety and depression during the pandemic, patients may have had worsening or increasingly complex presentations of existing diagnoses that required changes to their psychotropic management. Furthermore, the child psychiatry team typically coordinates with outpatient providers to initiate and monitor psychotropic medications. However, the increase in LOS during the pandemic, in tandem with limited outpatient provider availability, may have allowed patients to reach a nutritional status at which psychotropic medications, particularly SSRIs, may be more effective and therefore could have been started and monitored for side effects before discharge from the hospital.

We found that during the pandemic, fewer patients were determined to be clinically ready to return home for outpatient treatment, instead requiring referrals to PHPs or other higher levels of care. The effects of a perceived higher acuity at discharge and longer LOS may be interdependent. On one hand, the concomitant increase in patient numbers, higher acuity at discharge, and decrease in bed availability at treatment facilities during the pandemic led to longer inpatient stays. Conversely, although medical parameters were similar pre- and during the pandemic, longer LOS may have influenced providers’ clinical impressions of patient acuity. Regardless, these disposition challenges raise concerns about the impact of delayed ED-specific care on recovery course and future presentations as well as the impact on inpatient hospital bed capacity.

An ongoing intervention to combat limited availability in ED treatment facilities is the inclusion of “bridging services” in disposition planning. “Bridging services” refer to the use of outpatient resources to monitor patients while awaiting bed availability at ED treatment facilities and may include in-person or virtual monitoring by primary care providers, mental health providers with or without ED specialization, and outpatient dieticians. During the pandemic, our providers tried to use “bridging services” when a PHP level of care was clinically indicated but unavailable for several weeks. However, success was limited because of the lack of outpatient multidisciplinary resources, likely exacerbated by the pandemic. Centralized databases of community providers able to provide “bridging services” and alternate methods of outpatient ED-focused care, such as virtual visits, may improve the care of patients awaiting intake at a treatment facility. The impact of delayed admissions to PHPs on clinical outcomes and readmission rates for ED patients remains unknown and should be monitored closely.

With respect to limitations, we had a relatively small sample size, and our study was conducted at a single center limiting generalizability to other hospitals with different patient demographics and resources. The availability and wait times for PHP and residential facility beds in our state may be different than other states. We may have underestimated the number of hospital admissions, because patients may have been readmitted at other institutions. We analyzed data for patients who were admitted for inpatient management of an ED; therefore, patients with presentations to the emergency department that did not result in hospital admission or with admissions primarily for non-ED diagnoses were excluded from this study. Lastly, our study relied on information available in the electronic health record, which may have contained incomplete, absent, or incorrect documentation.

The unrelenting COVID-19 pandemic has contributed to a behavioral health crisis in pediatrics, exacerbated by competing demands on health care resources. Children with eating disorders are among those who have been adversely affected by the pandemic. At our institution, an increase in the volume of ED admissions and a greater number of patients unable to be safely discharged with outpatient services resulted in significant disposition challenges and longer hospital stays. Ongoing interventions that may improve the care of this vulnerable population include the generation of databases of community providers who can safely provide “bridging services” while awaiting higher levels of care and the expansion of models of ED-focused virtual visits. In a time of unprecedented strain on healthcare systems, pediatric hospitalists have an opportunity to prepare for future crises by recognizing the evolving needs of children with EDs and the associated implications on inpatient management and disposition planning.