Pediatric Mental Health Presentations and Boarding: First Year of the COVID-19 Pandemic

The study was conducted in a United States Northeast state where a COVID-19 state of emergency was declared in March 2020 and lifted in June 2021. A statewide lockdown that included school, religious, workplace closures, limits on in-person gatherings, and other restrictions, began in March 2020. A gradual, 4-phased reopening was initiated in May 2020, continuing beyond the first 12 months of the pandemic. The state experienced the first virus surge from April to June 2020, and a second, larger virus surge (with higher volumes of reported cases) from October 2020 to February 2021. All public and private elementary and secondary schools closed in-person instruction in March 2020, and full in-person schedules did not return for the next 12 months.

The study institution is a large, urban, freestanding children’s hospital. It has an emergency psychiatry service that sees patients with mental health presentations in the ED, and a separate psychiatry consultation service that sees patients on inpatient medical/surgical units. Both services consist of multidisciplinary teams of psychiatrists, psychologists, social workers, and a psychiatric nurse practitioner, with consultation available to the ED and inpatient units 24 hours a day, 7 days a week. The services use a standardized psychiatric evaluation template, and diagnoses are made using Diagnostic and Statistical Manual of Mental Disorders Fifth Edition (DSM-5) criteria. DSM-5 diagnoses and corresponding International Classification of Diseases, 10th Revision (ICD-10) codes¹⁵  are documented in the electronic medical record. After the psychiatric evaluation is completed and, where indicated, the patient is medically stabilized, a resource specialist team is consulted to assist with program referrals. These may include higher levels of psychiatric care at inpatient, acute residential, partial hospital/day treatment programs, or intensive outpatient or home-based services.

Boarding begins once a medically stable patient has received a psychiatric evaluation and recommendation for referral to a psychiatric treatment program. The resource specialist referral includes a completed form with a date stamp in the electronic medical record corresponding to the date of boarding onset. Some patients require medical stabilization before referral to psychiatric treatment programs (eg, patients with acute eating disorder symptoms or severe suicide attempts). Such patients are sometimes admitted directly to the inpatient medical units, bypassing the ED, and their boarding begins after medical stabilization and psychiatric evaluation on the inpatient units. For patients whose boarding starts in the ED, the institution has a practice of admitting them to inpatient medical/surgical units to continue boarding after several days of boarding and unsuccessful placement efforts in the ED.

We performed a retrospective chart review of all youth aged ≤18 years with ED or inpatient medical/surgical admissions from March 2019 to February 2021 related to mental health complaints. We studied identical 12 month calendar periods before and during the pandemic to capture similar seasonal variabilities. The hospital’s institutional review board approved the study, with a waiver of informed consent. The institution maintains a quality improvement dashboard, which includes demographic information and presenting problems, to track patients presenting to the hospital with mental health-related concerns, using data from the hospital’s electronic data warehouse. Both ED and inpatient medical/surgical admissions were included in the sample obtained from the quality improvement dashboard to capture all mental health-related presentations to the hospital. For patients admitted to inpatient medical/surgical units after an ED presentation, the entire encounter counted as 1 admission. We defined length of admission as the total amount of time spent in the ED and/or inpatient medical/surgical unit, including time spent on medical stabilization and time boarding. We defined length of boarding as time spent in the ED and/or an inpatient medical/surgical unit once medically cleared and awaiting placement at a psychiatric treatment program. Length of boarding was calculated in days from date of resource specialist referral (corresponding to boarding onset) to date of discharge from the hospital (ED or medical/surgical floor). Demographic (age, sex, insurance type), psychiatric diagnoses, suicide ideation/attempts, and admission (length of boarding, length of admission) characteristics were abstracted from the medical record. Psychiatric diagnoses (with corresponding ICD-10 codes) were categorized according to the DSM-5 classification system and suicide ideation/attempts were categorized using ICD-10 codes (Supplemental Table 2).

We compared the characteristics of pediatric mental health presentations during the 12 months before the COVID-19 pandemic (March 2019–February 2020) and the first 12 months of the pandemic (March 2020–February 2021). Inferential statistics included χ² and t tests, and level of significance was P = .05 for a 2-tailed test. Benjamini-Hochberg correction was conducted to adjust for experimenterwise error, with a false discovery rate equal to the level of significance (ie, 5%).

Interrupted time series analyses were conducted to evaluate longitudinal trends of specific outcomes of interest (psychiatric diagnoses, suicide ideation/attempts, length of boarding) throughout the 12-month period before and during the pandemic and to compare changes in the trend levels and slopes of these outcomes during both years.

During the prepandemic period, the expectation was that, for most psychiatric diagnoses except depressive disorders, there will be nonsignificant or ∼ 0 slopes showing some stability and equilibrium. For depression, suicidal ideation, and attempts, the expectation was that there would be positive slopes reflective of increasing trends before the pandemic. During the pandemic period, the expectation was that there would be positive slopes for suicide ideation/attempts, most psychiatric diagnoses, and length of boarding reflective of the mental health impact of the pandemic. We expected a decline in attention-deficit/hyperactivity disorder (ADHD) diagnoses during the pandemic period because of school closures.

There were 3799 pediatric mental health admissions to the ED and inpatient medical/surgical units during the 2-year study period. The most common diagnoses in both years included: depressive disorders (66.7%), anxiety disorders (48.1%), ADHD (30.6%), disruptive/impulse control/conduct disorders (29.5%), trauma and stressor-related disorders (23.3%), autism spectrum and other neurodevelopmental disorders (11.8%), eating disorders (10.1%), and substance-related disorders (7.8%).

Table 1 summarizes the differences in demographic and clinical presentations in both years. Total number of presentations decreased during the pandemic year, particularly during the lockdown period. Findings describe proportion changes in both years.

Compared with the prepandemic year, during the pandemic, admissions for males declined (44.0% vs 34.2%) and females increased (56.0% vs 65.8%, P < .001), admissions for children declined (22.0% vs 16.2%) and adolescents increased (78.0% vs 83.8%, P < .001) and there was a decrease in patients with public insurance (46.0% vs 41.7%, P = .008). Length of admission (2.5 vs 5.5 days, P < .001) and length of boarding (2.1 vs 4.6 days, P < .001) more than doubled during the pandemic year. Of all the pediatric patients who presented with mental health-related complaints during the pandemic, 71.5% (n = 1272) boarded in the ED and/or inpatient units for ≥1 day and 50.4% (n = 896) experienced extended boarding periods of ≥2 days awaiting placement, compared with 56.9% (n = 1150) and 30.2% (n = 611), respectively, during the prepandemic year.

Admissions related to suicidality increased during the pandemic year compared with the previous year, and the majority of pandemic pediatric mental health presentations were related to suicidal ideation and/or suicide attempts (49.7% vs 60.3%, P < .001). Compared with the previous year, admissions for depressive (63.4% vs 70.4%, P < .001) and anxiety disorders (46.0% vs 50.5% P = .006) increased and admissions for eating disorders doubled (6.9% vs 13.8%, P < .001), whereas substance-related disorders (6.7% vs 9.2%, P = .005) and obsessive-compulsive–related disorders (3.5% vs 5.4%, P < .001) also increased during the pandemic. Conversely, admissions for trauma- and stressor-related disorders (24.9% vs 21.5%, P = .014) and disruptive/impulse-control/conduct disorders (32.1% vs 26.6%, P < .001) decreased during the first pandemic year compared with the previous year. There was no change in overall admissions for ADHD, autism, other neurodevelopmental disorders, bipolar and related disorders, and schizophrenia spectrum and psychotic disorders (Table 1).

The interrupted time series analysis showed several significant findings and trends over the course of the 2 year period. For anxiety disorders (Fig 1), there was a significant difference between the prepandemic and pandemic trends (b = 3.885, t = 2.549, P = .019), with a steep positive linear slope representing 52.4% of the increase in anxiety disorder presentations over the first 12 months of the pandemic, compared with a nonsignificant slope showing stability of anxiety disorder presentations prepandemic.

For eating disorders (Fig 2), there was a significant difference in the prepandemic and pandemic phases (b = 1.941, t = 3.673, P = .002), with a steep positive linear slope representing 68.2% of the increase of eating disorder presentations throughout the first pandemic year, compared with the presence of 0 slope and stability of eating disorder presentations prepandemic. The slopes for anxiety disorders and eating disorders began at lower-volume levels at the pandemic onset compared with the same period the previous year, corresponding to the lockdown period, with striking surges as the pandemic year progressed, exceeding prepandemic levels.

There was a significant difference between the prepandemic and pandemic phases for bipolar and related disorders (b = 0.626, t = 2.348, P = .029), and autism and other neurodevelopmental disorders (b = 1.713, t = 2.574, P = .018), with positive linear trends during the pandemic year, compared with negative slopes prepandemic. There were no significant differences in the prepandemic and pandemic slopes for other psychiatric diagnoses, including depression. For suicide ideation/attempts (Fig 3), there were positive linear slopes in both the prepandemic and pandemic years as expected, with no significant difference despite an increase in overall suicidal presentations during the pandemic year.

For length of boarding (Fig 4), there were positive slopes in both the prepandemic and pandemic years reflective of increasing boarding trends before the pandemic. There was, however, a significant difference between the slopes (b = 74.993, t = 4.176, P < .001), with a very steep positive slope during the pandemic representing 75.4% of the escalation in length of boarding over the 12-month period, compared with 25.4% prepandemic.

To the best of our knowledge, this is the first study to describe psychiatric diagnoses and psychiatric boarding among pediatric patients presenting to the hospital during the first 12 months of the COVID-19 pandemic. Our study revealed that, after the pandemic onset, there was a profound strain on the United States health care system that was already lacking access to pediatric mental health treatment, with higher volumes of patients and families experiencing boarding, and length of boarding more than doubling during the pandemic compared with the prepandemic year. The situation worsened as the pandemic progressed and pediatric mental health declined,³  with lengths of boarding skyrocketing over the 12 month study period. Patients with more severe illness (eg, more suicide attempts), compounded by less treatment availability and access to care because of psychiatric inpatient and outpatient programs experiencing service disruptions from pandemic-related staffing shortages and implementation of infection control protocols,^16–18  likely contributed to escalated boarding.

The current study also highlights important demographic and diagnostic changes during the first 12 months of the pandemic compared with the previous year. There was an increase in adolescent girls presenting in mental health crises, as well as an increase in hospital presentations for suicide ideation/attempts and several psychiatric disorders after the pandemic onset. Similar positive slopes were seen for suicidal presentations during the prepandemic and pandemic years, suggesting the pandemic accelerated the already increasing trend in youth suicidality. Different trends were, however, seen for youth with anxiety and eating disorders because both disorders were stable throughout the prepandemic year but soared as the pandemic progressed in the context of a second virus surge within the state, and uncertain timelines for schools reopening and vaccine availability. Decreased physical activity, increased sedentary behaviors, and increased screen time may have contributed to declines in adolescent emotional well-being and increases in depression and eating disorders.^19,20  The impact of remote learning on body image and disordered eating needs further study because spending significant amounts of the day staring at their own images on virtual platforms may be triggering for adolescents with vulnerability to eating disorders.

Notably, even though children stopped physically attending school during the first year of the pandemic, there were no changes in ADHD diagnoses of youth presenting to the hospital, suggesting that children continued to struggle with ADHD symptoms at similar rates during the pandemic in the context of remote learning. Children with neurodevelopmental and learning disabilities often rely on school-based services, in-home therapeutic supports, and community resources for crucial support.²¹  Despite disruptions in both school-based and in-home services during the pandemic year, there was no change in overall prevalence of patients with autism and other neurodevelopmental disorders; however, there was a positive trend indicative of rising hospital presentations as the pandemic progressed. This may be reflective of the long-term toll of the pandemic on parents/caregivers balancing work from home and caring for a child with high needs, with reduced access to extended family and community supports.

Interestingly, there was a decrease in disruptive/conduct disorders, and trauma and related disorders, after the pandemic onset. Since both are related to interpersonal stressors, the decreased opportunities for peer interactions and conflicts at school, as well as in extracurricular and other social settings, may have contributed to a decline in these presentations after the pandemic onset. Studies have also revealed fewer child protection referrals during the first year of the pandemic because of less referrals originating from schools.²²  Despite restrictions on social gatherings and potentially decreased access to substances during the first pandemic year, youth presenting to the hospital with substance use disorders increased. Heightened stress, anxiety, and boredom likely contributed to increased adolescent substance use during this period.¹⁶  Heightened anxiety, increased germ concerns, hand-washing, and other infection control protocols adopted during the pandemic likely also contributed to the overall increase in youth presenting with obsessive-compulsive disorder, although it was interesting that the trend of these presentations did not increase over the 12-month period even with the second, larger virus surge.

Because the pandemic is extending into a third year and continues to evolve with virus variants, vaccinations, and school reopenings, further longitudinal research is needed to determine long-term effects on pediatric mental health and health care systems. The current study design involved a retrospective chart review from 1 tertiary urban pediatric hospital in the US Northeast, which may limit generalizability because our findings may not be representative of overall population and pandemic experiences. Given the sudden onset and unpredictable nature of the COVID-19 pandemic, longitudinal studies were less feasible during the first 12 months, and the potentially variable impact of the pandemic across geographic regions and seasons underscores the need for local studies. Additionally, it was not possible to control for other contextual factors occurring during the pandemic that may also have affected pediatric mental health.

Trends from the current study forecast concerning future impacts of the pandemic on adolescent health and hospital boarding and raises an urgent call to reform pediatric mental health care. In addition to the impact on children and families, the toll of the ongoing mental health crisis on health care providers and potential for staff burnout in a workforce that was already facing critical shortages cannot be underestimated. By the 20^th month of the pandemic, a national emergency in children’s mental health was declared by the American Academy of Pediatrics, American Academy of Child and Adolescent Psychiatry, and Children’s Hospital Association.²³  Addressing mental health provider shortages, increasing outpatient mental health services, expanding community crisis resources, and increasing psychiatry inpatient and partial hospital programs are essential long-term solutions. Since boarding disproportionately affects youth with mental health conditions more than those with medical conditions,^9,24  such disparities must be addressed to create a health care system that can handle the long-term effects of the pandemic, so families can access appropriate mental health care in the right setting, at the right time.