Pediatric Emergency Department Visits at US Children’s Hospitals During the COVID-19 Pandemic

We conducted a cross-sectional study using data from the Pediatric Health Information System (PHIS), an administrative database including 50 tertiary care children’s hospitals in the United States.¹⁸  The PHIS database contains International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10-CM) diagnosis and procedure codes, billed services (eg, laboratory studies), and hospital charges. Data integrity is jointly monitored by participating institutions and the Children’s Hospital Association.¹⁹  We included all ED visits during the COVID-19 pandemic (March 15, 2020, to August 31, 2020) and a 3-year comparator period (March 15–August 31, 2017–2019) for 27 hospitals with complete administrative and billing data for the study periods. The Wayne State University Institutional Review Board deemed that this study did not qualify as human subject research.

Consistent with authors of a previous report, we defined the start of the pandemic period as the first Sunday (March 15, 2020) after the national emergency declaration on March 13, 2020.²⁰  To account for yearly variation in ED volume and case mix, we averaged visit numbers for the 3-year comparator period across the same calendar dates as the pandemic period. We also examined ED visits from January 1, 2020, to March 14, 2020, to characterize ED visit volumes before the defined pandemic period. To contextualize ED visits from January to August 2020, we obtained daily national SARS-CoV-2 case counts for pediatric and adult patients from the Centers for Disease Control and Prevention.^4,21 

Demographic characteristics for each ED visit included age (<1, 1–4, 5–9, 10–14, 15–18, and ≥19 years), sex, race and ethnicity (non-Hispanic white, non-Hispanic Black, Hispanic, or other), insurance coverage (government, private, or other), geographic region (Midwest, Northeast, South, West), distance from the hospital (calculated as the number of miles between the centroids of the patient’s and hospital’s zip codes), day of the week (weekend versus weekday), and time of arrival to the ED (8 am–3:59 pm, 4 pm–11:59 pm, and midnight to 7:59 am). We did not restrict the cohort by age to capture potential epidemiological changes in adult visits to pediatric EDs during the pandemic period. Race and ethnicity were included in the analyses as a social construct that has previously been associated with differences in ED care–seeking behavior.^22,23 

We described patient complexity for each visit using a previously developed 3-tiered categorical variable: (1) no chronic condition, (2) noncomplex chronic condition, or (3) complex chronic condition (CCC).^24–27  First, we characterized patients as having a CCC, defined as any medical condition expected to last at least 12 months with multiorgan or severe single-organ involvement necessitating pediatric subspecialty care and hospitalization, using a previously validated approach.^24,25  For those without a CCC, we used the Agency for Healthcare Research and Quality’s (AHRQ) Chronic Condition Indicator to categorize patients using ICD-10-CM diagnosis codes into 2 groups: with or without a chronic condition.²⁶ 

We examined primary discharge diagnoses, ED management, disposition, and select quality metrics for each visit.²⁷  Principal ICD-10-CM discharge diagnosis codes were classified by AHRQ’s Clinical Classifications Software Refined.²⁸  ED management included (1) performing laboratory testing or diagnostic imaging; (2) administering medications (other than ibuprofen or acetaminophen); and (3) performing procedures.^27,29  Laboratory testing, diagnostic imaging, and medications administered for each visit were identified through billed transactions. Any charge within each category was categorized as resource utilization within the ED if the charge occurred on day 0 or before 6 pm on day 1 of hospitalization.³⁰  Procedures were identified by current procedural terminology codes, with any assigned current procedural terminology code in the range 10 000 to 69 990 considered a procedure.³⁰  Disposition was categorized as discharged from the hospital, transferred, hospitalized (with “observation” admissions categorized as a hospitalization), or died.³¹  Consistent with previous work, ED visits that did not result in hospitalization and in which no laboratory tests, diagnostic imaging, or procedures were performed were defined as a “low–resource-intensity visit.”²⁹  ED visit charges billed to the patient were inflated to the 2020 US dollar (USD) on the basis of the Consumer Price Index for Hospital and Related Services.³² 

Patient demographic and clinical characteristics were summarized by using frequencies for categorical variables and medians with interquartile ranges (IQRs) for numerical variables. Categorical variables were compared with Rao-Scott χ² tests, adjusting for clustering at the hospital level. Numerical variables were compared with Wilcoxon rank tests. Year-over-year comparisons of daily volumes were shifted to compare similar days of the week to account for the variation in volume by day of the week. For instance, the daily volume change of January 25, 2020, was compared relative to the most proximal Saturday in previous years: January 26, 2019; January 27, 2018; and January 28, 2017. Total charges and charges stratified by discharge status in 2020 were compared with those in each of 2019, 2018, and 2017 with paired signed rank tests. SAS 9.4 (SAS Institute, Inc, Cary, NC) was used for all analyses.

ED visits across 27 pediatric hospitals revealed a sharp decline starting March 12, 2020, ∼1 month before SARS-CoV-2 cases started accelerating (Fig 1). By April 13, 2020, ED visits had declined to a low of 70.5% of previous average years’ volume and remained at 27.2% below prepandemic volumes into July 2020. In total, during the pandemic period, ED volume decreased from 911 026 ED visits averaged over 2017–2019 to 495 052 visits in 2020 (45.7% decline). The decline in volume across sites ranged from 36.1% to 69.9%, with average decreases across institutions by geographic region as follows: the Midwest decreased by 49.3%, West by 46.2%, Northeast by 46.0%, and South by 42.8%.

Compared with previous years, the demographic characteristics of children visiting the ED differed during the 2020 pandemic (Table 1). The proportion of children 15 to 18 years of age increased from 10.3% to 12.9%, and the proportion of adults aged ≥19 years increased from 4.1% to 6.3% during the pandemic (P < .001 for both). The proportion of children with a chronic condition increased from 23.7% to 27.8%, and there was an increase in the proportion of visits among non-Hispanic white children (32.3%–35.9%), those with private insurance (29.8%–32.0%), and patients residing >10 miles from the hospital (43.7%–47.3%) (P < .001 for all).

There were changes in the distribution of the types of encounters during the pandemic, as shown in Table 2 (Supplemental Table 5). The most notable were decreases in the overall number of visits among children with respiratory illnesses (70.0%), including respiratory infections and asthma (69.6% and 73.9%, respectively). Sharp declines in ED visits were observed for less urgent conditions, such as otitis media and upper respiratory infection (75.1% and 69.6% decreases, respectively). Less pronounced decreases were observed for ED visits resulting from injuries and/or poisoning (33.1%) and mental health disorders (29.0%) and among pregnant patients (23.0%). Within the injuries and/or poisoning diagnostic chapter, the weighted average decrease in diagnostic codes related to poisoning (20.2%) was less pronounced than the decline in diagnostic codes related to injury (34.6%) (Supplemental Table 5). The most frequent poisoning-related diagnosis (poisoning by drugs, initial encounter) in the comparator period (n = 2414) declined 8.1% in the pandemic period (Supplemental Table 5). Among the mental health disorders, visits for depressive disorders decreased 38.8%, whereas visits for suicidal ideation or attempt declined 4.0% and stimulant-related disorders were unchanged (2.0%) (Supplemental Table 5).

ED visits during the COVID-19 pandemic were associated with increased resource use, revealed by relative increases in laboratory testing (11.1% [95% confidence interval (CI): 10.9 to 11.2]), diagnostic imaging (5.5% [95% CI: 5.3 to 5.6]), medication administration (2.2% [95% CI: 2.0 to 2.3]), and performance of procedures (3.2% [95% CI: 3.1 to 3.3]). Among the laboratory tests performed, testing for SARS-CoV-2 accounted for 5.2% (26 809 of 490 052) of tests during the pandemic period. Compared with previous years, hospitalizations increased 3.3 and intensive care admissions increased 0.4 percentage points between the comparator and pandemic periods (Table 3). Low–resource-intensity visits decreased from 30.7% of all ED visits in the prepandemic period to 23.7% during the pandemic period (−7.0% [95% CI: −7.1 to −6.9]).

Overall, total charges from cases originating in the ED decreased from an average of $5.7 billion in 2017–2019 to $4.6 billion in 2020. This represents a 20.0% decrease in total charges during the pandemic period. Hospitals experienced a median (IQR) change in charges from 2017–2019 to 2020 of −$35.1 million (−17.8 to −63.8) (P < .001; Table 4). In terms of percentage decreases, the hospitals saw charges decrease by a median (IQR) percent of −33.7% (−26.7 to −38.2) from patients who were discharged from the ED and −16.0% (−6.5 to −23.2) from patients who were admitted (both P < .001).

In this large multicenter study of 495 000 visits across 27 US EDs, we found that the COVID-19 pandemic profoundly impacted pediatric EDs. The overall number of ED visits declined rapidly and consistently across 27 pediatric hospitals, even before SARS-CoV-2 infections began to rise in the United States. It is suggested in our data that although this widespread decline impacted a broad range of pediatric conditions, pediatric EDs continued to care for higher-acuity patients during the pandemic period.

Notably, pediatric ED visits began to decrease steadily before the surge of SARS-CoV-2 cases, suggesting that public health measures, including emergency declarations to control the spread of disease, likely influenced health care–seeking behaviors.¹⁷  SARS-CoV-2 circulated with marked regional differences across the United States,^3,21  but the timing of the impact on pediatric ED volumes was uniform, with all 27 sites experiencing an immediate decrease in the overall number of visits. These declines are consistent with smaller international studies of pediatric hospitals, which revealed rapid declines in ED visits and hospitalizations around the time of each country’s emergency declarations, well before peaks in SARS-CoV-2 cases.^33–35  Yet, it is unclear from our data analysis and the current literature whether visits for SARS-CoV-2 or other common acute and nonacute pediatric conditions contributed most substantially to ED volume recovery.^{14,15,33–35} 

Although our results highlight that ED volumes are revealing slow recovery, our analysis underscores that the marked decline in pediatric ED visits and subsequent hospitalizations had a notable financial impact for children’s hospitals in the United States. Although numerous factors contribute, cancelled nonurgent and elective care visits coupled with the rising costs associated with disrupted medical and equipment supply chains led to a financial crisis for many US health care systems.³⁶  By the end of 2020, the Children’s Hospital Association projects that losses for pediatric hospitals will reach $10 billion,¹⁶  but with our data, we are the first to quantify the true financial impact of the COVID-19 pandemic on pediatric EDs.

Our data reveal the impact of public health initiatives to curb disease spread and directives to avoid nonessential care¹¹ : low–resource-intensity visits dropped significantly during the COVID-19 pandemic. As the low–resource-intensity visits declined during the COVID-19 pandemic, our data suggest a disproportionate decline in pediatric ED visits for potentially vulnerable children, characterized by race and ethnicity, insurance status, and proximity to the hospital. Although initiatives exist to encourage nonurgent visits in the primary care setting,³⁷  the decline in these visits in the context of the pandemic is concerning because pediatric practices also saw marked decreases in routine patient visits, caring for only 30% of their typical clinic volume.³⁸  Health care–seeking behavior is multifactorial, influenced both positively and negatively by public health strategies to contain SARS-CoV-2, but our results draw attention to the fact that health care avoidance may be mediated by various sociodemographic factors.

With our results, we also provide initial insight into care delivery for pediatric patients coming to the ED during the COVID-19 pandemic. As expected, ED care for nonurgent conditions such as acute otitis media and upper respiratory infection declined; however, conditions commonly requiring ED care also had a steep decline in visits, including asthma, injuries, mental health conditions, and child abuse. Stay-at-home orders and school closures may have altered the epidemiology of other viral infections that commonly lead to asthma exacerbations requiring acute care.³³  These same directives contributed to a decline in motor vehicle travel, playground closure, and cancellation of sporting events, partly explaining a decrease in acute injuries and fractures.^39–41  Public health measures to control SARS-CoV-2 may have had unintended benefits for children by mitigating the need for medical care for common childhood conditions.

On the other hand, our findings also support that there may be negative unintended consequences arising from mitigating public health measures that disrupt normal social interactions. First, our data reveal that ED visits related to poisoning and substance use disorders were not as profoundly impacted by the pandemic. Our findings align with reports that calls to poison control centers rose during the COVID-19 pandemic, driven by accidental exposures to various household substances including disinfectants.⁴²  Substance use disorders are more common among adolescents,⁴³  who increasingly sought ED care during the pandemic period. Second, social isolation and familial socioeconomic struggles resulting from public health measures may have negatively affected children’s mental health, particularly for adolescents.⁴⁴  Yet, while 40% of adults reported declining mental health by June 2020,⁴⁵  pediatric ED visits for depression, anxiety, and suicide declined. The longer-term impact of decreased mental health visits to pediatric EDs remains unclear. Finally, we note a decrease in visits during the pandemic for child maltreatment and abuse, aligning with documented decreases in child protection reports. Experts are concerned that child abuse is occurring without intervention from medical and legal authorities⁴⁶ ; thus, it is possible that in the context of the COVID-19 pandemic, the pediatric ED as a safety net may have inadvertently failed.

This study has several important limitations. First, our data reflect the experience of 27 US children’s hospitals, which limits national and international generalizability. However, our results align with previous US and international studies suggesting that similar trends regarding ED volumes and visit intensity exist for general EDs.^{14,15,33–35}  Second, children transferred to the PHIS hospital were not excluded, and reported charges may not reflect overall health care expenditures because diagnostic testing that may have been performed before the ED visit is not included. Third, ED visit acuity cannot be characterized in PHIS by using measures such as the emergency severity index. Thus, proxy measures were used, including hospitalization and intensive care admission, which may not accurately reflect ED visit acuity. Our results, however, note an increase in acuity, which aligns with a single-center US report in which emergency severity index was included in the analysis.¹⁵  Fourth, PHIS is an administrative data set and does not contain detailed clinical data. As such, we are unable to characterize the reason for the ED visit or the clinical course, limiting our ability to examine potential escalations in ED clinical care or increased morbidity for visits occurring during the pandemic period. Given the large number of children included in PHIS, many comparisons were statistically significant but may not be clinically meaningful. Finally, our data do not account for alternative care availability, such as visits occurring in the primary care setting. Thus, we are unable to comment on the exact impact of the decline in ED visits on children’s overall access to health care, but outpatient visits also declined.³⁸ 

Pediatric EDs in the United States saw a dramatic decline in visits for both acute and nonacute conditions during the COVID-19 pandemic. Although visits declined across a broad range of pediatric conditions, the proportional decline for poisoning and mental health visits was less pronounced. Because low–resource-intensity visits decreased and resource use increased, further study is needed to determine if patient outcomes were influenced by care-seeking behaviors during the COVID-19 pandemic.

AHRQ

Agency for Healthcare Research and Quality

CCC

complex chronic condition

CI

confidence interval

COVID-19

coronavirus disease 2019

ED

emergency department

ICD-10-CM

International Classification of Diseases, 10th Revision, Clinical Modification

IQR

interquartile range

PHIS

Pediatric Health Information System

SARS-CoV-2

severe acute respiratory syndrome coronavirus 2

USD

US dollar