Results of COVID-19 Surveillance in a Large United States Pediatric Healthcare System over One Year

Background: The lack of SARS-CoV-2 antigen surveillance testing in the pediatric population has inhibited accurate infection and hospitalization prevalence estimates. We aim to report the estimated prevalence of and risk factors for COVID-19 infection, hospitalization, and intensive care unit (ICU) admission across the three United States (US) waves in one of the largest pediatric healthcare systems in the nation. Methods: Retrospective electronic health record (EHR) review of all COVID-19 surveillance data among children aged 0–19 years seeking healthcare at one pediatric healthcare system that serves predominantly Medicaid-dependent families from 1 March 2020 to 31 March 2021. COVID-19 infection status (Y/N), hospital admission (Y/N), and ICU admission (Y/N) are the main outcomes. Results: Of 22,377 children aged ≤ 19 years tested for SARS-CoV-2 infection from March 2020–March 2021, 3126 were positive (14.0%), and out of those positive, 53.7% were hospitalized and 2.9% were admitted to the ICU. Compared to Wave 1 (1 March 2020–31 May 2020), the risk of a positive test increased from 16% (RR 1.16, 95% CI, 1.07–1.26) in Wave 2 (1 June 2020–31 October 2020) to 33% (RR 1.33, 95% CI, 1.23–1.44) in Wave 3 (1 November 2020–31 March 2021). Similarly, compared to Wave 1, the risk for hospitalization increased 86% (RR 1.86, 95% CI, 1.86–2.06) in Wave 2 and 89% in Wave 3 (RR 1.89, 95% CI, 1.70–2.08), and the risk for ICU admission increased from 10% in Wave 2 (RR 1.10, 95% CI, 0.39–3.01) to 310% in Wave 3 (RR 3.10, 95% CI, 1.21–7.80). Children with asthma, depressive disorders, type 1 or 2 diabetes, and anemia were more likely to be hospitalized while children with diabetes, obesity, cardiac malformations, and hypertension were more likely to be admitted to the ICU versus children without these conditions. Conclusions: Children were cumulatively impacted by the COVID-19 pandemic through the three US waves with more than a third hospitalized in Wave 3. Children with underlying health conditions were particularly at risk for severe illness and should be monitored for any long-term impacts.


Introduction
As of 17 June 2021 more than 4 million total pediatric COVID-19 cases, caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, have been reported in the United States (US), representing 14.2% of cumulative cases [1]. However, this number is likely an underestimate due to a high proportion of mild and asymptomatic cases in the pediatric population and that testing was not widely available for children [2].
Of the more than 600,000 Americans who have died since the beginning of the pandemic, fewer than 450 have been under the age of 18 [1]. Pediatric COVID-19 hospitalizations are significantly lower than those for adults, further suggesting that children have less severe symptoms and illness from COVID-19 compared to adults [3]. However, of hospitalized children, one-third are admitted to intensive care where approximately six percent receive mechanical ventilation, and a small number develop multisystem inflammatory syndrome in children (MIS-C) that includes hypotension, severe abdominal pain, and cardiac dysfunction [4].
Reports by geographic locations indicate that pediatric COVID-19 cases are substantially greater in economically disadvantaged census tracts [5][6][7]. One of the main challenges, however, has been the absence of a denominator to estimate true incidence of COVID-19 in the pediatric population, in either community or healthcare based settings. Specifically, to date no pediatric COVID-19 surveillance data has been reported from large healthcare systems serving low resource patients and their families. Moreover, although several reports have documented clinical presentation of children with MIS-C [8], no reports to date have captured the proportion of children of all ages who tested positive stratified by hospitalization and intensive care unit (ICU) status through the third wave of the pandemic in the US. Here we report the trends of COVID-19 results and hospitalizations during a 13-month period in the 6th largest pediatric healthcare system in the country. It was hypothesized that pediatric cases would parallel similar trends as those seen in the US adult population.

Study Design
The UTHealth Committee for the Protection of Human Subjects reviewed and approved a retrospective cross-sectional electronic health record (EHR) (EPIC, Verona, WI, USA) analysis of pediatric patient records from 1 March 2020 to 31 March 2021. Office visit and hospital encounter data of all patients who had a COVID-19 test were extracted from the EHR including demographics, diagnoses, vital measures, body mass index (BMI), hospital and ICU admission status, hospital and ICU stay length, and outcomes.

Study Setting
Children's Health System of Texas (CHST; serving 68% Medicaid and 60% ethnic minority families) is located in Dallas, Texas, which was a COVID-19 hot spot [9] throughout late winter 2021 [10]. The Centers for Disease Control and Prevention (CDC) defines a hotspot county as meeting all four of the following criteria, relative to the date assessed: (1) >100 new COVID-19 cases in the most recent 7 days, (2) an increase in the most recent 7-day COVID-19 incidence over the preceding 7-day incidence, (3) a decrease of <60% or an increase in the most recent 3-day COVID-19 incidence over the preceding 3-day incidence, and (4) the ratio of 7-day incidence/30-day incidence exceeds 0.31. In addition, hotspots must have met at least one of the following criteria: (1) >60% change in the most recent 3-day COVID-19 incidence, or (2) >60% change in the most recent 7-day incidence [10].

Study Participants
Participants included any child who aged 0-19 years presenting for medical care at CHST's two main hospitals or ambulatory clinics for any reason from 1 March 2020 to 31 March 2021. All children, regardless of symptoms of COVID-19 were tested.

Exposures of Interest
Primary exposure of interest was SARS-CoV-2 across waves, typically defined by Texas-specific regional infectious patterns [10]. Secondary exposures included underlying medical conditions.

Covariates
Covariates of interests include age, sex, race/ethnicity, BMI, insurance, and comorbidities. Participants included any child who aged 0-19 years presenting for medical care at CHST's two main hospitals or ambulatory clinics for any reason from 1 March 2020 to 31 March 2021. All children, regardless of symptoms of COVID-19 were tested.

Exposures of Interest.
Primary exposure of interest was SARS-CoV-2 across waves, typically defined by Texas-specific regional infectious patterns [10]. Secondary exposures included underlying medical conditions.

Statistical Analysis
Categorical variables were presented as frequencies and percentages, and continuous variables were summarized as means (standard deviation [SD]). The demographic and baseline characteristics of pediatric patients with COVID-19 positive results were compared by hospital admission (Y/N) and ICU admission (Y/N) status by Pearson Chi-square tests, Fisher's exact tests or two-sample t-test with equal or unequal variance, respectively. Age and BMI were analyzed as both continuous variable and categorical variable with three levels. A CDC SAS macro program was used to calculate BMI percentiles based on US national growth charts and obesity was defined as ≥95th%ile adjusted for age and sex [13]. COVID-19 related symptoms were compared by COVID test results (positive vs. negative) via Pearson Chi-square tests. Generalized linear models with normal distribution and identity link function were built to assess time trend of COVID-19 test results by the total of number tests and percentage of positive results. By using patients from Wave 1 as the reference group, relative risks of Wave 2 and Wave 3 for positive tests, hospital admission, and ICU admission were computed by the following equation, respectively: Relative Risk = Incidence rate in Wave 2 or Wave 3 Incidence rate in Wave 1 Three stepwise logistic regression models using p-value < 0.1 and < 0.05 as entrance and retention criteria were created to explore possible predictors of a positive test (Y/N), hospital admission (Y/N), and ICU admission (Y/N), separately. Independent variables included age, sex, race/ethnicity, BMI, and comorbidities (Y/N), such as asthma, anemia, anxiety, cardiac malformation, depressive disorders, diabetes (type 1 and type 2), epilepsy, hypertension, neurodevelopmental disorders, obesity, substance use disorders, and injury/trauma. All statistical analyses were performed using SAS v9.4 (SAS Institute, Cary, NC, USA) and R v4.0.2 (R core team, 2020). Two-sided p-value < 0.05 is considered significant.

Results
Between 1 March 2020 and 31 Mach 2021, a total of 23,719 CHST unique patients had a COVID-19 PCR test. After excluding those aged greater than 19 years old (n = 1342), the analytical sample included 22,377 unique patients. Out of those, the analytical sample used for analysis here included 3126 (14.0%) children with a positive test. Among those with a positive test, 1679 (53.7%) children were hospitalized and 49 (1.6%) were admitted to the ICU. (Supplemental Figure S1). The absolute number of children with a positive test ranged from 593 in March 2020 to 60 in March 2021 (Figure 1, Panel A). Over time, the prevalence of positive tests was 17.1% in March 2020, peaked at 30.1% in December 2020, then fell to 9.3% in March 2021. (Figure 1, Panel B).
Children with a positive COVID-19 test were more likely to report a symptom than those with a negative test (29.8% vs. 18.7%, p < 0.001). However, it is notable that the majority (70.2%) of children without any common symptoms had a positive test. Nasal congestion was the only symptom that did not differ between positive and negative groups (2.9% vs. 2.4%, p = 0.078) ( Table 2).    Children with obesity had more than four times higher odds of ICU admission (aOR 4.6, 95% CI 1.9-10.8). Comorbidities dramatically increased the odds of ICU admission. Specifically, children with diabetes (aOR 25.72, 95% CI 10.49-63.06), cardiac malformation (aOR 46.69, 95% CI 13.5-161.9), or hypertension (aOR 11.21, 95% CI 1.27-99.04) had 11 to 46 times higher odds of ICU admission, respectively, compared to their counterparts (Table 3).

Discussion
This analysis shows the proportion of children of all ages who tested positive for COVID-19 stratified by hospitalization and intensive care unit (ICU) status through the US third wave of the pandemic. In this analysis over 13 months in one large pediatric healthcare system serving predominantly low-income families, among the over 22,000 who received a COVID-19 test, results showed that the risk of a positive COVID-19 antigen test, hospitalization, and ICU admission increased over time with each subsequent wave. The majority (70%) of those with a positive test had no symptoms. Comorbidities including type 1 or type 2 diabetes, hypertension, obesity, and cardiac malformations increased the odds of ICU admission significantly, while children with asthma, depressive disorders, or diabetes were significantly more likely to be hospitalized. Collectively these findings show that the third wave of the US COVID-19 pandemic had a significant impact on children from low resource backgrounds and with underlying health conditions in particular.
Analysis of CDC COVID-NET data showed that in the first and second wave of the pandemic (March 1-25 July 2020) the cumulative COVID-19-associated hospitalization rate among children aged <18 years was 8.0 per 100,000 population, with the highest rate among children aged <2 years [14]. Reports from China's CDC corroborate those of US findings in that children were less susceptible to hospitalization including before COVID-19 hit this country (16 January 2020 to 8 February 2020) [15]. Systematic reviews that included studies from the US, Italy, and the UK among others during the first 6 months of 2020 also supported this finding [16,17]. There are few [18], if any, data on how the third and most deadly wave of the pandemic has impacted the pediatric population in all age groups. COVID-NET analysis of all three US waves among 12-17 year olds only showed that 31.4% were admitted to an intensive care unit (ICU) [18], similar to our results here among all age groups. Results here suggest that it was more consequential in terms of not only infections among children, but severe illness resulting in ICU admission versus the first wave for all age groups.
Results from the CDC COVID-NET analysis also showed that Hispanic and non-Hispanic Black children had higher cumulative rates of COVID-19-associated hospitalizations (16.4 and 10.5 per 100,000, respectively) versus non-Hispanic White children (2.1 per 100,000) [14]. In agreement with the COVID-NET studies, another study of 33 healthcare organizations showed the risk of hospitalization was greater in non-Hispanic Black and Hispanic children compared with non-Hispanic White children [19]. However, our results showed that non-Hispanic Black children were 32% less likely to be hospitalized versus non-Hispanic Whites and Hispanic and mixed race/other children were also slightly less likely to be hospitalized. Another analysis using the CDC data show that hospitalization and in-hospital death were rare in children diagnosed with COVID-19 but children at higher risk for these outcomes include those with an underlying medical condition, as well as non-Hispanic Black children [20]. We excluded children who died from this analysis. A more recent analysis through Wave 2 and partially into Wave 3 (March 2020 through January 2021) using the COVID-NET data found a higher risk of severe COVID-19 illness among children with medical complexity and certain underlying conditions, such as type 1 diabetes, cardiac and circulatory congenital anomalies, and obesity [21]. These results are similar to those found here in that children with either type 1 or type 2 diabetes were at increased risk for both hospitalization and ICU admissions [21].
Results here showed that children with obesity had increased odds of a positive COVID-19 test and IC admission, but not hospitalization. This finding may be due to children with obesity presenting with more severe illness requiring immediate ICU admission. Obesity is a known risk factor for severe COVID-19 illness in adults [21] and especially severe obesity [22] but not necessarily a COVID-19 diagnosis. Our results are also consistent with other studies that show that children with asthma, diabetes, and mental health disorders including depression are at increased risk for hospitalization [21,23]. Conversely, children with asthma in particular were less likely to have a positive COVID diagnosis, but more likely to be hospitalized, which is predictable given the pulmonary nature of the disease, but not in agreement with other national studies [18].
More than 70% of children with a COVID-19 positive diagnoses did not report any common symptoms suggested by CDC [12]. This is in agreement with other studies both in the US and globally [2,24]. This finding suggests that the virus had been circulating widely, and silently in the majority of children in the geographic catchment area that was a COVID-19 hot spot [9].

Limitations
This study had limitations. First, the use of ICD-10-CM diagnostic codes to identify COVID-19 cases might have resulted in misclassification bias, although there were few other cases of similar respiratory illness (e.g., influenza, RSV) during the same time period. Specifically, a COVID-19 specific ICD-10 code was not available until January 2021 [25] and thus may have not accurately captured children with a previous diagnosis. Second, the use of ICD-10-CM codes to identify underlying medical conditions may have also resulted in misclassification bias. Indeed, others have suggested that miscoding of medical claims can lead to misclassification bias from ICD codes [26]. Third, due to the cross-sectional study design and analysis, casual relationships cannot be inferred between previously existing medical conditions and diagnoses and severe COVID-19 illness. Fourth, because EHR data were used to retrospectively capture hospitalization and ICU admission status, the temporality may be affected. However, the pattern of hospitalization data in the present analysis is similar to the CDC's report indicting the accuracy of our results [24]. Fifth, the analysis only included frequent chronic conditions, and thus rare conditions were not assessed as potential risk factors for hospitalization and ICU admission, and thus the results presented here are not representative of all children with COVID-19. Finally, the characteristics of the pediatric population who had physical visits may be different from those who completed telehealth visits.

Conclusions
This analysis can be summarized in the following three conclusions: (1) the proportion of children of all ages who tested positive for SARS-CoV-2 stratified by hospitalization and ICU status increased across three waves; (2) children with diabetes, hypertension, obesity, and cardiac malformations were particularly at risk for ICU admission in the third wave; and (3) collectively these findings show that the third wave of the US COVID-19 pandemic had a significant impact on children with underlying health conditions and from low resource backgrounds and was more severe than earlier waves.
Author Contributions: S.E.M. was responsible for the conception and design of the analysis. M.S.M. led the data coordination components including the coordination of data acquisition from the electronic health record, data merging, and logistics. S.E.M. and L.X. were responsible for statistical analyses and table and figure design. H.W.K.III, G.L.D. and J.S.K. contributed epidemiological and clinical expertise to the manuscript based on analytical results. The first draft was written by S.E.M. All authors contributed to data interpretation, critically reviewed the first draft, and approved the final version and agreed to be accountable for the work. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Committee for the Protection of Human Subjects of the University of Texas Health Science Center at Houston (HSC-SPH-20-0944, date of approval 27 August 2020).
Informed Consent Statement: Patient consent was waived due to the retrospective nature of the analysis.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to ethical restrictions.

Conflicts of Interest:
The authors declare no conflict of interest.