Inpatient and Outpatient Differences in Pediatric Patients with Laboratory-confirmed COVID-19.

Among 30,286 pediatric inpatient and outpatient encounters with laboratory-confirmed COVID-19 seen at one of 40 US healthcare organizations, 1586 (5.2%) were inpatient. Encounter types varied by age and sex; the proportion of Black/African American inpatients was significantly higher than outpatients, and Hispanic/Latinx children made up nearly one-fourth of patients.

Studies indicate most children experience less severe coronavirus disease 2019 (COVID-19) illness than adults,[1] but rates of pediatric COVID-19 related hospitalizations increased in recent months.[2] Factors associated with hospital admission compared with outpatient encounters among children with COVID-19 are poorly understood. While 1 other study to date compared pediatric inpatients to those seen in outpatient settings,[3] we sought to conduct this study using a stricter definition of COVID-19, including requiring a positive test, and a larger time frame to better understand the risk factors for the hospital admission.

METHODS

We conducted a retrospective cohort study of 10,285,590 inpatient and outpatient encounters from 1/1/2020 to 8/31/2021 among patients ≤18 years old with COVID-19 seen at 1 of 40 healthcare organizations in the United States reporting data to TriNetX, a global health research network with continuously refreshed data for more than 45 million patients in the United States.[4] COVID-19 cases were defined by having an International Classification of Diseases (ICD-10) code (U07.1) and a positive severe acute respiratory syndrome coronavirus 2 test 2 weeks before or 3 days after the index encounter. Excluded were patients with an ICD-10 code of B97.89 or ICD-9 079.89 for other viral infections. Outpatient was defined as an encounter in the emergency department or outpatient clinic. Non-COVID-19 encounters were those without the codes defined above for cases. Comorbidities were defined using ICD-10 codes documented before or at the index encounter. Treatments given the day of up to 5 days after the index event were included. Descriptive statistics, t-test and χ2 tests were computed using SAS Enterprise Guide v7.1 (SAS Institute, Inc., Cary, NC). This study was deemed exempt by our organization’s Institutional Review Board.

RESULTS

As of August 31, 2021, 30,286 encounters among patients ≤18 years old with laboratory-confirmed COVID-19 were included in the study population; 1586 (5.2%) were inpatient encounters. The cohort was racially/ethnically diverse with 24% of patients Hispanic/Latinx and 23% Black/African American among patients with these data reported.

The distributions of age groups were similar between inpatient and outpatient encounters but differed when further analyzed by sex, with inpatient encounters among patients <13 years of age occurring more often among males and admissions among patients 13–18 years of age occurring more often among females. Among those with race reported, Black/African American patients had more frequent inpatient (33%) than outpatient encounters (29%) for COVID-19, and these were both higher than non-COVID-19 encounters among Black/African American patients (22%), whereas white patients were less likely to have inpatient (63%) than outpatient encounters (69%) for COVID-19 and these were both lower than non-COVID-19 encounters among white patients (73%) (Table). In regard to geographic differences, patients in the South were more likely to be admitted to a hospital than to be seen as an outpatient (59% vs. 53%) but these rates were similar to non-COVID-19 encounters in the South (59%) and patients in the West were less likely to have an inpatient admission than an outpatient encounter (3% vs. 16%) but these differed from the rate of non-COVID-19 encounters in the West (8%).

All major comorbidities analyzed were more common among inpatients compared with outpatients and among inpatients compared with non-COVID-19 encounters. Comparing inpatients, outpatients and non-COVID-19 encounters, notable differences were higher rates of metabolic disorders: 27% vs. 7% vs. 7%; diabetes: 6% vs. 1% vs. 0.5% and overweight/obesity: 10% vs. 6% vs. 4%; anemia: 12% vs. 2% vs. 1%; sleep disorders: 11% vs. 6% vs. 1%; and anxiety disorders: 15% vs. 7% vs. 6% (all P < 0.001). Antibiotics were more often given to inpatients than outpatients (48% vs. 7%, P < 0.001), as were glucocorticoids (43% vs. 5%, P < 0.001).

DISCUSSION

This study utilized a large sample of healthcare organizations throughout the United States to describe differences in pediatric laboratory-confirmed COVID-19 cases between inpatient and outpatient settings. We found that among younger age groups a larger proportion of inpatient encounters were males whereas among older age groups a larger proportion of inpatient encounters were females. We also found that the proportion of Black/African American inpatients was significantly higher than the proportion of outpatients and Hispanic/Latinx children made up nearly one-fourth of all pediatric patients with COVID-19. Previous studies have also observed these disparities among COVID-19 pediatric patients,[5] highlighting the urgent need for interventions to address this issue.

We observed differences in the proportion of children admitted to the hospital with COVID-19 based on geographic location. Disparate access to testing and different messaging based on geography has occurred during the COVID-19 pandemic.[6] The geographic differences observed in this study may reflect disparities in access to healthcare, higher rates of chronic conditions or other factors, and warrants further investigation.

Pediatric patients with multiple types of comorbidities were more likely to be inpatients, which is supported by previous studies.[7] Nearly half (48%) of all pediatric inpatients in this study received antibiotics, despite being hospitalized for a viral infection. Similar antibiotic prescribing rates were observed in adults hospitalized with COVID-19[8] despite low incidence of bacterial co-infections, indicating a strong need for active antibiotic stewardship programs during this pandemic.

Limitations of this study include race not reported for 21% of patients and the inability to include length of stay or adjusted analyses based on level of access to TriNetX data. Among the non-COVID-19 encounters, only a small subsample (4%) was available for analysis of comorbidities due to computational limitations, therefore this subsample may not represent all non-COVID-19 patients. The relatively low sample size of comorbidities also limited analysis by specific comorbidities. In addition, no previously published information is available on the representativeness of TriNetX among medical encounters in the United States. Advantages to this study include use of a large, national database with real-time updating, inpatient and outpatient encounters, and being able to include patients with both a positive severe acute respiratory syndrome coronavirus 2 test and corresponding ICD-10 code.

The impact of COVID-19 on children is not fully elucidated, and understanding the epidemiology of pediatric patients presenting to inpatient and outpatient settings may guide future prevention efforts and resource distribution. This pandemic exposed existing healthcare inequities among children that require corrective action.

Table 1.

Demographic and Comorbidity Differences Between Inpatient and Outpatient Encounters Among Pediatric Patients With COVID-19 as of August 31, 2021

	Inpatient Encounters (n = 1586)	Outpatient Encounters (n = 28,700)*	Non-COVID-19 Encounters (n = 10,255,304)
	n (%)	n (%)	n (%)
Age, in years
Mean (SD)	10.1 (6.3)	10.7 (5.7)	10.1 (5)
<1	107 (7)	733 (3)	133,956 (1)
1–5	376 (24)	6078 (21)	2,200,242 (22)
6–12	357 (22)	8507 (30)	4,098,291 (40)
13–17	552 (35)	10,380 (36)	3,164,680 (31)
18	194 (12)	3002 (10)	658,135 (6)
Sex†
Male	804 (51)	14,359 (50)	5,359,580 (52)
Female	779 (49)	14,274 (50)	4,864,481 (48)
Race‡
Black/African American	429 (33)	6437 (29)	1,722,191 (22)
Other race§	43 (3)	468 (2)	349,446 (5)
White	820 (64)	15,664 (69)	5,679,387 (73)
Ethnicity¶
Hispanic or Latinx	221 (18)	4219 (21)	1,674,474 (24)
Not Hispanic or Latinx	991 (82)	15,893 (79)	5,287,879 (76)
Geographic region‖
Northeast	253 (16)	4696 (16)	1,562,163 (16)
Midwest	355 (22)	4264 (15)	1,635,486 (17)
South	930 (59)	15,137 (53)	5,709,154 (59)
West	48 (3)	4603 (16)	790,140 (8)
Comorbidities**
Acute kidney failure and chronic kidney disease	89 (6)	219 (1)	1474 (0.3)
Anemia	195 (12)	677 (2)	5843 (1)
Anxiety disorder	239 (15)	2130 (7)	26,867 (6)
Asthma	227 (14)	2844 (10)	34,583 (8)
Cardiac arrhythmia	54 (3)	199 (1)	2410 (0.5)
Cardiomegaly	74 (5)	235 (1)	1714 (0.4)
Cerebral palsy	44 (3)	135 (0.5)	1581 (0.4)
Cerebrovascular diseases	49 (3)	151 (1)	1515 (0.3)
Congenital malformations	322 (20)	3169 (11)	50,402 (11)
Diabetes mellitus	97 (6)	251 (1)	2244 (0.5)
Epilepsy and recurrent seizures	94 (6)	403 (1)	5,534 (1)
Heart failure	42 (3)	90 (0.3)	576 (0.1)
Hypertensive diseases	138 (9)	450 (2)	3417 (0.8)
Ischemic heart diseases	38 (2)	64 (0.2)	233 (0.1)
Malnutrition	78 (5)	211 (1)	1419 (0.3)
Metabolic disorders	433 (27)	2044 (7)	29,593 (7)
Mood affective disorders	179 (11)	1,152 (4)	13,649 (3)
Neoplasms	136 (9)	1050 (4)	15,252 (3)
Neutropenia	80 (5)	201 (1)	1132 (0.3)
Overweight and obesity	153 (10)	1743 (6)	17,513 (4)
Purpura and other hemorrhagic conditions	109 (7)	249 (1)	1589 (0.4)
Sickle-cell disorders	47 (3)	196 (1)	1491 (0.3)
Sleep disorders	177 (11)	1665 (6)	2387 (0.5)

*Differences between inpatient and outpatient encounter variables were statistically significant (P < 0.05) except for sex.

†Not reported for n = 3 inpatient encounters, n = 67 outpatient encounters and n = 31,243 non-COVID-19 encounters.

‡Not reported for n = 294 inpatient encounters, n = 6041 outpatient encounters and n = 2,504,280 non-COVID-19 encounters.

§Includes American Indian or Alaskan Native, Asian and Native Hawaiian or other Pacific Islander.

¶Not reported for n = 374 inpatient encounters, n = 8588 outpatient encounters, n = 3292,951 non-COVID-19 encounters.

‖Not reported for 558,361 non-COVID-19 encounters.

**Reported for 446,290 non-COVID-19 encounters.