Common Diagnoses and Costs in Pediatric Hospitalization in the US

Abstract

This cross-sectional study ranks the prevalent diagnoses among hospitalized children and related costs by hospital type.

Approximately 2 million nonbirth pediatric hospitalizations occur annually in the US,¹ but scant evidence guides inpatient pediatric care. A critical first step in prioritizing research on hospitalized children is the identification of costly and common reasons for hospitalization,² but existing studies are either limited in scope or outdated.^1,2,3 In this study, we aimed to identify the demographic characteristics of hospitalized children nationally and rank the most common and costly diagnoses in this population.

Methods

We performed a cross-sectional analysis of data from the 2016 (most recent available) Kids' Inpatient Database (KID), the only all-payer, administrative data set designed for assessing hospitalizations of children in the US.⁴ We included all nonbirth hospital discharges for children aged 0 to 17 years, and we applied discharge weights to produce national estimates. Race and ethnicity were self-identified by parents and families using each hospital’s classification system (Asian or Pacific Islander; Black, non-Hispanic; Hispanic; Native American; White, non-Hispanic; and other [unlisted race and ethnicity category and more than 1 race and ethnicity]). Records were excluded if their primary diagnosis or cost data were missing (n = 32 310 excluded [1.2%, unweighted]). We grouped primary diagnoses using the Pediatric Clinical Classification System codes^3,5 and rank-ordered total prevalence and cost by hospital type. The eMethods in the Supplement provides additional information on our methods. This study was deemed exempt from review by the University of California, San Francisco Institutional Review Board, which also waived the informed consent requirement because the study used a deidentified data set. Data analyses were conducted using SAS, version 9.4 (SAS Institute Inc), from April 19 to July 28, 2021.

Results

An estimated 1 777 023 nonbirth pediatric hospitalizations occurred in 3768 US hospitals in 2016 (Table). Most hospitalizations involved children with 1 or more chronic medical conditions and children residing in communities with below-average median incomes (first or second quartile). We found that 54.3% of hospitalizations took place in urban teaching hospitals (n = 960 947), 9.6% in urban nonteaching hospitals (n = 170 531), 31.6% in freestanding children’s hospitals (n = 559 850), and 4.4% in rural hospitals (n = 78 695).

Table. Characteristics of Nonbirth, Pediatric Hospitalizations in the United States, 2016 (N = 1 777 023).

Patient characteristic	No. (%)
Age, y
<1	447 906 (25.3)
1-4	377 887 (21.3)
5-12	448 666 (25.3)
13-17	495 563 (28.0)
Sex
Female	863 941 (48.8)
Male	913 082 (51.2)
Chronic medical condition
None	753 900 (42.6)
Noncomplex, chronic condition	511 172 (28.9)
Complex, chronic condition	504 951 (28.5)
Median income quartile by zip code
1st	563 907 (31.9)
2nd	430 540 (24.3)
3rd	407 644 (23.0)
4th	341 711 (19.3)
Missing data	26 161 (1.5)
Primary payer
Government	976 467 (55.1)
Private	676 778 (38.2)
Self-pay, other payer, or missing data	118 354 (6.7)
Race and ethnicitya
Asian or Pacific Islander	54 349 (3.1)
Black, non-Hispanic	283 687 (16.0)
Hispanic	369 213 (20.9)
Native American	16 500 (0.9)
White, non-Hispanic	793 760 (44.8)
Otherb	89 074 (5.0)
Missing data	163 356 (9.2)
Elective hospitalization	331 608 (18.7)
Accompanying ED record	894 156 (50.5)
Transfer status
Not a transfer	1 465 707 (82.8)
Transferred inc	296 355 (16.7)
Missing data	8359 (0.5)
Discharge disposition
Routine	1 640 829 (92.7)
Transfer to another facility	65 741 (3.7)
Home health	49 283 (2.8)
AMA	2710 (0.2)
Death	9746 (0.6)
Unknown or missing data	1491 (0.1)
LOS, median (IQR), d	3 (1-5)

Abbreviations: AMA, against medical advice; ED, emergency department; LOS, length of stay.

^a Race and ethnicity were self-identified by parents and families using each hospital’s classification system.

^b Other included both unlisted categories and more than 1 race and ethnicity.

^c Transferred in refers to transfer from another acute care hospital.

Figure, A illustrates the diagnosis prevalence rankings. The most common diagnoses overall were bronchiolitis (n = 97 564 hospitalizations), pneumonia (n = 86 702 hospitalizations), asthma (n = 84 436 hospitalizations), major depressive disorder (n = 79 147 hospitalizations), and cellulitis (n = 42 556 hospitalizations). Both major depressive disorder (rank 1 in urban nonteaching hospitals; n = 15 201 hospitalizations) and neonatal hyperbilirubinemia (rank 5 in urban nonteaching hospitals; n = 7148 hospitalizations) were ranked higher outside of freestanding children’s hospitals (in freestanding children’s hospitals, rank 8 [n = 9657 hospitalizations] and rank 20 [n = 5267 hospitalizations], respectively). Chemotherapy, seizures, and respiratory failure were ranked higher in freestanding children’s hospitals (ranks 3, 5, and 7, respectively).

Figure. Most Common and Costly Diagnoses in Pediatric Hospitalizations, Overall and by Hospital Type.

NC indicates no cases.

^aOmitted because of the reporting or disclosure policies of the Healthcare Cost and Utilization Project.

Figure, B illustrates the diagnosis cost rankings. The costliest diagnoses overall were respiratory failure ($982.8 million), pneumonia ($785 million), chemotherapy ($781.4 million), bronchiolitis ($702.2 million), and septicemia ($697 million). Both asthma (rank 4 in rural hospitals; $14.3 million) and major depressive disorder (rank 5 in rural hospitals; $14.2 million) were ranked higher outside of freestanding children’s hospitals (rank 17 [$154.4 million] and rank 44 [$80 million], respectively). Respiratory failure, chemotherapy, septicemia, scoliosis, and hypoplastic left heart syndrome were ranked higher in freestanding children’s hospitals (ranks 1, 2, 3, 4, and 7, respectively). Diagnoses with both high prevalence and cost overall included bronchiolitis, pneumonia, asthma, and major depressive disorder.

Discussion

Most pediatric hospitalizations took place outside of freestanding children’s hospitals, and the most common and costly diagnoses differed by hospital type. We believe these findings help guide the selection of high-priority diagnoses for pediatric research and quality improvement efforts that would make the greatest impact for children across the US. Future research and quality improvement prioritization efforts are needed, which should (1) leverage the findings from this study; (2) engage a wide spectrum of stakeholders (eg, parents or caregivers, nurses, and administrators) in the care of hospitalized children; and (3) incorporate additional factors such as practice variation, health disparities, and preventable harms.

One important finding was the increase in hospitalizations for major depressive disorder since 2012 (from rank 9 [n = 28 914 hospitalizations] in 2012 to rank 4 [n = 79 147 hospitalizations] in 2016).¹ Recent data suggested that additional increases in mental health hospitalizations were associated with the COVID-19 pandemic.⁶ These findings reinforce the need to identify effective interventions for children who are hospitalized with mental health disorders.

Study limitations included the data in the KID database. Because the KID database contains discharge-level rather than patient-level records,⁴ the diagnoses represented by multiple hospitalizations from single patients could not be identified. In addition, the most recent KID data available were from 2016 and may not reflect current health care use. The study focused on primary diagnoses; therefore, it may have underestimated diagnoses that are less frequently selected as primary diagnoses.

Supplement.

eMethods. Detailed Study Methods