Abstract
We examined associations between a validated, multidimensional measure of social determinants of health and population-based hospitalization rates among children <18 years across 18 states from the 2017 HCUP State Inpatient Databases and US Census. The exposure was ZIP code-level Child Opportunity Index (COI), a composite measure of neighborhood resources and conditions that matter for children’s health. The cohort included 614,823 hospitalizations among a population of 29,244,065 children (21.02 hospitalizations per 1000). Adjusted hospitalization rates decreased significantly and in a stepwise fashion as COI increased (P<.001 for each), from 26.56 per 1000 (95% CI 26.41, 26.71) in very low COI areas to 14.76 per 1000 (95% CI 14.66, 14.87) in very high COI areas (IRR 1.8; 95% CI 1.78, 1.81). Decreasing neighborhood opportunity was associated with increasing hospitalization rates among children in 18 US states. These data underscore the importance of social context and community-engaged solutions for health systems aiming to eliminate care inequities.
INTRODUCTION
The Childhood Opportunity Index (COI) 2.0 is a validated, multidimensional measure of neighborhood resources and conditions that impact children’s health in the United States (US).1 Prior studies demonstrate associations with lower neighborhood opportunity and increasing emergency department (ED) and hospital utilization at freestanding children’s hospitals.2–4 However, most acute care hospitalizations of children occur at general hospitals,5 and it is unclear if the observed associations reflect differences in care-seeking behavior or true population differences. We aimed to define associations between COI and population-based hospitalization rates in US children.
METHODS
Data Source and Study Design
We performed a cross-sectional, population-based analysis of data from 18 states using the 2017 Healthcare Cost and Utilization Project (HCUP) State Inpatient Databases (SID)6 and 2017 ZIP code level US census data estimated from the 2015-2019 American Community Survey (ACS). The SID is an all-payer administrative database that includes inpatient records from all hospitals types in 49 US states.6 The SID does not include an observation status indicator,7 and the HCUP revisit variables (for tracking ED revisits and readmissions) are not available for all states.
Child Opportunity Index 2.0
The primary exposure was ZIP code level, nationally-normed COI quintile (very low to very high), a validated measure that includes 29 indicators of neighborhood conditions and resources across 3 domains—education, health and environment, and social and economic factors—that impact children’s healthy development.1 Each indicator is transformed to a z-score, standardized, and weighted by how strongly it predicts children’s long-term health and economic outcomes. The indicators are combined into three domain scores and an overall score. Overall scores may also be divided into nationally-normed quintiles. Prior studies suggest that ZIP code and census tract level COI yield similar results when using very large datasets like the SID.8,9
Population-Level Hospitalization Rate
The primary outcome measure was hospitalization rate per 1,000 children. The numerator included all inpatient hospitalizations captured in the 2017 SID from 18 states aggregated at the ZIP code level. The denominator was the 2017 pediatric population from the same 18 states estimated using the 2015-2019 ACS aggregated to the ZIP code level.7
To explore drivers of trends, we estimated hospitalization rates among a subset of common primary diagnosis groups. Diagnoses were selected by identifying Pediatric Clinical Classification System (PECCS) Codes within the dataset that accounted for ≥1% of admissions. PECCS classifies International Classification of Diseases, Tenth Revision, Clinical Modification diagnoses codes into mutually exclusive, clinically meaningful pediatric conditions.10 We further grouped these diagnoses into 7 diagnosis subgroups: Mood Disorders (major depressive disorder, mood disorders, bipolar disorder, suicide and intentional self-inflicted injury, adjustment disorder); Infectious Disease, Respiratory (bronchiolitis, pneumonia); Infectious Disease, Non-Respiratory (cellulitis, urinary tract infection, gastroenteritis); Asthma; Appendicitis; Diabetic Ketoacidosis; and Sickle Cell Disease with crisis.
Analysis
Associations between COI and hospitalization rates, overall and by diagnosis subgroup, were modeled using Poisson regression, adjusting for age, sex, and rurality (rural vs. urban).11 Race and ethnicity were not included due to inconsistent classification across the two databases. Incidence rate ratios (IRRs) comparing hospitalization rates in very low vs. very high opportunity neighborhoods were also calculated. A secondary analysis examined the effect of rurality on the observed associations between COI and hospitalization rates to explore potential disparities and inform future work. For this, the primary analysis was repeated after stratifying the cohort by rurality. This study was deemed exempt by the Vanderbilt Institutional Review Board. Analyses were conducted using SAS, version 9.4 (SAS Institute Inc).
RESULTS
The cohort included 614,823 hospitalizations among 29,244,065 children (21.02 hospitalizations per 1000 children, Table and Supplemental Table 1). Hospitalization rates decreased significantly and in a stepwise fashion as COI quintile increased (P<.001, Figure), ranging from 26.56 per 1000 children (95% CI 26.41, 26.71) in very low COI neighborhoods to 14.76 per 1000 children (95% CI 14.66, 14.87) in very high COI neighborhoods (IRR for very low vs. very high COI: 1.8, 95% CI 1.78, 1.81, Figure). This equates to an estimated 65,706 (95% CI 64 871, 66 541) excess hospitalizations annually for children in very low compared to very high COI neighborhoods. Similar patterns were seen across all diagnosis subgroups with IRRs ranging from 2.08 (95% CI 2.03, 2.13) for infectious disease, respiratory to 13.79 (95% CI 12.24, 15.55) for sickle cell disease with crisis. When stratified by rurality, similar decreases in hospitalization rates were noted as COI quintile increased (Supplemental Table 2), with a larger relative difference in urban (IRR for very low vs. very high COI: 1.82, 95% CI 1.81, 1.84) compared to rural areas (IRR for very low vs. very high COI: 1.54, 95% CI 1.48, 1.6).
Table.
State Inpatient Databases (SID) Patient-Level Demographics by COI 2.0 Child Opportunity Levels
| Overall | Nationally Normed COI 2.0 Child Opportunity Levels | |||||
|---|---|---|---|---|---|---|
| Very Low | Low | Moderate | High | Very High | ||
| SID Encounters, No. (%) | 614 823 | 160 018 (26) | 121 025 (19.7) | 121 268 (19.7) | 113 018 (18.4) | 99 494 (16.2) |
| Age, y, No. (%) | ||||||
| 0-4 | 242 580 (39.5) | 68 659 (42.9) | 49 422 (40.8) | 47 243 (39) | 42 255 (37.4) | 35 001 (35.2) |
| 5-9 | 96 904 (15.8) | 27 334 (17.1) | 19 365 (16) | 18 720 (15.4) | 16 837 (14.9) | 14 648 (14.7) |
| 10-14 | 138 771 (22.6) | 34 155 (21.3) | 26 628 (22) | 27948 (23) | 26 287 (23.3) | 23 753 (23.9) |
| 15-17 | 136 551 (22.2) | 29 864 (18.7) | 25 608 (21.2) | 27 353 (22.6) | 27 636 (24.5) | 26 090 (26.2) |
| Sex, No. (%) | ||||||
| Male | 317 335 (51.6) | 84 883 (53) | 63 019 (52.1) | 62 031 (51.2) | 57 389 (50.8) | 50 013 (50.3) |
| Rurality, No. (%) | ||||||
| Rural | 80 925 (13.2) | 21 931 (13.7) | 22 807 (18.8) | 19 932 (16.4) | 13 507 (12) | 2748 (2.8) |
| Urban | 533 898 (86.8) | 138 087 (86.3) | 98 218 (81.2) | 101 336 (83.6) | 99 511 (88) | 96 746 (97.2) |
| US Census Region of Hospital, No. (%) | ||||||
| Midwest | 120001 (19.5) | 20967 (13.1) | 13448 (11.1) | 22088 (18.2) | 31425 (27.8) | 32073 (32.2) |
| Northeast | 143479 (23.3) | 52800 (33) | 23153 (19.1) | 19420 (16) | 22092 (19.5) | 26014 (26.1) |
| South | 238862 (38.9) | 69646 (43.5) | 62107 (51.3) | 52805 (43.5) | 35005 (31) | 19299 (19.4) |
| West | 112481 (18.3) | 16605 (10.4) | 22317 (18.4) | 26955 (22.2) | 24496 (21.7) | 22108 (22.2) |
2017 SID states include: FL,GA, KY, MS, NC, AZ, CO, OR, UT, WA, NJ, NY, RI, WI, IA, MI, MN, NE
Figure.
Adjusted Hospitalization Rates per 1000 Children, Overall and by Clinical Diagnosis. X-axis represents COI 2.0 Child Opportunity Level quintiles from Very Low to Very High. All differences are significant at P<.001.
IRR = Incidence Rate Ratio (95% Confidence Intervals) comparing adjusted hospitalization rates between Very Low and Very High COI Opportunity Levels
a Infectious disease, respiratory includes acute bronchiolitis and pneumonia
b Infectious disease, non-respiratory includes cellulitis, urinary tract infections, and gastroenteritis
c Mood disorders includes major depressive disorders, mood disorders, bipolar disorders, suicide and intentional self-inflicted injury and adjustment disorders
e DKA=diabetic ketoacidosis
COI = Child Opportunity Index 2.0
DISCUSSION
In this population-based study across 18 US states, lower neighborhood opportunity was associated with higher hospitalization rates. Findings were consistent across 7 individual diagnosis groups, with the largest relative increases among children with asthma and sickle cell disease with crisis. Our findings support a link between neighborhood opportunity and inpatient care utilization across all hospital types, underscoring the importance of community investment to optimize population health and well-being.
These findings largely agree with a similar study that examined associations between COI and hospitalizations for ambulatory care sensitive conditions in two US metropolitan areas,4 but provides the first population-based analysis including all-cause hospitalizations at all hospital types. In both studies, the association between lower COI and higher hospitalization rates was evident, yet the prior study found higher overall hospitalization rates, most notably for asthma.4 Importantly, relative rate increases for asthma across the very low vs. very high COI neighborhoods were similar (IRR 4.6 in current study vs. 5.5 in Krager et al.).4 Rate differences between the studies may be due to heterogeneity in the underlying populations or types of encounters included (inpatient vs. observation hospitalizations). While Krager et al. included both inpatient and observation status hospitalizations, the SID does not reliably capture observation status encounters, resulting in an underestimation of overall hospitalization rates in our study. There is institutional variation in the use of observation vs. inpatient status, suggesting that any misclassification of hospitalizations is likely non-differential with respect to COI.12 The current study also used data from all acute care hospitals and was linked to census data to generate unbiased population-based rates across 18 states, whereas the Krager et al. study focused on the pediatric populations immediately surrounding one of two freestanding, academic children’s hospitals. These and other potential differences make direct rate comparisons difficult.
When stratified by rurality, associations between COI and hospitalization rates were similar to the main analysis, although the observed disparity was somewhat attenuated in rural areas. Reasons underlying these differences are likely multifactorial and were not explored within this analysis. A prior study suggests that differences in local inequity within metropolitan (i.e., urban) versus rural areas may partially explain these findings.13 Acevedo-Garcia et al. found that inequities in opportunity are larger within metropolitan areas than across the country, with 91% of variation in child opportunity occurring within metropolitan areas.13 The greater inequity in neighborhood conditions across metropolitan areas may drive the stronger relationship between neighborhood opportunity and hospitalization rates in metropolitan compared to rural areas.13 Future work is needed to better elucidate reasons underlying these differences.
Important disparities in hospital utilization for children driven by social and community factors are evident and underscore the need for partnerships between health systems, payors, community organizations, and public health entities to overcome these complex inequities. Children living in poverty are at higher risk for hospitalization for both acute and chronic conditions.14,15 Hospitalization for acute conditions such as gastroenteritis may be due to lack of access to outpatient primary care that leads to higher ED utilization and an increased risk for admission.16 Similarly, neighborhood-level factors (e.g., poor housing quality and environmental pollutants) directly influence the health of children with chronic conditions and lead to increased hospitalization rates.17–19 Prioritizing healthcare-community partnerships in areas of lower opportunity is one strategy to spurn innovation aimed at mitigating hospital utilization disparities and advancing child health equity.
The study has potential limitations. First, the SID only reliability includes inpatient hospital discharges likely resulting in an underestimation of hospitalization rates. Second, we were unable to further breakdown the composition of hospitals because the SID groups pediatric hospitals and academic medical centers together as “community” hospitals. Third, the magnitude of associations for asthma and sickle cell disease may be overestimated since the at-risk denominator population was unknown. Nonetheless, the directionality of association was consistent and significant and consistent across all diagnoses studied. Fourth, the data are from 2017, although we do not suspect that the observed trends differ markedly from today. Finally, hospitalization rates are influenced by factors beyond COI, such as individual or family health behaviors and health system capacity. The contributions of these and other patient- and population-level characteristics warrants further inquiry.
Higher hospitalization rates among children living in lower opportunity communities highlight important inequities and the need for continued investment and broad engagement to identify and eliminate their key determinants. The COI may help identify communities for targeted population-level interventions to improve child health and reduce healthcare utilization and costs.
Supplementary Material
Funding/Support:
Dr. Carroll was supported by grant number T32HS026122 from the Agency for Healthcare Research and Quality. The other authors received no additional funding.
Role of Funder:
The content is solely the responsibility of the authors and does not necessarily represent the official views of the Agency for Healthcare Research and Quality. The Agency for Healthcare Research and Quality did not participate in this work.
Abbreviations:
- COI
Child Opportunity Index
- ED
emergency department
- US
United States
- SID
State Inpatient Databases
- ACS
American Community Survey
Footnotes
Conflict of Interest Disclosures: The authors have no conflicts of interest to disclose.
REFERENCES
- 1.Noelke C, McArdle N, Baek M, et al. Childhood Opportunity Index 2.0 technical documentation. Published online 2020. Accessed February 2, 2023. chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://www.diversitydatakids.org/sites/default/files/2020-02/ddk_coi2.0_technical_documentation_20200212.pdf
- 2.Bettenhausen JL, Noelke C, Ressler RW, et al. The Association of the Childhood Opportunity Index on Pediatric Readmissions and Emergency Department Revisits. Acad Pediatr. Published online December 2021:S1876285921006306. doi: 10.1016/j.acap.2021.12.015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kaiser SV, Hall M, Bettenhausen JL, et al. Neighborhood Child Opportunity and Emergency Department Utilization. Pediatrics. 2022;150(4):e2021056098. doi: 10.1542/peds.2021-056098 [DOI] [PubMed] [Google Scholar]
- 4.Krager MK, Puls HT, Bettenhausen JL, et al. The Child Opportunity Index 2.0 and Hospitalizations for Ambulatory Care Sensitive Conditions. Pediatrics. 2021;148(2):e2020032755. doi: 10.1542/peds.2020-032755 [DOI] [PubMed] [Google Scholar]
- 5.Freyleue SD, Arakelyan M, Leyenaar JK. Epidemiology of pediatric hospitalizations at general hospitals and freestanding children’s hospitals in the United States: 2019 update. J Hosp Med. 2023;18(10):908–917. doi: 10.1002/jhm.13194 [DOI] [PubMed] [Google Scholar]
- 6.Healthcare Cost and Utilization Project (HCUP). HCUP Databases. Published online September 2021. Accessed February 28, 2023. https://www.hcup-us.ahrq.gov/sidoverview.jsp?_gl=1*sn4soy*_ga*MTQ3NjA2NDIzNi4xNjY5ODQyOTAx*_ga_1NPT56LE7J*MTY3NzYyMDQ4OS40LjAuMTY3NzYyMDQ4OS4wLjAuMA..
- 7.Coffey R, Barrett M, Steiner S. Final Report Observation Status Related to Hospital Records. Published online September 27, 2002. Accessed October 2, 2023. chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://hcup-us.ahrq.gov/reports/methods/FinalReportonObservationStatus_v2Final.pdf
- 8.Thomas AJ, Eberly LE, Davey Smith G, Neaton JD, Multiple Risk Factor Intervention Trial (MRFIT) Research Group. ZIP-code-based versus tract-based income measures as long-term risk-adjusted mortality predictors. Am J Epidemiol. 2006;164(6):586–590. doi: 10.1093/aje/kwj234 [DOI] [PubMed] [Google Scholar]
- 9.Lovasi GS, Moudon AV, Smith NL, et al. Evaluating options for measurement of neighborhood socioeconomic context: evidence from a myocardial infarction case-control study. Health Place. 2008;14(3):453–467. doi: 10.1016/j.healthplace.2007.09.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gill PJ, Anwar MR, Thavam T, Hall M, Rodean J, Mahant S. Pediatric Clinical Classification System for Use in Inpatient Settings. JAMA Pediatr. 2021;175(5):525–527. doi: 10.1001/jamapediatrics.2020.6007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Rural Health Research Center. RUCA Data. Accessed March 6, 2023. https://depts.washington.edu/uwruca/ruca-uses.php
- 12.Synhorst DC, Hall M, Macy ML, et al. Financial Implications of Short Stay Pediatric Hospitalizations. Pediatrics. 2022;149(4):e2021052907. doi: 10.1542/peds.2021-052907 [DOI] [PubMed] [Google Scholar]
- 13.Acevedo-Garcia D, Noelke C, McArdle N, et al. Racial And Ethnic Inequities In Children’s Neighborhoods: Evidence From The New Child Opportunity Index 2.0: Study uses the Child Opportunity Index 2.0 to examine geographic and racial/ethnic inequities children are exposed to in the one hundred largest metropolitan areas of the United States. Health Aff (Millwood). 2020;39(10):1693–1701. doi: 10.1377/hlthaff.2020.00735 [DOI] [PubMed] [Google Scholar]
- 14.Fieldston ES, Zaniletti I, Hall M, et al. Community Household Income and Resource Utilization for Common Inpatient Pediatric Conditions. Pediatrics. 2013;132(6):e1592–e1601. doi: 10.1542/peds.2013-0619 [DOI] [PubMed] [Google Scholar]
- 15.Herrod HG, Chang CF. Potentially Avoidable Pediatric Hospitalizations as Defined by the Agency for Healthcare Research and Quality: What Do They Tell Us About Disparities in Child Health? Clin Pediatr (Phila). 2008;47(2):128–136. doi: 10.1177/0009922807306166 [DOI] [PubMed] [Google Scholar]
- 16.Lowe RA, Fu R, Ong ET, et al. Community Characteristics Affecting Emergency Department Use by Medicaid Enrollees. Med Care. 2009;47(1):15. doi: 10.1097/MLR.0b013e3181844e1c [DOI] [PubMed] [Google Scholar]
- 17.Bryant-Stephens TC, Strane D, Robinson EK, Bhambhani S, Kenyon CC. Housing and asthma disparities. J Allergy Clin Immunol. 2021;148(5):1121–1129. doi: 10.1016/j.jaci.2021.09.023 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Goodman JE, Loftus CT, Liu X, Zu K. Impact of respiratory infections, outdoor pollen, and socioeconomic status on associations between air pollutants and pediatric asthma hospital admissions. PLOS ONE. 2017;12(7):e0180522. doi: 10.1371/journal.pone.0180522 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kanchongkittiphon W, Mendell MJ, Gaffin JM, Wang G, Phipatanakul W. Indoor Environmental Exposures and Exacerbation of Asthma: An Update to the 2000 Review by the Institute of Medicine. Environ Health Perspect. 2015;123(1):6–20. doi: 10.1289/ehp.1307922 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.