Related Experiment Video
Updated: Apr 28, 2026
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
Precision risk assessment for pediatric hospitalization using address-level data in Cincinnati, Ohio
Carson S Hartlage1,2, Qing Duan3, Erika Rasnick Manning3
1Department of Biostatistics, Health Informatics and Data Sciences, University of Cincinnati College of Medicine, Cincinnati, OH, USA.
Insights
This study precisely identified pediatric hospitalization risks by linking health data with neighborhood factors at the address level. Findings support targeted interventions for child health disparities.
Area of Science:
- Environmental Health
- Public Health
- Health Informatics
Background:
- Persistent child health disparities necessitate research with enhanced spatial precision.
- Traditional area-level analyses may not capture localized health risks effectively.
Purpose of the Study:
- To link socio-environmental data to healthcare data for precise pediatric hospitalization risk assessment at the address level.
- To develop and validate models for identifying high-risk residential addresses for children.
Main Methods:
- Linked hospitalization data (2016-2022) with parcel-level housing, crime, census, and eviction data in Cincinnati, Ohio.
- Utilized generalized random forest models to estimate address-level and birth-adjusted hospitalization risks.
- Assessed model performance using ROC-AUC and PR-AUC, and evaluated fairness by racial demographics.
Main Results:
- Successfully matched 81.5% of hospitalizations to residential addresses.
- Models accurately characterized high-risk addresses, with housing violations, violent crime, and property value as key predictors.
- The birth-adjusted model demonstrated high performance and moderate agreement with the primary hospitalization risk model.
Conclusions:
- Address-level modeling and multiscale data integration advance precision population health beyond traditional methods.
- This scalable approach precisely identifies pediatric health risks, informing targeted clinical and policy interventions.
- Future work includes geographic expansion, stakeholder engagement, and patient-level validation.
Introduction:
Persistent disparities in child health highlight the need for clinical and public health research approaches to identify and address risks with greater spatial precision. This study linked residence-and neighborhood-specific socio-environmental data to population-wide healthcare data to characterize pediatric hospitalization risk for every residential address in Cincinnati, Ohio.
Methods:
We linked hospitalization data (07/01/2016-06/30/2022) to parcel-level housing data from the Hamilton County Auditor and Cincinnati Department of Buildings & Inspections and street-range crime data from the Cincinnati Police Department. Addresses were localized to 2010 census tracts to join variables from the US Census American Community Survey and Eviction Lab. Generalized random forest models estimated address-level hospitalization risk and birth-adjusted hospitalization risk, accounting for child residency using vital birth records. Model performance was assessed based on varying diagnostic thresholds; fairness was evaluated by census block-level racial demographics.
Results:
We matched 81.5% of hospitalizations to residential addresses. Among 77,077 addresses, 7.4% had ≥1 hospitalization. Our model performed well (ROC-AUC: 0.98-0.99; PR-AUC: 0.65-0.72) in characterizing high-risk addresses, with housing violations, violent crime, and market total value among top features. The birth-adjusted model also showed high performance (ROC-AUC: 0.92-0.93; PR-AUC: 0.65-0.78) and moderate agreement with the hospitalization risk model (κ = 0.43).
Conclusions:
Our results highlight the potential of address-level modeling and multiscale data integration to build on traditional area-level analyses and advance precision population health. Future directions include geographic expansion, stakeholder engagement, and patient-level validation. This work offers a scalable approach to precisely identifying pediatric health risks, supporting targeted clinical and policy interventions.
