Capstone Abstract--
Eron Y. Friedlaender

Substandard Housing and Risk of Injury in Children

Eron Y. Friedlaender, MD, Jill C. Posner, MD, MSCE, Amy Hillier, PhD, Charles C. Branas, PhD, J. Felipe Garcia-Espana, PhD, and Flaura K. Winston, MD, PhD
December 2005

Abstract
Background: The association between substandard housing conditions and injury risk is undefined.
Objective: To identify remedial substandard housing conditions that may reduce injury risk.
Design/Methods: Nested case-control study within a retrospective cohort from an urban hospital-based primary care center (PCC) network with residence-level linkage to a database of housing conditions. PCC patients < 18 years old with injury diagnoses (defined by ICD-9 and E-Codes) between January and October 2004 were identified. Cases of abuse and intentional self injury were excluded. Controls were randomly selected from PCC patients with no visits for injuries in the 2 years prior to enrollment. Severity of injury was approximated as a secondary outcome measure using Injury Severity Scores (ISS). Each residence was linked to a property-level database of housing conditions maintained by the Cartographic Modeling Laboratory, University of Pennsylvania. Principle exposures of interest included open code violations, suspended utility service, or proximity to imminently dangerous housing (homes pending demolition, vacant lots, or high crime areas).
Results: The cohort included 1031 cases of injury and 4582 controls. Among cases, 83% of injuries were unintentional; 8% sustained a clinically significant injury (ISS>4). Many substandard housing conditions existed within the cohort: almost 50% of subjects lived close to a vacant lot, 75% lived close to a vacant residence, and 84% had 2 or more exposures to crimes against property. Bivariate analysis demonstrated associations between injury risk and proximity to a vacant lot (p = 0.04) and open code violations (p = 0.05). Further analysis revealed unadjusted associations between unintentional injury risk and open code violations (p = 0.03) and crimes against property (p = 0.05). Adjusted logit models demonstrated an increased risk of injury for children living with an existing open code violation (OR 1.19, 95% CI 1.02, 1.38) or in proximity to a vacant lot (OR 1.16, 95% CI 1.02, 1.33).
Conclusions: Investigation into the nature of the association between injury risk and open code violations, crimes against property, and proximity to vacant lots is needed to inform advocacy efforts.