Citation

Williams J, Petersen N, Stoler J. BMC Public Health 2018; 18(1): e1085.

Affiliation

Department of Public Health Sciences, Miller School of Medicine, University of Miami, 1120 NW 14th St, Miami, FL, 33136, USA.

Copyright

(Copyright © 2018, Holtzbrinck Springer Nature Publishing Group - BMC)

DOI

10.1186/s12889-018-5985-5

PMID

30170574

Abstract

BACKGROUND: Routine screening and intervention for intimate partner violence (IPV) in healthcare settings constitutes an important secondary prevention strategy for identifying individuals experiencing IPV early and connecting them with appropriate services. Considerable variation in available IPV-related healthcare services exists and interventions are needed to improve the quality of these services. One way to prioritize intervention efforts is by examining the level of services provided in communities most at risk relative to local incidence or prevalence of IPV. To inform future interventions, this study examined the spatial relationship between IPV-related healthcare services and IPV arrests in Miami-Dade County, Florida, and identified predictors of the observed spatial mismatch.

METHODS: Survey data collected in 2014 from 278 health facilities pertaining to IPV services were geocoded, computed into a density layer, and aggregated at the census tract level to create a population-based normalized comprehensiveness score (NCS) as a proxy for IPV-related healthcare resources. IPV arrests from 2011 to 2015, collected from the county court, were geocoded and summarized by census tracts to serve as a proxy for IPV prevalence. These measures were combined into a resource disparity score (RDS) that compared relative service density to relative arrest rates, where positive RDS represented over-resourced neighborhoods and negative RDS corresponded to under-resourced neighborhoods. We used correlation analyses and a two-phase spatial modeling approach to evaluate correlates of NCS and RDS.

RESULTS: A spatial lag model did not yield an association between NCS and IPV arrests, demonstrating a spatial mismatch, which we visualized using a Geographic Information System (GIS). A spatial error model revealed that the percentage of white non-Hispanic residents was positively associated with RDS, while percent black non-Hispanic, median age, ethnic heterogeneity, and economic disadvantage were negatively associated with RDS.

CONCLUSIONS: These findings underscore the need to further evaluate the adequacy of IPV-related healthcare resources for secondary prevention relative to local IPV arrest rates, particularly within economically disadvantaged neighborhoods. Our approach demonstrates the utility of GIS for identifying potential priority regions for IPV prevention efforts and resource allocation.

Language: en

Keywords

Domestic violence; Geographic information systems; Health services accessibility; Healthcare disparities