Abstract: The Incidence and Risk Factors of Child Maltreatment-Related Hospitalizations: A Population-Based Study (Society for Social Work and Research 24th Annual Conference - Reducing Racial and Economic Inequality)

The Incidence and Risk Factors of Child Maltreatment-Related Hospitalizations: A Population-Based Study

Friday, January 17, 2020
Marquis BR Salon 12, ML 2 (Marriott Marquis Washington DC)
* noted as presenting author
Rebecca Rebbe, MSW, PhD Student, University of Washington
Joseph Mienko, PhD, MSW, Senior Research Scientist, University of Washington, Seattle, WA
Melissa Martinson, PhD, MSW, Associate Professor, University of Washington, Seattle, WA
Background and Purpose: It is estimated that 1,720 children died in 2017 as a result of child maltreatment (CM), underscoring that in addition to poor psychological and behavioral outcomes, death and serious physical injuries are also products of CM. However, our knowledge regarding CM resulting in hospitalization is limited. Previous research has utilized standardized medical records to identify CM-related hospitalizations, but most of this work has relied on samples of hospitals containing cross-sectional sociodemographic variables. The reliance on these single data sources has restricted our understanding of the incidence rates and risk factors of CM that results in hospitalization. An approach to overcome the deficiencies of a single data source is utilizing population-based linked administrative datasets.

Therefore, the aim of this study was to employ a population-based linked administrative dataset to calculate the incidence and ascertain risk and protective factors of CM-related hospitalizations under the age of three for the population of a U.S. state.

Methods: Birth, hospitalization, and child protective services (CPS) administrative records were linked for all children born in Washington state between 1999 and 2013 (N = 1,271,416). ICD-9 diagnostic codes were used to identify hospitalizations related to CM, utilizing codes specifically related to CM and codes identified as suggestive of CM by Schnitzer et al. (2011).

The incidence was calculated using the entire population born in the state as the denominator. Risk and protective factors were identified using two methods. First, hierarchical linear modeling was employed to test community-level poverty/disadvantage simultaneously with sociodemographic variables from the birth record. Second, Cox proportional hazards models were used to test a prior CPS report as a risk factor while controlling for sociodemographic variables.

Results: A total of 6,944 CM-related hospitalizations were identified for a rate of 5.46 hospitalizations per 1,000 births. The most frequent subtype was for neglect at 5,272 (rate of 4.15 per 1,000 births), which is 75.9% of the total of hospitalizations. From the hierarchical linear modeling, maternal residence at the time of birth in a zip code with high concentrated disadvantage was identified as a risk factor for CM hospitalization (OR: 1.21, 95% CI: 1.09-1.35), in addition to low infant birth weight, teenaged mother, and male infant sex. A prior CPS report was also indicated as a risk factor (HR: 2.21, 95% CI: 2.05-2.37), a finding that was consistent on sub-analyses across the different subtypes of CM.

Conclusions and Implications: These results add to our knowledge about the incidence of CM, which is enhanced by the use of a population-based dataset. They provide utility for the targeting of prevention programs at birth because the risk and protective factors are identifiable at birth at both the individual and community levels. The finding that a prior CPS report is a risk factor for CM resulting in hospitalization is consistent with previous work finding prior CPS reports to be a risk factor for injury mortality and that it is an independent indicator of risk (Putnam-Hornstein, 2011).