A Retrospective Examination of Child Protection Involvement Among Young Adults Receiving Homeless Services

BACKGROUND AND PURPOSE. Despite a well-established relationship between involvement with child protective services (CPS) during childhood and subsequent homelessness as an adult, efforts to align services across these two systems has been challenging. Among a range of issues that have hindered the coordination of public programs across sectors, certainly one is the absence of data which would allow for an understanding of when, where, and for whom these cross-system interactions may be most salient. In short, we are missing data assembled in a manner that can inform investments in programs and preventive interventions based on patterns of service delivery. Fortunately, the increasing availability of high-quality administrative case management data, coupled with advances in technology, now allow for the systematic integration of records from CPS systems, with records from homeless service programs. This analysis provides a population-level, epidemiological characterization of the child protection histories of young adults accessing homeless services using linked, cross-system administrative data.

METHODS. This analysis is based on administrative records from San Franciso County's Homeless Managment Information System (HMIS) linked to statewide CPS records falling under the authority of the California Department of Social Services. HMIS records concerning all young adults between the ages of 18-24 who were waitlisted or received homeless services between 2012 and 2014 were extracted, cleaned, and standardized (n=3,792). These records were then probablistically linked to more than 1.5 million child protection records dating back to 1998. Records were linked based on a combination of unique (e.g., SSN) and non-unique (e.g., first name, last name, date of birth) personal identifiers common to both files. Chi-square tests were used to provide a descriptive comparison of the demographic characteristics of homeless youth with and without a CPS history. Multivariable models (log link, poisson distribution) were specified to differentiate among demographic factors. Among homeless young adults with a CPS history, we additionally examined the nature of earlier CPS interactions.

RESULTS. More than 50% of the population of young adults waitlisted or receiving homeless services in SF County between 2012 and 2014 had a history of CPS involvement. Rates of past CPS involvement were modestly higher among young adults who were waitlisted, but did not receive services (61% vs. 57%; p<.05). Being female and African American were associated with an increased likelihood of a documented history of reported abuse and neglect, both before and after adjusting for other factors (p<.05). Overall, more than 40% of homeless young adults had a CPS history originating in another county, although a majority fell within the broader Bay Area region.

CONCLUSIONS AND IMPLICATIONS. This analysis provides a epidemiological characterization of the child protection histories of young adults accessing homeless services using linked, cross-system administrative data. These data expand the evidence base to support the coordination of policies and programs between child protection and homeless service sectors. Importantly, this analysis also documents the potential for generating critical, population-level data concerning young adults accessing homeless services through linked data.