A Latent Class Analysis of Infants Reported for Maltreatment

Summary

Infants are at particularly high risk of child abuse and neglect, which can negatively affect development in the short and long term. Identifying infants who are at increased risk of maltreatment re-reporting would help the child protective services (CPS) system more accurately assess initial reports, allowing it to better tailor interventions and target supports. Children reported to CPS, however, rarely have just one risk factor present; they often have multiple risk as well as protective factors. And these various factors are constantly interacting, meaning that a variable may be a risk factor for one group of children but not for another. For this reason, identifying the defining characteristics of a group of children that goes on to have future reports of alleged maltreatment may be useful to ensuring resources are properly allocated.

The purpose of this study was to move from a focus on individual risk and protective factors to an examination of subgroups of infants with different risks of a follow-up maltreatment report during the first 5 years of life, a time of critical developmental importance. Birth records for all children born in California in 2006 were linked to statewide child protection records through 2012. The outcome of interest was a CPS report within 5 years of an initial report during infancy. Seven family and infant characteristics known to be associated with CPS involvement and four characteristics of the initial report were modeled. Latent class analysis (LCA) is an analytic approach that allows researchers to model data to define 'classes' or subgroups of individuals that share characteristics. In the present analysis, LCA was used to identify classes of infants with variable risk of a follow-up maltreatment report.

Overview of Findings

Despite relatively high rates of re-reporting generally, the LCA analysis identified four distinct subpopulations that varied significantly in their probability of a future CPS report during the first 5 years of life, from 44% to 78%. This indicates LCA can be useful when assessing the risk of re-report among infants with an initial report of maltreatment. The findings suggest that infants in Class 1 had a relatively low probability of re-report in comparison to the medium- and highest-risk classes, despite the presence of factors generally associated with CPS reports. Specifically, infants in Class 1, one of the two lowest-risk groups, had risk factors including the lowest level of maternal education and the highest rate of public health insurance utilization (proxies for lower income). But infants in this class also had higher rates of protective factors, like paternity established at birth.

Two birth risk factors clustered in families in the medium and highest-risk groups: lack of established paternity and delayed or absent prenatal care. Two factors from the initial CPS report (i.e., an initial allegation of neglect and a family history of CPS involvement) also strongly defined classes of infants with very high risk of re-report within the first 5 years of life. The clustering of these two risk factors was associated with a 78% chance of re-report during this period. This study demonstrates that in an already high-risk subpopulation of children (i.e., infants with an initial report of alleged maltreatment), data can help identify classes of infants that are more likely to be re-reported and may be in greater need of preventive supports and services.

Products

CDN Latent Class Analysis - CA Version
CDN Latent Class Analysis - LA Version
Investigators
  • Andrea Lane Eastman, MA
  • Michael N Mitchell, PhD
  • Emily Putnam-Hornstein, PhD
Funders
Data and Research Partners
Timeline
  • May 2015 - December 2015
1150 South Olive Street, Suite 1400
Los Angeles, CA 90015
General Questions
Connect
Newsletter
Sign up for our mailing list to stay up to date with the Children's Data Network's latest research and news developments.

©2019 Datanetwork.org
Designed by House Helios