Childhood maltreatment involves dynamics between the type of maltreatment experienced and the context of maltreatment. Reports of alleged maltreatment to child protective services may overlap and shift over time, complicating understanding of their independent and interacting nature, including how child protection systems respond. Latent class analysis (LCA) and latent transition analysis (LTA) were used to construct data-based models of longitudinal dynamics of alleged maltreatment throughout childhood. We sought to identify patterns leading to system decisions to substantiate allegations of maltreatment and place children in foster care. Using linked birth and child protection records, we defined a cohort of children born in California in 1999, 29.4% of whom had at least one referral for alleged maltreatment before their 18th birthday. Maltreatment and perpetrator indicators were coded, and LCA identified five alleged maltreatment classes and one class of children without referrals. LTA determined consistency of classifications and estimated transitions between classes over age periods. Children with multitype maltreatment patterns or experiences of neglect were most likely to experience future maltreatment allegations. Estimated probabilities of placement indicated children with Multitype Maltreatment allegations were more likely to experience substantiated maltreatment allegations and out-of-home placements. Findings identify a repeatable method for better understanding complex systems.
Bottom-up approaches to strengthening child protection systems: Placing children, families, and communities at the center
