Beyond prediction: Advancing risk assessment and addressing the evidence deficit in the statutory child maltreatment...

Structured professional judgement (SPJ) tools are increasingly used in statutory child welfare contexts, yet evidence for their predictive validity remains limited.

To evaluate whether two SPJ instruments-the Child Abuse Risk and Resilience Assessment (CARRA) CARRA and the Child Protection Removal Assessment (Chi-PRA)- adequately predicted subsequent judicial outcomes in court-involved families with child maltreatment related concerns.

Archival data from 433 parent-child dyads assessed at the Victorian Children's Court Clinic between (2017-2022) in Australia, were deterministically linked with court records to examine post-assessment order breaches of court orders and escalation to more restrictive orders.

Binary logistic regression models examined associations between risk classifications and post-assessment breach and escalation outcomes.

The most consistent associations were observed for order escalation. The CARRA short-term (CAT-1) and long-term (CAT-2) risk ratings and the Chi-PRA rating were associated with higher odds of escalation, with moderate- and high-risk classifications showing substantially elevated odds relative to low risk. These associations were robust across sensitivity analyses; however, overall predictive performance was modest, with limited discrimination and calibration. In contrast, the CARRA severity judgement (CAT-3) was not reliably associated with escalation, and neither tool demonstrated consistent associations with breach outcomes.

These findings underscore the importance of evaluating validity using multiple performance indices. While SPJ tools may support structured reasoning in statutory contexts, claims about their predictive capabilities should be made cautiously and transparently, with explicit attention to base rates, calibration, and outcome choice.