Building an effective casework system for child maltreatment is a global issue. We estimated the effect of household dysfunction (i.e., interparental violence, caregiver mental health problems, and caregiver substance abuse) on child maltreatment to understand how to advance the current framework of child welfare.
The sample comprised 759 children (1- to 17-year-old; mean age was 10.6; 404 boys and 355 girls) placed in temporary custody units (one of the strongest intervention of the Japanese child protection system). Caseworkers from 180 units across 43 prefectures completed questionnaires on children and their family and were asked whether a child maltreatment report had been made after cancelation of custody in a 15-month follow-up period. The relations of household dysfunction and maltreatment reports were assessed using the Cox proportional hazard model.
About half (48.4%) of the children had been placed in the unit because of maltreatment, and 88.3% had a history of victimization. Seventy-six cases had maltreatment reports after cancelation. We entered household dysfunction variables individually into the model, and each had a significant relationship with maltreatment reports (hazard ratios for interparental violence, caregiver mental health problem, and substance abuse were 1.69, 1.69, and 2.19, respectively) after covariate adjustment. When treating these three variables as cumulative risk score model of household dysfunction, the hazard ratio increased with increasing number of score (1.96 for score two; 2.35 for score three; score 0 as reference).
Greater household dysfunction score is a risk of maltreatment after intensive intervention. It is imperative to construct systems facilitating cooperation between child and adult service sectors and to deliver seamless services to children and families. Our findings provide child protect services with risk-stratified interventions for children at victimization risk and promote adult-focused services to be proactive in prevention or intervention for adults with perpetration risk.