Can Two-Generation Approaches Improve Child Welfare Outcomes? Evidence on Safety and Permanency

Background/Purpose: Critical measures of success for child welfare agencies are reducing the risk of maltreatment recurrence and preventing the removal of children from the biological family home. Perhaps because many interventions with child welfare involved families are narrowly focused on individual parenting behaviors, success in the domains of safety and permanency remains limited. The current study rigorously evaluates an intervention designed to address a broad constellation of risk factors associated with maltreatment and foster care placement. Rather than focusing only on parenting behaviors, the Two-Generation (2Gen) approach seeks to improve safety and permanency outcomes by promoting protective factors and improving economic conditions by supporting and engaging both caregivers and their children.

Methods: The sample was comprised of 2,945 children associated with the 2Gen program and 9,053 children associated with other prevention-oriented programs serving similar families. Propensity Score Matching (PSM) was used to account for potential differences between the treatment and comparison groups, differences that could bias the estimates of both safety and permanency outcomes. Post matching, there were no significant differences between the treatment (2,945) and the comparison (2,945) groups in terms of age at referral, reason for referral, prior maltreatment history, race or sex. The safety outcome was defined as a substantiated report of maltreatment subsequent to program referral. Permanency was defined as the placement of the child in foster care subsequent to program referral. Data were analyzed using Cox proportional hazards on the matched sample. Cox regression is well-suited here because children enter observation at different times (e.g., referral date) and thus are observed for different durations. This method appropriately accounts for right-censoring and unequal exposure time rather than treating all children as if they had identical follow-up.

Results: Cox proportional hazards models indicated that participation in the MiFamily program was associated with a lower hazard of subsequent maltreatment (HR = .859, p = .030) and subsequent foster care placement (HR = .767, p = .019). Age of the child at the time of referral was associated with decreased hazard of both outcomes (Safety: HR = .950, p < .001; Permanency: HR = .922, p < .001). Children referred to prevention services for physical abuse were significantly less likely to experience a subsequent foster care placement in comparison with children referred for neglect (HR = .652, p = .008). Children with a history of maltreatment complaints were significantly more likely to be removed and placed in foster care (HR = 5.371, p < .001).

Conclusions/Implications: The findings indicate that the 2Gen approach significantly improves child safety and permanency outcomes, underscoring the potential of the 2Gen approach to provide integrated support for both children and caregivers to promote child well-being. These findings highlight the value of coordinated, two-generation approaches in advancing child welfare practice and supporting family well-being. Findings may inform the development of interventions that better address the complex needs of families and enhance coordinated service delivery across domains. Future research should examine underlying mechanisms and assess whether these effects are equitable across populations.