Abstract: Predicting Family Progression through Voluntary Prevention Services: A Sequential Analysis of Service Uptake and Engagement (Society for Social Work and Research 31st Annual Conference)

Predicting Family Progression through Voluntary Prevention Services: A Sequential Analysis of Service Uptake and Engagement

Schedule:
Sunday, January 24, 2027
Golden Gate 3 (Hilton San Francisco Union Square)
* noted as presenting author
John Prindle, PhD, Research Associate Professor, University of Southern California, CA
Lindsey Palmer, PhD, Assistant Professor, University of Utah, Salt Lake City, UT
Lacey Jenkins, PhD, Assistant Professor, University of Alabama, Tuscaloosa, AL
Claire McNellan, PhD, Assistant Professor, University of Colorado Anschutz, Aurora, CO
Jacquelyn McCroskey, PhD, Professor, University of Southern California, Los Angeles, CA
Background/Purpose: Family engagement in voluntary child welfare prevention services unfolds across multiple stages, from initial outreach through service completion. Yet most existing research treats engagement as a binary outcome, obscuring the distinct barriers and facilitators that operate at each stage of the service continuum. Understanding which family and referral characteristics predict progression at each point is essential for designing targeted strategies to reduce attrition and improve service reach. This study identifies and describes a large cohort of families offered voluntary prevention services following a child welfare referral in Los Angeles County and examines how demographic and referral characteristics predict the likelihood of advancing through each sequential stage of service engagement.

Methods: Prevention and Aftercare (P&A) program data from Los Angeles County (N = 24,404) were merged with child welfare referral data. Generalized linear models were used to examine predictors of progression through three sequential stages of the service cascade: (1) successful contact versus no contact; (2) among those reached, agreement to services versus refusal (n = 16,903; 69.3% of the full sample); and (3) among those who agreed, successful program completion versus non-completion (n = 11,387; 46.7% of the full sample).

Results: Greater neighborhood-level racial and ethnic heterogeneity was associated with lower odds of successful contact (OR = 0.80, 95% CI: 0.74–0.86) and lower odds of service agreement among those reached (OR = 0.46, 95% CI: 0.42–0.50), but higher odds of successful completion among those who agreed (OR = 2.45, 95% CI: 2.22–2.70; all p < .001). Compared to evaluated-out referrals, investigated referrals were associated with modestly higher odds of progression at the contact and agreement stages (ORs: 1.06–1.11) but lower odds of successful completion (ORs: 0.86–0.90; all p < .001).

Conclusions/Implications: Predictors of early-stage contact do not uniformly predict later-stage success, underscoring the importance of modeling engagement as a sequential process rather than a single outcome. Community-level racial and ethnic diversity appears to shape engagement differently across service stages, suggesting that contextual and structural factors warrant greater attention in program design. Referral characteristics similarly shift in their influence across the cascade, pointing to the need for stage-specific outreach and retention strategies. These findings have direct implications for how agencies allocate resources and tailor engagement efforts across the prevention service continuum.