Do Higher Foster Care Placement Rates Prevent Child Maltreatment Fatalities, or Do Rising Fatalities Drive More Placements?

Background/Purpose

Child Protective Services (CPS) aims to prevent maltreatment recurrence and escalation, including child maltreatment fatalities (CMFs). Foster care (FC) placement is the most intensive child safety intervention available within CPS. Higher FC placement rates may help prevent CMFs, but CPS systems may also respond to rising fatalities by increasing use of intensive interventions. This study examined these potentially bidirectional associations using state-level panel data. First, we assessed whether higher concurrent (t) and lagged (t−1) FC placement rates were associated with lower current CMF rates (FC-to-CMF model). Second, we examined whether higher concurrent and lagged CMF rates were associated with higher current FC placement rates, indicating a possible reactive system response (CMF-to-FC model).

Methods

We constructed a state-year panel for all 50 U.S. states and the District of Columbia from 2000 to 2024. For the FC-to-CMF model, FC placement rates (per 1,000 children) were decomposed into within-state variation (annual deviations from each state’s long-term average) and between-state variation (state-specific long-term averages). Concurrent and lagged within-state FC placement rates were categorized as “average” (approximately 25th–75th percentile), “higher” (>75th percentile), or “lower” (<25th percentile) relative to each state’s own historical distribution. The same approach was used in the CMF-to-FC model for CMF rates (per 1,000,000 children). We estimated linear mixed-effects longitudinal panel models with state random intercepts and an AR(1) correlation structure. Models adjusted for within- and between-state CPS report rates, substantiation rates, and key socioeconomic and demographic characteristics.

Results

In the FC-to-CMF model, neither higher concurrent nor higher lagged FC placement rates were significantly associated with lower CMF rates. Likewise, lower concurrent or lagged FC placement rates were not significantly associated with higher CMF rates. In the CMF-to-FC model, concurrent CMF rates were not significantly associated with FC placement rates. However, higher lagged CMF rates were associated with modest increases in current FC placement rates. Specifically, when a state’s prior-year CMF rate was higher than its usual level (approximately >75th percentile), the current-year FC placement rate increased by 1.95 per 10,000 children (95% CI = 0.74–3.16), or about a 5% increase relative to the mean FC rate. Lower lagged CMF rates were not significantly associated with FC placement rates.

Conclusions/Implications

We found no evidence that higher FC placement rates were associated with lower CMF rates at the state level. In contrast, elevated CMF rates in the prior year were followed by modest increases in FC placements, suggesting that CPS systems may respond to worsening child safety outcomes with greater use of intensive interventions. These findings do not establish causality and should not be interpreted as evidence for or against the effectiveness of individual FC placements. Rather, they suggest that broad shifts in placement intensity alone may be insufficient to reduce fatalities. Efforts to prevent CMFs may require earlier, targeted prevention strategies, stronger service quality, and closer attention to the circumstances preceding severe harm.