Child Maltreatment Reports and Investigation Outcomes during the COVID-19 Pandemic

Background & Purpose:

Many states have identified that reports to child protective services (CPS) dropped with the stay-at-home orders and the closure of in-person schooling related to the COVID-19 pandemic. However, there is little information regarding if those drops in reports have been consistent across all reporter types. Further, it is unknown if CPS responses to child abuse and neglect reports have changed since the onset of the pandemic. Thus, the purpose of this study was to identify if there were changes after COVID-19 in-person school closures regarding the count of CPS reports by reporter type, screening decisions of CPS reports, and allegation investigation outcomes.

Methods:

We included all CPS administrative records from the State of Washington from January 2016 through October 2020 in the analysis, including all CPS reports during this period (N = 720,178). We examined trends before and after COVID-19 for three outcomes: the count of intakes, screening decision (screened in or out), and allegation investigation outcome (founded or unfounded). Specifically, we plotted weekly counts of CPS intakes, percent of intakes screened in, and percent of allegations founded over time. Interrupted Time Series (ITS) models were run for each outcome, controlling for seasonality, and stratified by reporter category (education, health, law enforcement, social service, and non-mandated). We subsequently conduct Chi-squared tests to determine differences in the screening and substantiation rates for relevant reporter types.

Results:

Counts of intakes dropped dramatically in mid-March 2020, coinciding with the pandemic-related in-person school closures. However, when we disaggregated the results by reporter type, we found that the reduction in reports by mental health professionals and educators accounted for almost all intake reduction. Our ITS models confirmed a statistically significant drop in intakes controlling for the factors mentioned above (p < 0.001).

Subsequent Chi-squared analyses revealed that only reports from educators had a corresponding reduction in the likelihood of screening or substantiation. Specifically, we identified a 15% to 17% reduction in the screening rate of reports from educators during the pandemic compared to trends before the pandemic (p < 0.001). We observed no meaningful differences in the screening rates of any other reporter type and no differences in the substantiation rates of any reporter type (including reports from educators).

Conclusions:

The reduction in both intakes and screening rates from educators is similar to patterns observed in unpublished trends from other parts of the country. This shift suggests that the change in intakes was both quantitative and qualitative. Specifically, it indicates that the system received intakes from educators that the state assessed, on average, lower risk than the educator intakes received before the pandemic. In conjunction with the lack of significant change in the substantiation rate for educator intakes, we could interpret this as an indicator of a reliable screening mechanism in the state child welfare system. Future research should explore additional case-level details to better understand the system’s ability to reliably assess risk in the context of the resource constraints imposed by the pandemic.