What Counts? the Use of Structured Decision-Making in Child Welfare

Purpose:

            Improving decision-making in social work represents a grand challenge for the field. The Structured Decision-Making Model (SDM) has been implemented in over 30 states to make decisions more objective and empirical.  This study asks what are the intended and unintended consequences of using actuarial-based risk assessments to frame decision-making in child welfare? While methodological studies of the SDM have been conducted, very little research exists on how these models are used in practice. This research seeks to fill a gap in our understanding of how evidence-based practices are translated on the ground, in the context of the complexities of child welfare practice.

Methods:

         Data from this study draws from observation, analysis of state policies and decision-making procedures and 35 semi-structured interviews with CPS workers. Interviews were audio-taped and transcribed with analysis conducted in accordance with the principles of grounded theory. Using Nvivo 10, interviews and primary source materials were coded thematically by a research team to uncover the interactions between policy, informal practice, and the SDM in setting case trajectories.

Findings:

            Findings demonstrate that in the service of avoiding type 1 errors (leaving a child in the home who should have been removed), type 2 errors (removing a child who was safe to remain at home) may be more likely to occur. Data suggests there are two factors that contribute to this likely overestimation of risk: (1) state policy that explicitly ties the score on the risk assessment to a case’s trajectory without providing any formal procedures to lower the risk score and (2) the collection of information only on the presence or absence of risk factors without a more dynamic understanding of the relationship these factors have to the presenting complaint. Some examples include cases that are assigned a high-risk score based on demographic information (such as number of children and number of past complaints), whether the parent was maltreated as a child, and whether the family has a history of domestic violence without relating these factors to the current investigation.

Implications:

            The very real need for accountability, to reduce the risk of making an incorrect decision, and to address the variability of authority-based decisions facilitates not only the adoption of actuarial-based tools but also policies that limit ecological evaluation of cases. While use of actuarial-based risk assessments reduce some forms of error, this research demonstrates that without evaluation of case context, they may also cause one type of preventable error. Several policy recommendations are made based on these findings: (1) procedures should be introduced to permit workers to both raise and lower the risk level of a case. (2) Formal committees of experts should be established to review these override requests. (3) The risk assessment should be updated to collect information not just on the presence or absence of risk factors but also on their relationship to the current complaint.  This research addresses one grand challenge for social work: how evidence-based practices can be adapted to account for the nuances of real world settings.