Background: The Illinois Department of Children and Family Services Office of the Inspector General (OIG) investigates child deaths in families already known to the Department. Impetus for this project was the need for a more systematic and reliable approach to uncovering multi-level causes of case decision errors leading to child deaths.

Methods: The method chosen for examining these cases, RCA, was originally designed to reveal factors contributing to adverse outcomes in other high risk enterprises (e.g., chemical factory explosions, airline crashes, and nuclear power plant accidents). RCA allows the user to build a model of contributing factors on a case-by-case basis, conduct a quantitative analysis, and use the results to improve organizational performance at the individual, management, and policy levels. REASON was selected as the software tool because it can be used to guide child death investigations and produce parsimonious results that preserve rich contextual detail. Key elements include: 1) defining a problem statement, 2) creating a causal tree, 3) logic testing, 4) termination of cause-effect chains, and 5) identification of prevention principles. RCA moves the investigation of child deaths beyond placing individual blame to the identification of system problems contributing to decision errors (e.g., organizational culture, policies, procedures, supports, constraints and incentive structure). We can determine the proportion of outcome explained by each root cause, permitting solutions to be selected that will have the greatest impact. For each case, a reliability value is calculated describing the proportion of cause that remains unanswered by the model. For pools of cases, common error patterns are revealed leading to recommendations that affect the largest number of cases.

Implications: Not only does RCA offer analytic strategies, it is also a disciplined approach for guiding the investigation of case decision-making. The structured format provides uniformity in the manner in which events are identified and analyzed, contributing to highly reliable investigations. The multi-level approach reveals the impact of policies, rules and procedures, and individual actions on child abuse investigations. The ultimate goal is to design an organizational environment that produces better outcomes for children. This project represents an initial step toward understanding error-producing conditions in child protection services and provides useful strategies for the study of decision errors in human services more generally.