A System Dynamics Simulation of Three Child Neglect Prevention Interventions Utilizing a Group Model Building Approach

Introduction: Bronfenbrenner’s ecological theory of child development highlights the influence of multiple levels of factors affecting child well-being, but most interventions only focus one or two levels at a time. Further, most interventions are delivered in contexts that differ from those in which they were developed, leaving decision makers to make educated guesses about how the interventions might perform in their contexts. While guidelines and checklists exist, rigorous, evidence-based decision making tools are needed to help decision makers choose the programs that are most likely to have significant population level health impact and are aligned with community context. The purpose of this study was to develop an evidence-based system dynamics model that could simulate 1) the complexity of risk and protective factor interactions across the ecological model of child development and 2) the effectiveness of three evidence-based child neglect prevention programs: Nurse Family Partnership, Triple P, and Family Foundation.

Methods: Utilizing a Group Model Building (GMB) approach to engage eight stakeholders, we developed a systems-informed framework for child maltreatment and a quantitative system dynamics simulation model. The model simulated effects of the three evidence-based interventions on the number of families at low or high risk, the number of parents or caregivers receiving trauma treatment services, and the predicted number of child neglect incidents based on the proportion of families at high or low risk. We explored how outcomes varied by the adult caregivers’ trauma history and trauma treatment status. We parameterized the model using data from the National Survey of Child and Adolescent Welfare waves I and II (NSCAW). Tests of structural validity, behavior reproduction, and robustness were also performed.

Results: Simulation experiments suggested key risk and protective factors such as maternal stress, housing insecurity, trauma provider availability, and parent peer support should be targeted for intervention. Stakeholders lamented the lack of peer support specialists that were utilized in the interventions.

Conclusion: We find that the reach and effectiveness of parenting interventions to prevent child neglect are influenced by factors across the socio-ecological levels. Stakeholders were crucial in identifying the intervention factors that should be presented to decision makers in order to select which interventions should be implemented. As communities respond to legislation such as the Family First Prevention Services Act that require only evidence-based programs be eligible for reimbursement, simulation models could be helpful for making evidence-informed decisions with respect to a given community context.