Session: Developing Empirically-Based Risk Screening Tools for Use in Pediatric Settings: Implications for Preventing Externalizing Psychopathology (Society for Prevention Research 25th Annual Meeting)

The session will contain a series of studies outlining how secondary analysis of existing longitudinal data can be used to develop and refine screening strategies to identify youth in the community who should be referred to specialized interventions designed to prevent the development of severe externalizing psychopathology (e.g., violence, heavy substance use). The presenters will address a complementary set of challenges associated with developing precise screening tools using data collected on a sample of boys who have been repeatedly assessed from childhood into adulthood (i.e., Pittsburgh Youth Study). The symposium supports the conference theme of developing strategies for identifying youth in need of prevention services in primary care settings.

The first paper, “Accurate Identification of Boys At-Risk for Serious Aggression and Violence – Challenges in Classification Methods,” addresses the importance of evaluating violence risk screening strategies for children in terms of classification accuracy, including examining whether screening precision varies based on the informant (i.e., parent vs. teacher), the child’s grade, and the child’s race.

The second paper, “Accurately Identifying Adolescents Who Will Exhibit Persistent Frequent Substance Use in Adulthood: The Importance of Replicating Findings Across Studies” examines whether a previously validated screener for substance use problems exhibits equivalent levels of classification accuracy when implemented in an independent longitudinal sample.

The third paper, “Potential Contributions of Machine Learning Methods to Screening Efforts in Pediatric Settings” describes how complex data-driven algorithms can be used to develop pediatric violence risk screening tools with increased predictive accuracy. The absolute and relative performance of several algorithms are compared to holdout (i.e., new) data from the same sample.

Following the presentations, a discussant will summarize the findings within the context of his extensive program of research developing and evaluating interventions targeting youth at risk for developing externalizing psychopathology. The discussant will then moderate discussions with audience members focused on implementing empirically-supported screening tools in pediatric care settings.