attention. Project SAFESPACE has recently initiated screening and assessment for all children entering out of home care in Kentucky with the goal of providing assessment-driven, evidence-based treatment. The Child and Adolescent Needs and Strengths (CANS) assessment-the instrument adopted in this service system-is designed to link the assessment process with the selection of appropriate interventions. Informed by naturalistic decision making, this dissertation sought to investigate the relationship between responses on the CANS assessment and the empirically supported interventions subsequently prescribed by the assessing clinician. Using secondary data maintained in the statewide automated child welfare information system, a number of descriptive analyses were conducted in order to better understand this redesigned service system including the frequency of specific treatment prescriptions at the case, clinician, and agency levels as well as population-level CANS findings. Building upon these initial inquiries, four multilevel logistic models were developed to examine the relationship between assessment-derived predictors and the v prescribed treatment modality along four dimensions: trauma exposure, emotional/behavioral problems, family functioning, and substance use. Descriptive findings reveal an out of home care population with prevalent traumarelated symptoms and extensive emotional/behavioral needs. Considerable variation was observed in the treatment prescription patterns of assessing clinicians. Likewise, there is variation in the degree to which assessment findings align with prescribed treatment modalities. Findings suggest that assessment-derived information may be an important consideration in the prescription of a trauma-focused treatment and a minor, but still possibly salient, factor in treatment decision making related to emotional/behavioral needs and substance use. There appears to be no meaningful relationship between assessment responses and recommendations for family-focused treatment. As interpreted through the lens of naturalistic decision making, the results suggest that while assessment responses may one source of information utilized by clinicians in the field when making treatment decisions, there is likely a differential degree of influence depending on the specific dimension of treatment. Findings provide a foundation for future research into the situation-action matching decision rules employed

No abstract