Summary
A practice-based research study to investigate integration of mental health in primary care was conducted generating a cohort of 2495 cases obtained from an FQHC in Montana during primary care visits. The results from electronic self-report administration of the PHQ-9 with a QPD in-depth psychological assessment administered randomly before or after the PHQ-9, are analyzed from multiple perspectives and used for training. First, using CJ Peek’s three world view, the effectiveness and cost-effectiveness of the clinical depth, breadth, accuracy and automation used in the study are analyzed demonstrating the need for dynamic administration of latent trait measures. Second, the study dataset is analyzed to produce the evidence-based AI network weights used for probabilistic, higher accuracy, item-by-item scoring, enabling more effective dynamic administration. The evidence-based AI network weights are also generated between latent traits making the comorbid relationships in the dataset explicit and available to drive the AI interview process. Finally we will use the study dataset to analyze effective use of a data warehouse integrated with an EMR and how to automatically support practice-based research. Based on the information learned, attendees will then design a screening solution for their practice that covers detection, in-depth assessment, actionable results and potential research, outcomes and QI initiatives.
Objectives
- Able to design a cost-effective in-depth detection and assessment solution integrated with their EMR
- Extend the design to include practice-based research, outcomes and QI initiatives
- Explain the need for evidence-based AI interviewing using standard measures