Abstract for: Enhancing AI-Based Clinical Interventions through Community-Based System Dynamics
Rapid adoption of artificial intelligence (AI) in healthcare is hindered by insufficient clinician engagement, limited understanding of systemic health disparities, and unintended intervention outcomes. This project employs Community-Based System Dynamics (CBSD), a participatory method, to equip clinicians with skills to collaboratively analyze complex healthcare systems and enhance equitable, effective AI implementations addressing structural healthcare disparities. We delivered structured CBSD workshops involving 25 clinicians across healthcare disciplines within the Clinicians Leading Ingenuity in AI Quality (CLINAQ) fellowship. Fellows were trained in systems thinking, causal loop diagramming, and participatory modeling. This approach emphasizes stakeholder engagement, fairness, and equity, fostering a systemic understanding of healthcare disparities and facilitating co-design of AI interventions grounded in real-world clinical practice. Preliminary results from CBSD workshops revealed clinicians effectively constructed causal loop diagrams identifying systemic factors reinforcing healthcare disparities, such as mistrust, social determinants of health, provider bias, and clinician burnout. Stakeholders identified key intervention leverage points including transparency enhancement, bias reduction, equitable access policy development, and clinician burnout mitigation, guiding further formal system dynamics modeling. Our initial findings underscore the critical importance of clinician involvement and systems thinking in AI design. By systematically validating causal loop diagrams and prioritizing leverage points, the project advances toward robust AI interventions. This integrative approach ensures the resulting AI solutions are socially acceptable, equitable, and capable of addressing complex systemic healthcare disparities.