Engaging Community to Co-design a Multilevel Intervention to Reduce Lung Cancer Disparities in Persistent Poverty Tracts in California through Group Model Building and Simulation

Introduction: Despite a 13.9% decline in lung cancer incidence and a 33.4% drop in mortality in California (1988–2021), screening uptake remains low (13.9%), with a five-year survival rate of just 14.8%. Notably, one-fourth of patients receive no treatment, and over 80% are diagnosed at advanced stages. These poor outcomes are exacerbated in persistently poor communities. This study examined social determinants and structural barriers to lung cancer outcomes in Kern and Fresno Counties of California and co-developed a multi-level intervention strategy informed by community perspectives. Methods: This multi-approach study employed community engagement, participatory modeling, and system dynamics (SD) simulation to develop a model to test scenarios and identify a multi-level intervention strategy. Results: We identified 181 variables suggested by the group model building (GMB) participants. These variables were categorized into five domains and 12 thematic areas, which shaped the system boundary and model structure. Of the 16 single and three bundled scenarios tested, seven showed a cumulative reduction in the at-risk population and increases in screening, diagnoses, treatment, and cancer-free survival. Following discussions, participants selected a multi-level strategy focused on expanding public health and insurance education and advocating for air pollution-related screening within existing protocols. Discussion: Rural residence and poverty interact to heighten lung cancer risk, yet intervention studies in rural settings remain scarce. Our study advanced a multi-level understanding of causes leading to lung cancer disparities and designed an intervention strategy to address obstacles along the cancer care continuum through simulation. While other scenarios showed greater impact, participants chose to intervene through public health education and screening, likely reflecting bounded rationality in decision-making. By using SD modeling as an information-processing tool, we supported stakeholders in navigating complexity and refining their decisions. Conclusion: Community engagement is essential for expanding system boundaries, understanding lung cancer disparities, and designing practical, multi-level interventions. Scenario testing enables informed planning to improve long-term population health outcomes.