A Simulation-Based Diagnostic Stewardship Framework for Imaging Utilization in Primary Care: A Model Using 100 Common Clinical Conditions

Background: The increasing utilization of diagnostic imaging has raised concerns regarding imaging overuse, unnecessary radiation exposure, and downstream diagnostic cascades. Because primary care physicians serve as the first point of contact for most patients, diagnostic decisions made in primary care may substantially influence healthcare resource utilization at the system level. This study aimed to develop and evaluate a conceptual diagnostic stewardship framework for primary care using a simulation-based modeling approach. Methods: A synthetic dataset consisting of 100 common primary care conditions was developed across ten clinical domains. Model parameters, imaging utilization probabilities, and diagnostic pathway assumptions were derived from literature-informed estimates and multidisciplinary expert judgment. Each condition was assigned diagnostic attributes including World Health Organization age group classification, commonly requested laboratory tests, preferred imaging modality, and imaging necessity classification (Class A: imaging usually unnecessary; Class B: conditional imaging; Class C: imaging usually required). Using this dataset, a simulation model representing one million hypothetical primary care visits was constructed. Imaging utilization, modality distribution, radiation burden index, incidental diagnostic cascades, and a relative diagnostic resource utilization index were estimated under a baseline diagnostic scenario and a framework-guided diagnostic stewardship scenario. Results: In the baseline scenario, the model generated 412,000 imaging examinations across one million simulated visits (41.2% imaging rate). Within the simulation model, application of the framework was associated with an estimated reduction in imaging examinations to 258,000, corresponding to a 37% reduction in imaging utilization. The estimated population-level radiation burden index decreased from 285,000 to 179,000 units, representing a 37% reduction in radiation exposure. The number of incidental diagnostic cascades decreased from 48,200 to 29,700 events, while the relative diagnostic resource utilization index decreased from 2,480,000 to 1,690,000 units. Sensitivity analyses confirmed the robustness of these findings across alternative model assumptions. Conclusions: Within the assumptions of this simulation model, the proposed diagnostic stewardship framework generated modeled reductions in imaging utilization, radiation burden, and downstream diagnostic consequences. These findings illustrate the potential impact of structured diagnostic stewardship strategies and provide a hypothesis-generating basis for future validation using real-world clinical data.