Simulation Consultation: San Francisco Department of Public Health

The Syndemics Lab, led by CHERISH Population Data & Modeling Core Director Benjamin Linas, uses simulation models to investigate population-level outcomes, costs, and cost-effectiveness of interventions and care delivery models to treat HIV, hepatitis C virus (HCV), and substance use disorder. In addition to their core research, the Boston-based team collaborates with real-world decision-makers to apply simulation models to support policy making. Such partnerships generate bidirectional insights: the modeling work has the potential to provide policy makers data they need for both advocacy and planning, while policy makers provide critical perspective on the simulation and help researchers ensure that their work asks the best questions and generates useful outcomes.

In March 2024, the team was excited to consult with the San Francisco Department of Public Health (SF DPH) to demonstrate how one of their models, Researching Effective Strategies to Prevent Opioid Death (RESPOND), can inform opioid use disorder (OUD) treatment capacity needs and project outcomes of expanding treatment capacity in San Francisco. SF DPH provided local data, including information on overdoses and engagement with medications for OUD (MOUD), that the Syndemics team used to reflect San Francisco’s OUD and overdose epidemiology in RESPOND.

Since the RESPOND model was originally designed for a population within Massachusetts, it was also calibrated to reflect the population in San Francisco that uses opioids. This highlighted a challenge common to the modeling community: balancing real-world situations with available data. “The model did not take into account some important factors influencing our local opioid overdose crisis,” a representative from SF DPH said. Specifically, “secular and policy-related differences that are difficult to incorporate such as huge changes in the average potency and variability in potency of street products, policing, etc.”

In response, analyses evolved throughout the life cycle of the collaboration to include the impacts of the local factors as best as possible. This meant that the process was extremely dynamic, allowing both teams to adapt to emerging results, refine future analyses, and deepen their understanding of the data and the model. To account for the gaps between real-world situations and available data, the team ran various sensitivity analyses to consider possible scenarios and finalized results in December of 2024.

Reflecting on the collaboration and analysis, a team member from SF DPH said, “the process has been helpful to guide our decisions and determine where to focus [MOUD and financial] resources.” As for the Syndemics Lab, identifying gaps between data and realistic events remains a critical component in modeling. However, through collaborative data input processes and feedback, models such as RESPOND can be adapted and utilized for scenarios outside their original designs.

The CHERISH Consultation Service offers consultation to researchers on the design and implementation of observational and interventional studies related to treatment interventions for substance use disorder, HCV, and HIV. Learn about how to receive consultation support in simulation modeling by visiting https://cherishresearch.org/what-we-do/consultation-service/.