In this study we will develop an agent-based simulation model (ABM) to help policy makers and health professionals in North Carolina identify the best mix of cost-effective interventions to reduce opioid overdoses (ODs) and related deaths. Interventions are identified in the NC Opioid Action Plan and cover the Three Pillars: prevention, connection to care, and harm reduction. Our ABM will represent a community (e.g., a town) of individuals (patients, physicians, dealers, etc.), and simulate how proposed interventions affect individual pathways to opioid misuse and other outcomes (i.e., OD death). The estimation of transition probabilities between the states in these pathways will be based on data from several sources: North Carolina dashboard, national studies, and published literature. The model will rely on a representative synthetic population, which allows multiple data types (e.g. prevention, treatment) to be probabilistically connected in one model.
Aim 1. To develop a North Carolina-specific ABM that describes multiple pathways of opioid use in the context of prescription practices, treatment modality and availability, the illegal drug market, prevention policies, and other factors affecting the parameters of the various pathways that lead to OD fatalities. Besides OD deaths, we will investigate multiple other sources of morbidity. We will leverage existing national models and a representative synthetic population to examine spatial (community-level) and temporal (short- and long-term) effects of prevention and treatment interventions on opioid misuse and ODs. We will validate the model on North Carolina data from the past 13 years and will evaluate the sources of prediction uncertainty.
Aim 2. To predict the response to the mix of interventions specified in the North Carolina Opioid Action Plan at the local level (e.g., county). The policies include reducing the over prescription of POs, increasing naloxone availability, increasing community awareness, and expanding treatment and recovery care. We will estimate the uncertainty of the forecasts accounting for the changing policy and environmental factors and will refine the model on the basis of new data from the NC DHHS. We will discuss the results with the expert panel and will disseminate data-driven recommendations to North Carolina stakeholders to generate public health impact.
Aim 3. To estimate the cost and cost-effectiveness of the key interventions in Aim 2 and compare them with the status quo. For each intervention, we will work with the NC DHHS to estimate costs and cost variation by county characteristics (e.g., population density, poverty). The model will address a significant public health problem and will inform policy on the short- and long-term cost-effectiveness of these interventions.
This study will examine the effects and cost of opioid interventions at the state and local level to assist policy makers and health professionals in North Carolina. Using publicly available data and published literature we will build a geographically explicit agent-based model (ABM) that will simulate individual trajectories to opioid overdose and death under different intervention scenarios. These interventions will reflect North Carolina Opioid Action Plan 2 and focus on ?three pillars? prevention, connection to care, and reduction of harm. We will use representative synthetic populations to link individual trajectories with local environment and context.
