An Innovative Time Series Approach to Dynamic Population Modeling for Tobacco Harm Reduction Assessment

Smoking is a leading cause of serious disease among smokers, with the greatest risks arising from the combustion of tobacco and inhalation of smoke. Tobacco Harm Reduction (THR) encompasses strategies aimed at lowering the health risks associated with smoking, while potentially involving continued use of nicotine or tobacco products. Next Generation Products (NGPs), such as Electronic Vapor Products (EVPs), deliver nicotine without combustion and may play a role in THR. To predict the long-term population impact of NGPs, mathematical approaches such as Dynamic Population Models (DPM) have been widely used. These models typically rely on longitudinal data to simulate individual transitions between smoking and NGP use over time. However, such longitudinal data are often difficult to obtain. To address this limitation, we propose a novel modeling approach based on prevalence data. Historical prevalence trends were modeled using time-series forecasting methods and combined with relative risk estimates to project mortality under two scenarios: a cigarette-only market and a market with both cigarettes and EVPs. Validation against US data and comparison with a reference transition-based model demonstrated both accuracy and computational efficiency. When applied to the United Kingdom, the model predicts that EVP availability could prevent between 360,000 and 850,000 premature deaths by 2040, with up to 1.53 million deaths averted under higher switching assumptions. The largest reductions are observed in lung cancer, cardiovascular disease, and stroke mortality. These findings suggest that EVPs could substantially reduce smoking-related mortality. Furthermore, the proposed time-series DPM provides a practical and scalable framework for assessing harm reduction, particularly in settings where longitudinal data are limited.