📄 medicine

Population health impact of non-tobacco flavour restrictions in electronic vapour products: a dynamic modelling study in the United Kingdom

This dynamic modelling study in the United Kingdom predicts that the population-level health impact of banning non-tobacco flavours in electronic vapour products depends critically on user behavioural responses, particularly the extent to which current users relapse to smoking rather than switching to tobacco-flavoured products or quitting entirely.

Original authors: Thomas Verron, Guo Mengran, libby Clarke, Thomas Nahde, Xavier Cahours

Published 2026-08-06

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Thomas Verron, Guo Mengran, libby Clarke, Thomas Nahde, Xavier Cahours

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the human body as a bustling city, and think of smoking cigarettes as a massive, dirty factory that chokes the air with smoke and toxic fumes. This factory is notorious for causing fires, illness, and premature deaths. For years, the only way to leave this city was to quit the factory entirely, but many people found that incredibly hard to do. Enter a new kind of vehicle: the Electronic Vapour Product (EVP), or e-cigarette. Think of these as a cleaner, less smoky bus that still takes you to the same destination (getting nicotine) but without the combustion and the heavy smoke. Public health experts have been watching this "bus" closely, hoping it could help people escape the dirty factory.

However, there's a catch. These buses come in all sorts of fun flavors—like strawberry, bubblegum, or mint—which makes them very popular with young people who have never been to the factory at all. Regulators are worried that these tasty flavors might be an "on-ramp," luring new people into the city of nicotine use. To stop this, some governments are considering a rule: ban all the fun flavors and only allow the bus to run with a plain, tobacco-flavored engine. The big question is: if you take away the fun flavors, what happens to the grown-ups who are already riding the bus to escape the factory? Will they happily switch to the plain bus, or will they get so frustrated that they jump back into the dirty factory?

This paper, written by a team of researchers, uses a sophisticated computer simulation to answer that exact question. They built a "time machine" model of the United Kingdom's population, projecting how people's habits might change from now until the year 2040. They didn't just guess; they ran thousands of different scenarios to see what happens if the government bans non-tobacco flavors. They tested three main possibilities: the "Static" scenario (everyone just switches to the plain tobacco bus), the "Optimistic" scenario (everyone quits the bus and stops using nicotine), and the "Pessimistic" scenario (everyone gets angry, quits the bus, and jumps back into the smoking factory).

The results of their simulation suggest that the outcome depends entirely on how people react, but there is a very clear warning. The study found that the biggest danger isn't people quitting vaping; it's people relapsing back to smoking. The researchers discovered that even a small number of people switching from vaping back to smoking could cause a massive spike in premature deaths, far outweighing any health benefits gained from people quitting vaping entirely. In their simulations, if a ban causes even a modest number of current vapers to return to cigarettes, the result is a net loss of health for the population.

Specifically, the model showed that under a "pessimistic" scenario where 100% of non-tobacco flavor users switch back to smoking, the UK could see an increase of over 600,000 premature deaths by 2040 (assuming vaping is 95% less harmful than smoking). Even in more mixed scenarios, where some people quit and some switch back, the risk of increased smoking remains the dominant factor driving harm. The authors suggest that banning flavors might accidentally create a "gateway" back to smoking for adults who were successfully using vaping to avoid cigarettes. While the study acknowledges that banning flavors might stop some young people from starting, their simulation indicates that for the current adult population, the risk of driving people back to the "dirty factory" is a serious, potentially deadly consequence that policymakers need to weigh carefully. The paper concludes that the health impact of such a ban is not a simple math problem but a complex puzzle where the behavior of current users holds the key.

Technical Summary: Population Health Impact of Non-Tobacco Flavour Restrictions in Electronic Vapour Products

Problem Statement
Cigarette smoking remains a leading cause of preventable mortality globally, primarily due to the combustion of tobacco. While Tobacco Harm Reduction (THR) strategies advocate for the substitution of combustible cigarettes with less harmful Electronic Vapour Products (EVP), regulatory bodies have expressed concern that non-tobacco flavours (e.g., fruit, candy) may increase the appeal of EVPs to youth and non-smokers. Consequently, jurisdictions are proposing bans on non-tobacco flavours. However, such restrictions may inadvertently affect current adult EVP users, potentially altering their behaviour by causing them to quit vaping entirely, switch to tobacco-flavoured products, or relapse to smoking. The core problem addressed is the uncertainty regarding the net population-level health impact of these bans, specifically how behavioural responses among existing adult users might drive changes in smoking prevalence and mortality.

Methodology
The study employs a Dynamic Population Model (DPM) previously developed by the authors, applied to the United Kingdom context with projections extending to 2040. The methodology consists of three primary stages:

Prevalence Forecasting: Historical smoking and vaping prevalence data from the UK (2014–2022) were sourced from the Office for National Statistics. A Prophet time-series model was trained on this data to forecast baseline trends for nine combined smoking/vaping statuses (never, current, former) up to 2040.
Behavioural Scenario Simulation: The model simulates the impact of a non-tobacco flavour ban starting in 2023. The analysis assumes that initiation rates into EVP use remain constant (a conservative assumption), focusing exclusively on the behavioural transitions of current users of non-tobacco-flavoured EVP (approximately 70% of vapers). Three parameters define user transitions:
- $\alpha$ : Proportion switching to cigarette smoking.
- $\beta$ : Proportion switching to tobacco-flavoured EVP.
- $\gamma$ : Proportion quitting all nicotine use.
- Constraints: $\alpha + \beta + \gamma = 100\%$ .
The study evaluates three extreme scenarios and a continuum of intermediate combinations:
- Static ( $\beta=100\%$ ): All users switch to tobacco flavours; vaping prevalence remains constant.
- Optimistic ( $\gamma=100\%$ ): All users quit vaping; no increase in smoking.
- Pessimistic ( $\alpha=100\%$ ): All users relapse to smoking.
Mortality Estimation: Projected prevalence trajectories are integrated with a mortality model (based on Rogers et al.) and Representative Sample Trajectories (RSTs). Excess Relative Risk (ERR) assumptions for EVP use relative to smoking are varied from 5% to 50% to test the robustness of outcomes under different risk profiles. Cumulative premature deaths are calculated relative to a "no ban" baseline.

Key Results
The study demonstrates that population-level health outcomes are highly sensitive to behavioural responses, particularly the rate of relapse to smoking ( $\alpha$ ).

Dominance of Relapse: The transition from vaping to smoking is the dominant driver of harm. Even small increases in the proportion of users relapsing to smoking result in substantial increases in cumulative premature deaths, far outweighing the health gains from users who quit vaping entirely.
Scenario Outcomes (ERR = 5%):
- In the Optimistic scenario (complete cessation), the model projects a reduction of approximately 26,900 premature deaths by 2040.
- In the Pessimistic scenario (complete relapse), the model projects an increase of approximately 605,500 premature deaths.
- Intermediate scenarios show a linear relationship between the proportion of relapse and mortality outcomes. For instance, a scenario where 40% of displaced users relapse and 60% quit results in an estimated 226,100 additional premature deaths.
Risk Sensitivity: While increasing the assumed ERR of vaping (e.g., to 50%) narrows the gradient between scenarios (reducing the relative benefit of continued vaping), the dependence on behavioural patterns remains strong. Even at an ERR of 50%, the pessimistic scenario yields 318,700 additional deaths, while the optimistic scenario yields 269,100 fewer deaths.
Probability of Negative Impact: Under an ERR of 5% (consistent with literature), the model indicates that a flavour ban would result in a net negative impact on population mortality in approximately 76% of the 21 tested behavioural scenarios. Across all simulations (varying ERR and behavioural mixes), 71% of cases showed detrimental health effects for the cohort of existing EVP users.

Significance and Claims
The paper claims that the public health impact of banning non-tobacco flavours is not intrinsic to the policy itself but is critically dependent on how current adult users adjust their behaviour. The authors argue that:

Unintended Consequences: Restrictive flavour policies carry a substantial risk of unintended negative consequences, primarily driven by increased smoking relapse among adults who currently rely on flavoured EVPs as a harm reduction alternative.
Gateway Reversal: The study suggests that such bans could effectively "design a gateway into smoking" for EVP users by removing a product feature that supports sustained abstinence from combustible tobacco.
Methodological Contribution: The study underscores the value of synthetic population approaches in dynamic modelling. This method allows for the simulation of long-term outcomes across a broad range of behaviourally plausible scenarios while preserving real-world demographic structures and privacy, addressing limitations in longitudinal evidence for tobacco harm reduction.
Policy Implication: Policymakers are urged to balance potential youth protection benefits against the risk of increased tobacco-related mortality among adults. The authors emphasize that policy assessments must account for the full spectrum of behavioural responses (relapse, switching, cessation) rather than focusing solely on vaping prevalence metrics.

The authors conclude that while the scenarios serve as structured approximations rather than precise predictions, the strong imbalance between the harms of smoking relapse and the benefits of cessation suggests that flavour bans could significantly undermine tobacco harm reduction goals if they trigger even modest increases in smoking relapse.

Technical Summary: Population Health Impact of Non-Tobacco Flavour Restrictions in Electronic Vapour Products

More like this