The Spreadsheet at a Knife Fight

There is an old saying that you should never bring a knife to a gunfight, but in tobacco harm reduction some of the world’s most respected scientists made a more subtle and more consequential mistake: they brought spreadsheets to a knife fight they did not realise was happening.

For more than two decades, researchers studying reduced-risk nicotine products did exactly what modern academia trained them to do, designing randomised controlled trials, running toxicology assays, measuring biomarkers, mapping exposure pathways, reviewing epidemiological evidence, and publishing meticulous papers in peer-reviewed journals, refining uncertainty instead of slogans and assuming that if the evidence became strong enough and the methodology careful enough then public health institutions would eventually adjust course.

That assumption has quietly collapsed under its own weight.

It rests on a belief that evidence is not only necessary but sufficient, that truth once assembled carefully enough naturally rises to the surface of policy, and that is a comforting belief but also structurally false.

Most scientists are still socialised into a world where evidence is the currency and disagreement is resolved by accumulation of data, where if switching from combustible cigarettes to smoke-free nicotine products reduces toxicant exposure, improves biomarker profiles, and lowers risk indicators across multiple physiological systems, then policy eventually reflects that gradient of harm, which is the implicit promise of evidence-based medicine: better data leads to better decisions.

Tobacco harm reduction research operated inside that promise for years as researchers documented dramatic differences between combustion and non-combustion nicotine use, showed repeatedly that nicotine, while addictive and pharmacologically active, is not the primary driver of smoking-related disease, and isolated combustion as the central mechanism of harm, the delivery system that turns a psychoactive substance into a multi-system carcinogenic exposure event.

The evidence did not arrive as a single revelation but accumulated slowly across jurisdictions, methodologies, and disciplines as reviews from the United Kingdom, New Zealand, Sweden, and independent research groups converged on the same basic architecture of risk: smoke-free nicotine products sit dramatically lower on the harm spectrum than cigarettes, even if they are not harmless, and over time the direction of travel became increasingly difficult to dispute.

So the researchers did what scientists are supposed to do: they published again and again and again, each paper another increment of refinement, another reduction in uncertainty, another tightening of confidence intervals around a conclusion that was already becoming clear, and then they waited for the world to respond as though it was a seminar room.

But it was not.

It was a stage.

What they had entered was not a neutral information environment where competing hypotheses are weighed on evidentiary merit but a political theatre shaped by institutional incentives, media dynamics, moral framing, and entrenched narratives about nicotine, addiction, and corporate power, where evidence is not absent but it is not sovereign and instead is one instrument among many and often not the loudest.

On one side of the stage stood researchers carrying spreadsheets, forest plots, systematic reviews, and carefully caveated language, while on the other stood coalitions carrying slogans that required no statistical training to understand and no patience to remember: “Nicotine is highly addictive,” “We don’t know the long-term effects,” “Think of the children,” and “This is Big Tobacco all over again.”

These are not arguments in the scientific sense but narrative compression devices that reduce complexity into emotionally transmissible units designed for repetition rather than precision, and repetition in public discourse behaves like evidence even when it is not.

A researcher might spend years conducting a controlled study demonstrating reduced exposure to carcinogens in users who switch away from combustion, but that paper can be neutralised in public perception by a single headline, “Vaping could be worse than smoking,” which travels faster than the research ever could, requires no context to be believed, and requires no nuance to be shared.

By the time correction arrives, if it arrives at all, the narrative has already hardened because science is cumulative and cautious while public discourse is episodic and reactive, science rewards precision while media ecosystems reward salience, science is slow while narrative is instantaneous, and in that gap, policy is often made.

Scientists are trained to speak in conditional statements, to acknowledge uncertainty and limitations and distinguish what is known from what is not, producing formulations like “the current evidence suggests” or “confidence intervals indicate,” which are epistemically honest but politically inert because no campaign slogan begins with a confidence interval and no headline is built on a systematic review.

This mismatch becomes most visible when preliminary or low-quality evidence enters the public domain and is amplified far beyond its methodological weight, circulating globally within hours, embedded in policy discussions, cited in parliamentary debate, and echoed through advocacy networks, only to be later corrected or retracted in a slower and quieter process that rarely dislodges the original claim.

The claim becomes memory while the correction becomes archive, and this pattern is not rare but structural, particularly in tobacco harm reduction where alarming studies are amplified rapidly, reassuring studies struggle for visibility, dramatic claims are institutionalised, and corrections are treated as administrative rather than epistemic events, producing a slow sedimentation of outdated or incorrect claims into something resembling consensus.

Australia illustrates this dynamic clearly, with independent researchers repeatedly presenting evidence on relative risk, smoking cessation trends, international comparisons, and unintended consequences of restrictive nicotine policies, pointing to countries where smoking has declined alongside regulated access to lower-risk alternatives, while political narratives have instead centred on youth vaping panic, prohibition framing, and moral urgency that flattens complexity into binary harm.

When signals emerged suggesting that smoking rates among younger adults may have stabilised or increased following intensified enforcement against vaping products, there was little appetite for reflection, not necessarily through explicit rejection but through institutional inertia that slows, dilutes, or fails to amplify destabilising interpretations.

Over time, this creates an imbalance where some forms of evidence travel easily through media and policy systems while others struggle regardless of methodological strength, because the determinant is not only quality but compatibility with existing narrative structures.

The mistake many reduced-risk researchers made was not epistemic but architectural, assuming they were operating inside a system where evidence accumulates toward consensus in a linear way and that institutions claiming to be evidence-based would structurally prioritise evidence over narrative stability, while in reality institutions are actors with incentives, reputations, political constraints, and risk thresholds that respond not only to what is true but to what is communicable and defensible.

Some scientists adapted to this reality by engaging more directly in communication and policy discourse, while others continued to behave as though publishing another high-quality paper would be sufficient to correct a public narrative that was never built from papers in the first place.

Because the conflict was never purely scientific but moral, institutional, and political, centred on addiction, responsibility, corporate history, youth protection, and the symbolic meaning of nicotine itself, where evidence competes not only with other evidence but with identity narratives that are far more resistant to revision.

You cannot dismantle a moral frame with a meta-analysis, you cannot override a fear-based narrative with a confidence interval, and you cannot out-publish a slogan.

The tragedy is not that the science was weak but that it was strong in a system that does not privilege strength alone.

None of this invalidates the scientific work because the evidence on relative risk has mattered enormously in real-world behaviour, contributing to millions of individuals shifting away from combustible cigarettes, and the science has real consequences even when it is politically underpowered.

But the gap between evidence generation and evidence impact remains the central unresolved problem because evidence does not enforce itself; it must be carried, translated, defended, and made to survive contact with institutions that are not neutral conduits but selective filters.

The tobacco harm reduction field has experienced this tension more visibly than most areas of public health, but it is not unique, it is simply an early example of a broader reality in which contested policy domains do not reward the best evidence automatically but instead reward the evidence that best survives institutional, media, and narrative selection pressures.

The scientists were not wrong to bring spreadsheets, they were wrong to assume that everyone else was playing a game in which spreadsheets mattered most, they were not, they never were, and by the time that became clear, the knife fight was already well underway.