When Evidence Becomes Selective

Simon Chapman has once again reached for a familiar line: emotional testimonials from grateful ex-smokers are the weakest possible form of evidence for the value of vaping.

“Any radio discussion of ecigarettes inevitably attracts evangelical callers wanting to tell their story of the miracle they have experienced”

On the surface that sounds impeccably scientific. Anecdotes are not data. Personal stories are not randomised controlled trials. Fair enough.

But the selectivity is difficult to ignore.

Public health has never hesitated to amplify personal testimony when it serves the anti-smoking cause. For decades, campaigns have featured powerful individual stories of lung cancer, regret, suffering and loss. These stories are not randomised controlled trials either. They are persuasive narratives. They are designed to move people. And rightly so, because lived experience matters.

Yet when smokers describe quitting after decades of failed attempts with patches, gum or going cold turkey, testimony suddenly becomes epistemologically suspect. If anecdote is weak evidence, it is weak across the board. It cannot be compelling when it warns against smoking but meaningless when it describes escaping it.

Chapman then leans heavily on early PATH cohort data from 2018 to argue that vaping produces sobering or even negative net effects. But those waves captured a particular moment in time: early generation devices, weaker nicotine delivery, high levels of experimentation, widespread dual use and a market still finding its shape. That is not today’s product landscape. Presenting those transitions as definitive proof of long-term population harm assumes a static market in what has clearly been a rapidly evolving space.

More importantly, the way outcomes are classified does much of the argumentative work. A smoker who reduces from twenty cigarettes a day to two while vaping is counted as a failure. A dual user who moves back and forth within twelve months is counted as relapse. Someone abstinent for eleven months but smoking on the survey date is a negative outcome. Public health counts only perfect, persistent abstinence as success. Harm reduction is not binary, yet the scoring system is.

The frequently cited figure that nearly a quarter of adult never-smokers who vaped later reported smoking sounds alarming in isolation. But the absolute numbers are small, misclassification in longitudinal surveys is well documented, and overall adult smoking prevalence in the United States continued to fall over the same period. If vaping were recruiting large numbers of new adult smokers at scale, we would expect smoking rates to rise. They did not.

Then comes the celebration of unaided cessation. The PATH analysis showed higher persistent abstinence rates among those who used no aid at all compared with those who used e-cigarettes. But most smokers attempt to quit without assistance. When a method is used by the largest share of quitters, it will inevitably produce the largest absolute number of successes. That does not necessarily make it the most effective per attempt, particularly among highly dependent smokers who turn to aids precisely because quitting unaided has repeatedly failed them.

Randomised trials have limitations. Observational studies have limitations. Cross-sectional surveys are imperfect. Wastewater analysis is imperfect. But dismissing trials as artificial, dismissing testimonials as weak and leaning exclusively on selectively interpreted cohort findings creates the impression not of balance, but of evidential filtering.

Relapse is real and common. It dominates cessation across all methods. That is precisely why lower-risk substitutes are relevant. The central question is not whether every person who vapes achieves permanent abstinence within twelve months. The question is whether access to non-combustible nicotine shifts behaviour away from smoking at scale. In several countries, smoking prevalence declined during periods of rapid growth in vaping. That does not prove causation. But neither can it simply be brushed aside.

Ultimately, this is less about relapse curves and more about philosophy. An abstinence-only lens will regard anything short of complete nicotine cessation as inadequate. A harm-reduction lens asks whether people are moving down the risk continuum. Those are fundamentally different frameworks.

Testimonials are not proof. But nor are they meaningless noise. They are signals. When large numbers of long-term smokers report a similar experience that vaping succeeded where other methods failed, the appropriate response is not reflex dismissal. It is curiosity, scrutiny and proportionality.

The real debate is not whether anecdotes outrank cohort data. It is whether vaping is being evaluated with the same consistency, humility and openness that public health expects everywhere else.