‘Adults could join’: does a quit-smoking study represent all adults?

01

Follow an invented adult-only headline

This is a wholly fictional study, with no real programme, institution or effect estimate. Its criteria allow adults aged 18 or over who smoke cigarettes, can read the study's single language and can attend the city venue. Campus and workplace noticeboards recruit volunteers. The baseline table lists 240 participants aged 20–59; none are 65 or over. The records do not tell us how many older adults saw the notices, were assessed or declined.

The study randomly offers four evening group workshops plus a brief consultation, versus the same consultation alone. Its report describes four workshops being delivered, but the attendance distribution is not supplied in this exercise. The defined outcome is no cigarette smoking during the preceding seven days, assessed at six months; no outcome numbers are provided.

Now imagine a public service question about daytime neighbourhood workshops, including people aged 65 or over and additional languages. ‘Open to adults’ does not establish evidence in those older groups. Yet their absence does not demonstrate no benefit, a safety problem or a prohibition on help. Nor do we know why they were absent. The change of language and delivery needs its own assessment, not an invented biological age effect.

A faithful note is: ‘The criteria had no upper age limit, but the enrolled sample was 20–59. Evidence for the proposed older, multilingual community setting is less direct; recruitment reach, adaptation and the new comparator remain to be examined.’ Do not fabricate a new quit rate from these characteristics. The table below records questions about public study and service documents, not about the reader.

02

Three records that cannot substitute for each other

Read the eligibility criteria and their rationale, the recruitment setting and methods, then the baseline characteristics by assigned group. Where recorded, the participant-flow diagram distinguishes assessed people, exclusions, refusals, randomised participants and those with follow-up. Those stages are not interchangeable; unavailable counts should remain unavailable.

Random allocation between study arms is not random sampling from everyone who smokes. It helps the internal comparison when properly performed, but does not by itself make volunteers representative. Likewise, similarity to a target population does not repair a biased or poorly measured study result. Keep the credibility of the original result separate from its applicability.

Investigate material differences rather than counting matches. Evidence from other populations or settings may inform the question, with suitable uncertainty. Do not declare every demographic difference an effect modifier or demand an exact personal match to the sample. A missing subgroup is an evidence gap, not a subgroup result.

[1][2]

03

Carry both sides and the outcome into the new question

Compare what was actually delivered with what the proposed service would provide: content, language, location, scheduling, personnel and access conditions. Translate the comparator too. An addition to a brief consultation is not automatically the same contrast as an addition to an established intensive programme. The common intervention name does not preserve that comparison.

Keep the tobacco-product scope, abstinence definition and assessment time. A six-month seven-day measure does not answer a question about a year of continuous abstinence or service reach. Even if a relative effect were transportable, absolute differences can change with the comparator outcome rate; without justified data, do not assign a precise local effect. Seek relevant studies, implementation information and explicit discussion of unresolved differences, not an applicability score.

[1][2]

04

Study entry rules do not decide who deserves help

Cochrane cautions against rigidly turning study inclusion and exclusion criteria into clinical application rules. A research requirement can serve measurement, scheduling or design rather than establish a personal contraindication. This page needs no age, diagnosis, smoking history or other personal inputs, and does not decide anyone's service eligibility or treatment.

For personal help, NHS routes differ across the UK nations; check the appropriate local service or qualified professional. A service's actual eligibility and available alternatives must be confirmed there. Being absent from one study is not a reason for this page to deny you support.

[1][3]

05

What to keep in mind

• Read criteria, recruitment and actual participants separately.
• Explain material differences without inventing subgroup effects.
• Keep the comparator and outcome; do not turn study absence into denial of care.