Three results that must retain their labels
Real published estimates, not local service ratings or individual predictions. A confidence interval crossing 1 includes no relative difference.
| Evidence | Endpoint and comparison | Estimate and limit |
|---|---|---|
| 2024 review; search to September 2023 | Self-report assessed ≥6 months after intervention began; ≤3-minute advice vs no smoking advice/no contact | Publication-bias-adjusted RR 1.17 (1.07–1.27); moderate certainty |
| Biochemical sensitivity analysis; 6 studies | Confirmed abstinence assessed ≥6 months after intervention began; advice vs no smoking advice/no contact | RR 1.53 (0.98–2.40); no clear difference established |
| Older Cochrane comparison; search 2013 | Abstinence assessed at ≥6-month follow-up; brief physician advice vs no advice/usual care, 17 trials | RR 1.66 (1.42–1.94); different eligibility, not a trend estimate |
What counts as ‘brief’ matters more than the label
WHO describes encounters lasting 30 seconds to three minutes. The 2024 review specifically examined advice lasting no more than three minutes, mostly delivered by doctors in clinical settings. Asking about smoking, advising stopping and offering a route to support are examples of studied components—not a script that this page asks you to deliver.
Some studies also supplied a leaflet or contact card, and some used follow-up advice. A study of this package does not isolate the causal contribution of a single sentence, a professional's job title or the leaflet. It is also different from a scheduled course of counselling.
Try reading two invented summaries
Summary A says ‘a 30-second spoken reminder’. Summary B says ‘the reminder, a card with service contacts, and another call a month later’. These are fictional descriptions for a reading exercise, not care instructions or outcome evidence. Even if both headlines say ‘brief advice’, B tests a different package. Underline all three parts before reading B's result; assigning it entirely to the spoken reminder would erase the other parts.
The newer review follows people beyond the conversation
Cheng and colleagues searched through 30 September 2023 and included 13 randomized trials with 26,437 participants. Their main endpoint was self-reported tobacco abstinence assessed at least six months after the intervention began. That assessment time does not mean everyone had been continuously abstinent for six months; the underlying trials used different definitions.
After adjustment for possible publication bias, the pooled risk ratio was 1.17 (95% confidence interval 1.07–1.27), with moderate-certainty evidence. This is a relative comparison with the control condition, not a 17-percentage-point rise or a statement that 17 of every 100 recipients quit. An absolute difference needs the corresponding control proportion.
Why you may also encounter a larger, older estimate
The Cochrane physician-advice review's website carries a 2022 publication display, but its reported search is January 2013, with an additional Latin American search in February. It includes 42 trials from 1972–2012; the brief-advice versus no-advice/usual-care comparison draws on 17 of them. Its broader intervention definition is not identical to the newer three-minute limit.
Do not average the two ratios or conclude that advice has become less effective over time. Different eligibility rules, outcome handling and bias adjustments can produce different estimates. The table preserves their distinct questions rather than ranking them.
Self-report and biochemical confirmation are not interchangeable
In the newer review, the six-study analysis using biochemical confirmation had RR 1.53, but its interval was 0.98–2.40 and included no difference. It does not independently establish a clear benefit on that endpoint, nor does it overturn the self-report result by itself. It is a less precise, different analysis.
The reviewers also noted limited evidence from nurses, dentists, community settings and non-cigarette tobacco users. A headline saying ‘all professionals, all products, all situations’ goes beyond that evidence. Moderate certainty is meaningful support for a population finding, not certainty about every encounter.
A brief offer can be a doorway, not the end of available help
A short recommendation does not imply that longer support is unnecessary or that someone who continues smoking failed to listen. WHO separately recommends more intensive behavioural support for adults interested in quitting. Questions about personal care, medicines or ongoing difficulty need a qualified professional, not a pooled ratio.
In England, NHS Better Health links to local Stop Smoking Services. Elsewhere, use the service available in your own jurisdiction and check access arrangements. A useful question is ‘What further support is available here?’—not ‘Does this statistic prove that your advice will work for me?’
What to keep in mind
Common questions
Does RR 1.17 mean my chance of quitting is 17%?
No. It compares study-group proportions, rather than supplying a personal probability. Neither an individual's baseline probability nor a local service's absolute result follows from that number alone.
Sources
The central claims on this page were checked against the sources below.
- Journal of General Internal Medicine: Cheng et al. (2024): very brief advice, methods, outcomes and limitations
Sources checked: 2026-10-05
- Cochrane: Physician advice for smoking cessation — search through January/February 2013
Sources checked: 2026-10-05
- World Health Organization: WHO clinical treatment guideline for tobacco cessation in adults (2024), §3.1, print pp. 11–13
Sources checked: 2026-10-05
- NHS Better Health: Ready to quit smoking — local support in England
Sources checked: 2026-10-05
General adult evidence literacy, not a personal treatment recommendation or instructions for a clinical consultation. No personal smoking or health information is needed or collected here. Individual questions belong with qualified local care.