One "Smoking Rate", Eight Comparability Breaks: What WHO's Own Metadata Says
In one line: Line up national "smoking rates" in a table and they look like the same metric. But in its own indicator metadata, WHO lists exactly what breaks that comparability: product scope (smoked only, or all tobacco — and whether e-cigarettes are counted), the age floor (15+ or 18+), self-report under-reporting, the reference window for "current use", and whether figures are age-standardised. Most fundamentally: the number WHO publishes is a statistical model estimate, not an official statistic reported by countries to WHO. Eight breaks to clear before any cross-border comparison.
1. What the indicator actually measures
WHO's indicator is: "Age-standardized prevalence of tobacco use among persons 15 years and older (%)".
| Item | Content |
|---|---|
| Numerator | People who currently use any tobacco product — both daily and non-daily/occasional users count |
| Denominator | Population aged 15 and over |
| Products included | Cigarettes, cigars, pipes, waterpipe (hookah/shisha), bidis, kretek, heated tobacco products, and all forms of smokeless tobacco (oral and nasal) |
| Explicitly excluded | E-cigarettes (they contain no tobacco) |
| Data source | Nationally representative household surveys |
| Presentation | Age-standardised to the WHO Standard Population |
"Tobacco use" and "smoking" are not the same thing in this indicator — that distinction drives every break below.
2. The eight breaks, each from WHO's official metadata
| # | Break | What WHO states |
|---|---|---|
| 1 | Product scope | Some countries monitor only cigarettes, others all tobacco. "In many countries where other forms of tobacco are also being consumed, smoking rates will be lower than tobacco use rates" |
| 2 | E-cigarettes | Excluded from this indicator, but "countries have begun to monitor use of e-cigarettes and other emerging products, which may confound countries' definitions of tobacco use" |
| 3 | Age floor | This indicator uses 15+; the WHO Health 2020 definition for the same topic uses 18+ — and WHO annotates this directly: "the definition applied in the source … is not completely identical" |
| 4 | Self-report bias | "Surveys ask people to self-report their tobacco use, which can lead to under-reporting" |
| 5 | Reference window | The definition of "current use" is not always restricted to the 30 days prior to the survey |
| 6 | Age standardisation | Standardised rates exist only to compare across countries or periods; WHO states plainly that they should not be used to estimate the number of smokers |
| 7 | Estimate vs official statistic | "The data presented here are WHO estimates, and not official statistics reported by countries to WHO" |
| 8 | Borrowed data | Where a country has fewer than two nationally representative surveys in different years, no attempt is made to fill gaps and no estimates are calculated; where gaps are filled, information is "borrowed" from countries in the same UN subregion |
Break 8 is the one most often missed: a figure in the table may derive partly from a neighbouring country's survey, not that country's own measurement.
3. The consequence: one country, one year, several different rates
Combining the breaks, a single country in a single year can carry at least these versions:
| Version | Break |
|---|---|
| Cigarette smoking rate vs all-tobacco use rate | 1 |
| Including vs excluding e-cigarettes | 2 |
| 15+ vs 18+ | 3 |
| Survey measurement vs model estimate | 7 |
| Crude vs age-standardised rate | 6 |
| Own survey vs borrowed regional data | 8 |
Any two combined produce a gap large enough to look like two different problems in two different places — when they may be one thing computed two ways.
4. WHO states there is no standard survey protocol
WHO's metadata says it directly:
"There is no standard protocol used across Member States to ask people about their tobacco use."
WHO's Tobacco Questions for Surveys (TQS) have been adopted in many surveys, which helps improve comparability, but adoption is not universal.
This means that even where two countries both ran a "nationally representative survey", the questions asked may differ.
5. Comparable statements from other institutions
- Nuffield Trust (UK health policy think tank), explaining international comparison of preventable risk factors: "International comparability may be affected by a lack of standardisation in the measurement of smoking habits in health interview surveys across OECD countries."
- World Bank, World Development Indicators: its adult tobacco-use indicator draws on the WHO Global Health Observatory and notes the data are harmonised to adjust for differences in age structure.
- WHO European Region health indicators: the same topic is maintained under both 15+ and 18+ definitions, with the divergence annotated.
6. How to read the indicator without error
| Practice | Why |
|---|---|
| Always cite year, age floor and product scope | Change any one and the figure changes |
| Distinguish "estimate" from "official statistic" | The WHO series is a model estimate |
| Distinguish age-standardised from crude | One is for comparison, the other approximates the actual burden |
| Never derive smoker counts from standardised rates | WHO explicitly prohibits this use |
| Check uncertainty ranges | Countries with irregular surveys have wide intervals; WHO advises caution |
Boundaries and scope
- This piece compares measurement methods and indicator definitions only. It does not compare national results, rank countries, or evaluate trends. That is a hard rule for this topic: cross-national health statistics are easily read as "grading governments", whereas the subject here is measurement, not governance.
- No country-level figures are listed. The reason is not that the data are unavailable, but that placing numbers side by side before the definitions are aligned is itself misleading.
- All definitions are taken from WHO's official indicator metadata, confidence A; national survey instruments were not read, so the actual degree of divergence between questionnaires is an unverified item.
- This is a population-level statistical tool and does not apply to any individual determination.
- No commercial data-vendor sources were used.
Sources
Sourcing floor: all data here comes from public international organisations and intergovernmental bodies. No commercial data vendors.
| Data | Source | Grade | Confidence | Record ID |
|---|---|---|---|---|
| Indicator definition (age 15+, current use including daily and non-daily, product list, exclusion of e-cigarettes); numerator and denominator | WHO Data indicator page | L1 | A | — |
| Eight comparability breaks: product scope, e-cigarette definitional confounding, 15+ vs 18+ divergence, self-report under-reporting, non-uniform reference window, restrictions on age-standardised rates, estimate versus official statistic, subregional data borrowing | WHO Data indicator page (Comparability / Denominator / Method of estimation sections) | L1 | A | — |
| "There is no standard protocol used across Member States"; TQS helps improve comparability | World Health Organization (WHO) | L1 | A | — |
| International comparability may be affected by lack of standardisation in survey measurement | Nuffield Trust (UK health policy think tank) | L2 | B | — |
| Indicator sourced from WHO Global Health Observatory; harmonised for age structure | World Bank, World Development Indicators (official blog) | L1 | B | — |
| Same topic maintained under both 15+ and 18+ definitions, divergence annotated | WHO European Region health indicators gateway | L1 | A | — |
Grade key: L1 = government / international organisation / intergovernmental body / customs; L2 = international standards bodies / regulated exchanges / national think tanks / national academic societies; L3 = industry associations / company statements / official-media relays (grey, original source must be named); L4 = commercial data vendors / self-media (not used by this project).
Rows marked "—" are published here for the first time and are not yet in the site's record index.
Data as of 2026-10-11. Individual records can be retrieved by Record ID via the site's data index or the MCP endpoint.