Do we really know how many people smoke in a general population? The case of Italy (mini-review)

Smoking prevalence is usually assessed by mailed questionnaire or personal interview on general population samples. However, participation in epidemiological surveys is decreasing throughout the world. This declining trend not only reduces precision in prevalence estimates, but can also bias these estimates, as current smokers tend to be late responders to epidemiological surveys. In Italy estimates of smoking prevalence, yielded by different surveys, are rather consistent, irrespective of the methodological approach used: samples withdrawn from the general Italian population or within participating centers; mailed questionnaires or personal interviews; replacing or re-contacting non-available subjects. However, inconsistency can occur among surveys performed in the same period, or even within the same type of survey repeated one-year apart. Hence combining the results of different surveys rather than relying on the results of a single survey is advisable.