[The truthfulness of the answers of children, adolescents and young people to surveys on tobacco consumption conducted in schools].

BACKGROUND

The prevalence of tobacco consumption in scholars has been determined by self-governed inquiries, methodology that presents limitations of reliability and validity. This obstacle has been endeavoured to overcome through methodological inspections such as the test-retest valuation. On the contrary, the available analytical methods (i.e., determination of cotinine in urine or determination of carbon monoxide [CO] in exhaled air) have not been used up to now in this type of studies. Their use together with the inquiries could complement the results of the questionnaires and verify their reliability.

POPULATION AND METHODS

814 scholars took part in the survey, being valid the results of 809, 385 male students and 424 female students, 13 to 24 years old (mean 15.90, SD 1.73) from an urban high school and another in the rural area. The survey included a survey and the measurement of CO levels in exhaled air in the same scholar centre.

RESULTS

257 scholars stated that they had never tried tobacco and in 15 of them (5.8%) levels of CO were equal or higher than 10 ppm (14.26, SD 4.35 ppm). 589 students stated that were not regular smokers; in 46 of them (7.8%) levels of CO were superior to 10 ppm (13.87, SD 4.68). The morning when the survey was held 629 students assured that they had not smoked: in 52 of them (8.3%) levels of CO higher than 10 ppm were observed. The mean level of CO observed in these students was 13.5 (SD 4.4) ppm. In our survey, the measurement of CO in exhaled air allowed us to show percentage between 5.8 and 8.3% of students whose answers to the different questions were not reliable.

CONCLUSIONS

The reliability of the results, observed through the carrying out of surveys, can be determined with the use of analytical methods that, owe to their simplicity, can be used massively. We have not found any relation between the shortage of reliability and the variants studied (age, sex or habitat), which reaffirms the difficulty in identifying variants that can determine those "deceives" and demands for its identification the need of practicing analytical determinations.