[What is the prevalence of smoking in China].

OBJECTIVE To determine the accuracy of prevalence data sets on tobacco use so as to measure the risk of tobacco use and the impact of tobacco control in China. METHODS Three published data sets on nation-wide survey were reviewed and compared. Two principles were applied to determine the accuracy of the data on prevalence: i ) The estimated consumption of cigarettes based on the current prevalence rate on smokers should have been close to the actual cigarette consumption level; ii) change on the annual prevalence of male current smokers should be around 1% in China, since the international experience on the prevalence of current smokers tended to decrease at a rate of around 1% per year in the presence of comprehensive tobacco control strategies. RESULTS The differences between the estimated cigarette consumption and the actual cigarette consumption for the three surveys were 51.07 billion through Behavior Risk Factor Surveillance System (BRFSS 2002), 62.94 billion through the Chinese Citizen Nutrition & Health Survey (CCNHS 2002), and 217.11 billion through the China Health Service Survey (CHSS 2003). In comparison with the national tobacco use survey in 1996, the prevalence of male current smokers apparently dropped by 0.9% in BRFSS 2002, 2.2% in CCNHS 2002 and 2.0% per year in CHSS 2003. Thus, the prevalence of current smokers in BRFSS (2002) was more reliable, comparing to the results from the other two surveys. CONCLUSION The prevalence of current smokers as determined by the BRFSS should be used to reflect the epidemic of tobacco use when implementing the FCTC in China. However, the reporting prevalence rates of tobacco use were different in the different surveys regarding tobacco use, suggesting that the capacity of surveillance on tobacco control should be strengthened, including the standardization of definitions on 'ever-smoker' and 'current smoker' , as well as on standardized questionnaire, sampling strategy and the process of data analysis, quality of field work etc. Precise estimation of prevalence appears to be the key point for understanding how many current smokers so as to develop control policy, including setting up 'quit' clinics and evaluating the impact of tobacco control programs. There is an urgent need to establish a national standardized surveillance system to monitor the tobacco epidemics.