WPC 2B ZR ,#|`HP LaserJet 4HPLAS4.PRSx  @\ubX@ 3'3'Standard6&6&StandardHPLAS4.PRSx     USUK#x6X@8;2X@#  X` hp x (#%'0*,.8135@8:DgFdddgF9CgggEEggzzddd~rrEgdF"m^gCdddgC9CgggEEggzzdddsrrEgdC2 a #Xi\  P6ƒXP#@VI - @   6&6&StandardHPLAS4.PRSx  6&6&StandardHPLAS4.PRSx       #{\  P6MP#     #Xi\  P6ƒXP##{\  P6MP#  CHAPTER VI  X #{\  P6MP##Xi\  P6ƒXP#@VI - VI - @   #Xi\  P6ƒXP#MONITORING THE TOBACCO EPIDEMIC: PRINCIPLES AND ISSUES  X  #Xi\  P6ƒXP#WHY MONITOR THE TOBACCO EPIDEMIC? 55Since the 1950s, an overwhelming body of scientific evidence on the health consequences of tobacco use has accumulated. These laboratory, clinical, and epidemiological studies have repeatedly demonstrated that tobacco in all its forms greatly increases the risk of premature death from a number of chronic diseases, including coronary heart disease, stroke, chronic bronchitis and emphysema, and cancers of the lung, larynx, mouth, oesophagus, pharynx, pancreas and bladder. Tobacco use is also a contributing factor for cancers of the kidney and cervix. 55Numerous other adverse health conditions including other respiratory diseases, peptic ulcers and pregnancy complications are caused or exaggerated by tobacco use. The adverse effects of smoking on pregnancy outcome range from low birth weight, (one of the strongest predictors of infant mortality), to increased incidence of spontaneous abortions, prematurity, stillbirths and sudden infant death syndrome. Tobacco used in smokeless forms, such as for chewing or snuff taking, is a major cause of oral cancer. These health complications are compounded in developing countries due to limited access to health care and often scarce resources to treat diseases resulting from tobacco use. 55Exposure to environmental tobacco smoke (ETS) is a cause of disease, including lung cancer, in nonsmokers. Children of parents who smoke have an increased risk of lower respiratory tract infections, such as bronchitis and pneumonia, compared with those of nonsmoking parents. The incidence and severity of asthma in children is also greatly aggravated by passive smoking. There may be a cumulative effect of ETS in that those children exposed to maternal smoking before birth generally continue to be exposed to environmental tobacco smoke after birth. *-++Ԍ55Policies and programmes to control tobacco use can be assisted by reliable, relevant and timely information about the pattern, extent and trends of tobacco use in the population, the health and economic consequences of tobacco use, and the sociocultural factors which underlie it. Given the already prominent role of tobacco as a major health hazard, and the likelihood that the health effects will increase dramatically in the future, it is clear that the regular assessment of tobacco use and associated disease trends should become an integral part of a country's health information system and updated regularly. 55WHO and its affiliated organization, the International Agency for Research on Cancer. together with other agencies concerned with tobacco control, (particularly the International Union Against Cancer (UICC)), have previously addressed this issue by proposing questionnaires, definitions and guidelines for conducting tobacco use surveys among the general population (adults and young people), as well as among health professionals and school teachers. While these documents are intended to facilitate comprehensive data collection efforts, it has become necessary to update them to reflect advances in knowledge and survey procedures. The following chapters are intended to meet this need.  X-  EPIDEMIOLOGICAL SURVEILLANCE OF THE TOBACCO EPIDEMIC  X  55Unlike the natural history of infectious diseases, where the time between infection and disease manifestation is typically short, there is a long delay between persistent smoking and full health effects. The risk of lung cancer, for example, depends strongly on the duration of smoking. As a result, current lung cancer rates are largely determined by smoking patterns two to three decades ago. This long delay, illustrated schematically in Figure VI.1, is a chief source of misunderstanding about the health effects of tobacco. The model depicted in the Figure is an approximation based on the experience of the developed countries and shows the various phases of the cigarette epidemic with exposure first rising among males, (with a 23 decade delay in death caused by tobacco), and then among females (with a comparable delay before the health effects are felt). The implications of this model for developing countries are clearthe health hazards of smoking are not yet evident simply because men in these countries have not been smoking manufactured cigarettes for sufficiently long to experience the full health effects. There is an even greater urgency to prevent the epidemic among women in developing countries who, by and large, do not smoke.C*-++Ԍ X  ÙThe need for standardization 55Standardized approaches facilitate global, regional and national monitoring of the tobacco epidemic, and the evaluation of the effectiveness of policies and programmes to control it. In order for WHO to effectively comply with the World Health Assembly's request to assist countries to implement comprehensive tobaccocontrol policies and to closely monitor the evolution of the global epidemic of tobaccorelated diseases, it is necessary that national policies and programmes be based on internationallycomparable principles and procedures. A  X major aim of this book is to promote the use of standardized measures and approaches for the continuous assessment of tobacco use and its impact on health.   X  The need for appropriate data 55For policies and programmes to be effective in reducing overall tobacco use and subsequent health consequences, information needs to be collected on tobacco use within specific population groups. The most important categories of a population for which specific data are required include:  X  sex :55NN& &  men usually begin to smoke before women do and in much larger numbers. Thus female prevalence generally does not reach the same level as that for men, and only reaches a peak several years later (see Figure VI.1). Combined data for men and women will mask these major sex differences in smoking behaviour and should be avoided.#  X  age :NN& &  the classical pattern of agespecific prevalence, at least in developed countries, is for younger age groups to begin smoking first, with older women smoking much less. In some southern European countries, typically about half of all adolescent girls smoke but less than 5% of women above age 55 or 60 do so. As these cohorts of younger smokers age, prevalence also rises among older age groups as well, in the absence of smoking cessation. Hence it is absolutely essential to collect prevalence data by broad age groups.# C*-++Ԍ X ԙ socioeconomic & & prevalence has been found to vary markedly according to#&  X]  status:NN& &  socioeconomic status. At the earlier stages of the smoking epidemic, it is frequently the higher social classes who can afford cigarettes who smoke most. As health campaigns take effect, prevalence tends to fall first among this better educated group with the result that lower socioeconomic classes have the highest prevalence, with the gap widening over time.#&  X  The need for disease surveillance 55In addition to collecting and monitoring data on groupspecific tobacco use patterns, it is equally important to monitor trends in diseases related to tobacco use. This is much more difficult and expensive and as a result, reliable cause of death data to assess changes in tobaccoinduced diseases are not widely available in developing countries. 55It is important to keep in mind that monitoring diseases in populations is at best an indirect indicator of the health hazards of smoking since some diseases are caused by several risk factors or exposures. However, among all diseases, lung cancer can be a useful index of  X- the overall extent of the health consequences of smoking in populations where it is known that  X lung cancer is strongly associated with smoking.  XJ    MEASURING THE USE OF TOBACCO 55In most countries, smoking is the foremost way in which tobacco is used and is the cause of most of the disease burden arising from tobacco. Thus, the term "smoking" is used in this section. However, in some countries, smokeless tobacco use is widespread, and is a major cause of disease, such as oral cancer. In these countries, smokeless tobacco use should be monitored separately. 55Globally, manufactured cigarettes are the most common tobacco product. Therefore, the extent of tobacco use is most satisfactorily and easily measured in terms of cigarette consumption. However, vast amounts of tobacco are also smoked in ways that are difficult to measure because the quantity of tobacco used by individuals during a single smoking occasionI*-++ cannot be easily assessed. This is particularly true of handrolled cigarettes, bidis, pipes, and a range of other handmade smoking devices.  X 55The two ways of measuring smoking levels and patterns are by population surveys and  Xv by per capita adult consumption (based on sales data or trade and production data). Provided that the limitations of each approach are taken into consideration, both methods can be valuable, each complementing the other to give a more complete understanding of the extent of the tobacco epidemic.  XI  Smoking prevalence surveys 55Surveys of the prevalence of tobacco use in a population are discussed in more detail in Chapter VIII but are briefly referred to here to distinguish them from methods to ascertain tobacco consumption. Prevalence surveys measure individual or group behaviour, such as differences according to gender, age group, socioeconomic status, race/ethnicity, immigrant status, educational level or occupation. They can be influenced by such factors as cultural norms or the social acceptability of smoking, or by perceived confidentiality or fear of disclosure. Differences will arise according to the manner in which information is obtained " whether via face to face interview, questionnaire, or telephone survey. Biochemical validation of data, for example, by testing for cotinine in the saliva although difficult and costly to  XH obtain, is scientifically valuable and should be encouraged whenever feasible. 55National probability surveys provide reasonably valid estimates of the prevalence of cigarette smoking (usually within 13 percentage points of the biochemically validated estimates). Underreporting of prevalence may be more common in societies where, for example, smoking is unacceptable among women or among certain religious groups. Prevalence is likely to be underreported among groups where the demand for abstinence is high, such as cardiac patients and pregnant women who have been advised to quit. Smokers may underreport the number of cigarettes they smoke each day, most likely by rounding down. 55Underreporting of smoking also appears to be more common among adolescents thanG*-++ among adults. Prevalence estimates for adolescents, especially younger adolescents, are lower in household settings than in schoolbased surveys, most likely because of the respondents' concerns about privacy. Recent innovations to protect confidentiality in the household setting have resulted in estimates of prevalence among adolescents that are about twice as high as were previously observed.  X.  Per capita cigarette consumption data  X  55 Per capita consumption (the total number of cigarettes sold or otherwise estimated to have been consumed, divided by the population size) can be used to assess trends in the  X smoking of manufactured cigarettes within the population concerned. In the past, it has been  X common to estimate per capita consumption based on the total population. However, it is much more appropriate to relate consumption to the age group most likely to smoke  X cigarettes. In most societies, children do not account for a very significant proportion of all tobacco consumed. Since most population statistics are available for 5year age groups, the most convenient denominator (as used by WHO) is the population 15 years old and over. To assess the consumption of cigarettes (or other tobacco products) in a population, it is recommended to calculate consumption per adult aged 15 years and over. This will facilitate global monitoring of the epidemic, although it is recognized that the age chosen often differs between countries. While this does not invalidate the estimates, it makes comparisons between countries very difficult. 55Assuming that excise tax and trade data are reliable, and that the population census data are accurate, the estimated per capita data can give a fairly accurate measure of the smoking behaviour of a society as a whole. Per capita consumption can be used as an indicator of changes resulting from increases in price or taxation, or as a measure of the effectiveness of antismoking campaigns or legislation. However, it cannot provide information on the smoking behaviour of specific groups, nor identify any changing patterns of smoking within different groups that may be contributing to overall changes among the population as a whole. For example, over a period of years there could be a decrease in consumption among men and an increase among women, yet there may be no apparent change in the overall per capita consumption figures. Similarly, per capita consumption could be falling, despite very markedK*-++ increases in smoking by members of younger age groups. CONCLUSION: KEY ISSUES FOR EPIDEMIOLOGICAL SURVEILLANCE  Xt  * 55Data and information on tobacco use and its impact on health, and particularly how the epidemic is evolving, are useful for supporting tobacco control efforts. Countries are encouraged to collect data for their own populations: local evidence is both scientifically interesting and politically useful.#5  XE  * 55The complexity of the tobacco epidemic, with the characteristic long delay between persistent smoking and its full health effects, means that data need to be collected on current tobacco use patterns as well as on diseases causally related to smoking.#5  X  * 55Both tobacco use prevalence and per capita adult consumption are key indicators of current population exposure to tobacco and should be monitored regularly.#5  X  * 55Other data and information are also important as part of the policy support system, including economic data, public opinion surveys and information about the implementation of tobacco control measures, specifically legislation on advertising and marketing practices.#5  X  * 55It is critical that data on prevalence and consumption (as well as subsequent mortality data) be collected for key subgroups of the population. Tobacco use patterns vary widely among population subgroups, as does mortality from tobaccorelated causes. Effective policy responses should be based on reliable local information about tobacco consumption, smoking prevalence, and the health consequences of tobacco use.