METHODS

Sample selection

The GYTS is a school based survey of a defined geographic site that can be a country, a province, a city, or any other geographic entity. The following steps are followed for the sample selection.

• First, because the GYTS focuses on students aged 13–15 years, the country research coordinator identifies the grades in their educational system that correspond to these ages.

• Second, the research coordinator prepares a database of schools that include the identified grades. Each school is designated a unique identifier to facilitate school selection. The number of students enrolled in each school grade to be surveyed is added to the database. This database forms the survey sample frame. The amount of work involved in creating this database varies from country to country. In some countries, the creation of the sampling frame was the most labourious and time consuming part of the GYTS (for example, the individual states in India).

• Third, the database is sent to the CDC, where the GYTS sample is drawn using a two stage cluster sample design. Schools are selected with probability proportional to school enrolment size during the first stage, and then classes within participating schools are selected as a systematic equal probability sample with a random start during the second stage. All students in the selected classes are eligible to participate in the survey. For this two stage sample design, statistical analysis conducted by CDC^10,11 has found that, for most sample designs, a minimum of 1500 completed student interviews is needed to obtain a precision level of ± 5% for a given estimate. WHO and CDC use this information to work with the countries to determine the sample size of schools and students for each site. The desired sample size is then adjusted for anticipated non-response at the school, class, and student levels. The very large samples of schools in South Africa, USA, and Philippines were done to provide regional or population subgroup estimates within the country.

GYTS questionnaire

The GYTS questionnaire is a self administered, school based instrument consisting of a “core” set of questions that are used by all countries, unless the information is not relevant in the country (for example, pro-cigarette advertising is not permitted in Singapore).* In addition, there is an optional set of questions from which a country can draw depending on its needs and priorities. The 2001 core questionnaire consists of 56 questions and includes items on the following topics: prevalence of tobacco use, age of initiation, exposure to tobacco advertising, perceptions and attitudes on behavioural norms with regard to tobacco use among young people, media and advertising, legislation, economics, school curriculum, and ETS. The 2001 core questionnaire differs from the 1999 and 2000 core GYTS questionnaire, by the addition of the economics questions and the deletion of the question on alcohol and/or drug use when last smoked cigarettes. Specific guidelines are followed for questionnaire translation into local languages and pilot testing. The final questionnaire is the responsibility of each participating country. This paper only includes data from the core questions.

Since classes were carefully identified to correspond to students 13–15 years of age, the majority of selected students were in this age group. However, all students in the selected classes were eligible to participate, without regard to their age, therefore there were some students who were younger than 13 years or older than 15 years. Because the objective of this paper is cross country comparisons of same aged children (13–15 years), respondents younger than 13 or older than 15 years have been excluded from the analysis.†

Survey administration procedures

WHO, CDC, and the research coordinators from the countries who participated in the 1999 surveys developed a GYTS research manual, which includes detailed procedures for administering the GYTS in schools. The manual is modified for each subsequent GYTS training to meet the specific needs of the countries in those trainings. The manual includes information on obtaining school participation, procedures for completing all survey forms, protocol in the classroom, and instructions for returning the completed forms to CDC for data processing. The GYTS uses a generic answer sheet, which allows for 99 questions, with eight response categories available per question. There are no open ended questions, no skip patterns, and no multiple response questions in the GYTS. The completed answer sheets are scanned through an optical reader. Edits for consistency and out-of-range responses are performed for each question. The quality of the GYTS data has been very high. Consistency failures or out-of-range responses rarely exceed 5% per question.

The GYTS is administered during one class period. GYTS administration procedures were designed to protect students' privacy by assuring that student participation was anonymous and voluntary. Before the survey was administered each country followed local procedures for obtaining parental permission and institutional review.

Analysis

The GYTS data are weighted to adjust for sample selection (school and class levels), non-response (school, class, and student levels), and post-stratification of the sample population relative to the grade and sex distribution in the total population. The computer program SUDAAN¹² was used to compute standard errors, 95% confidence intervals, and weighted prevalence estimates. The weighting factor consisted of the following formula:where

• W1 = the inverse of the probability of selection for each school

• W2 = the inverse of the probability of selection of each classroom within each selected school

• f1 = a school level, non-response adjustment calculated by school enrolment size category (small, medium, large); school non-response is calculated within each tertile

• f2 = a class level, non-response adjustment factor calculated for each school

• f3 = a student level, non-response adjustment factor calculated by class

• f4 = a post-stratification adjustment factor calculated by sex and grade.

Training and follow up

WHO and CDC developed the GYTS to enhance the capacity of countries to design, implement, and evaluate their tobacco prevention and control programmes. There are four phases to this capacity building process‡:

• First, the methodology and procedures for conducting the GYTS are taught to country research coordinators at regional workshops.

• Second, upon completion of their GYTS, country research coordinators participate in a workshop which focuses on data analysis, report writing, and dissemination.

• Third, regional programme development workshops are held to assist countries in identifying potential programmes and interventions that can be included in their tobacco control programme.

• Fourth, ongoing technical assistance is provided to countries by WHO and CDC as the countries implement programmes, conduct repeat GYTS, and evaluate their programme's effectiveness.¹³

RESULTS

This section presents cross country comparisons for students aged 13–15 years across the 75 sites in 43 countries and the Gaza Strip/West Bank region, including core questions in each topic area. The number of sites included for each topic may vary for two reasons: (1) some sites did not ask all of the core questions; and (2) in some sites the student sample size for a given table cell was less than 35 cases, which is considered statistically unstable.

Response rate

Table 1 shows the school, student, and overall response rates for each country or site within each country. The school response rate ranged from 100% to 68.8% (median 98.0%); the student response rate ranged from 99.7% to 70.1% (median 86.8%); and the overall response rate (school rate * student rate) ranged from 96.8% to 55.2% (median 84.1%). The number of students aged 13–15 years who completed the GYTS by country/site ranged from 129 in Montserrat (only one school was open on the island because of volcanic activity) to 16 416 in the USA. In total, over 230 000 students in nearly 3 500 schools have completed the GYTS.

Prevalence

Ever smoked cigarettes§

The overall median per cent of students who had ever smoked cigarettes, even one or two puffs, was 33.0% (table 2). The highest per cent who ever smoked cigarettes was in the Northern Mariana Islands (79.8%), and the lowest in Tamil Nadu, India (3.4%). Over 70% of students reported having ever smoked cigarettes in three of the 75 sites (Santiago, Chile; Kiev, Ukraine; and Northern Mariana Islands) and 50% reported having ever smoked in 18 sites. Only five sites reported ever smoking rates less than 10% (four states in India and Nepal).

Age of initiation

The overall median per cent of students who ever smoked cigarettes, who smoked their first cigarette before age 10 years, was 23.9% (table 2). Manipur, India (87.8%) had the highest rate of smoking initiation before age 10, and the lowest was Buenos Aires, Argentina (6.1%). A total of eight sites, all in India, reported that of the students who smoked, more than half smoked their first cigarette before the age of 10 years. Only five sites reported a prevalence of under 10% for students smoking their first cigarette before the age of 10 years.

Current any tobacco use

The overall median per cent of current use of any tobacco product (smoked cigarettes or used other tobacco products on one or more days in the 30 days preceding the survey) was 18.7% (table 2). The highest per cent currently using any tobacco product was in the Nagaland, India (62.8%), and the lowest in Goa, India (3.3%). Over 50% of the students reported current use of any tobacco product in six states in India; Northern Mariana Islands; and Palau. Less than 10% of the students currently used any tobacco product in nine of the 75 sites (Virgin Islands (Am.); three states in India; Nepal; Sri Lanka; Shandong and Tianjin, China; and Singapore).

Current cigarette smoking

Current smoking was defined as having smoked on one or more days in the 30 days preceding the survey. The overall median rate for current cigarette smoking was 13.9% (table 2). The highest rate was found in Coquimbo, Chile (39.6%), and the lowest in Goa, India (0.5%). More than one third of students were current smokers in six sites (Coquimbo, Santiago, and Valparaiso, Chile; Moscow, Russian Federation; Kiev, Ukraine; and the Northern Mariana Islands); whereas less than 10% were current smokers in 24 of the 75 sites.

Current use of other tobacco products¶

Students were asked if they had used any form of tobacco other than cigarettes on one or more days in the 30 days preceding the survey. The overall median rate for current use of other tobacco products was 8.8% (table 2). The highest rate of use of other tobacco products was in Palau (53.5%), and the lowest in Jakarta, Indonesia (2.5%). Over one third of the students currently used other tobacco products in nine sites: seven states in India; Northern Mariana Islands; and Palau. Use of other tobacco products was less than 10% in 46 of the 73 sites.

Smoked six or more cigarettes per day

The overall median per cent of current cigarette smokers who, on the days they smoked, smoked six or more cigarettes per day was 9.4% (table 2), while the highest rate was in Moscow, Russian Federation (29.9%), and the lowest rate was in Tarapoto, Peru (1.0%). In six sites (Cross River State, Nigeria; Buenos Aires, Argentina; USA; urban Poland; Moscow, Russian Federation; and Singapore) over 20% of the students smoked six or more cigarettes per day; whereas less than 10% smoked six or more cigarettes per day in 37 of the 68 sites.

Perceptions and attitudes

Students who smoke have more friends

The students were asked a series of questions regarding their perception and attitudes toward smoking (table 3). The overall median per cent of students who responded “boys who smoke have more friends” was 28.0%, with Sikkim, India (60.7%) reporting the highest rate, and Colonia, Uruguay (8.7%) the lowest. More than 30% of the students responded “boys who smoke had more friends” in 29 of the 71 sites, whereas this figure was less than 10% in three of the 71 sites (Buenos Aires, Argentina; Colonia, Uruguay; and Guangdong, China). When asked if students thought “girls who smoke have more friends” the overall median was 16.8%, the highest rate was in Sikkim, India (50.2%), and the lowest in Jakarta, Indonesia (3.9%). More than 30% of the students responded “girls who smoke had more friends” in six sites (Ghana; South Africa; three states in India; and Fiji), whereas this figure was less than 10% in 10 of the 71 sites, including all four sites in China.

Students who smoke are more attractive

Students were also asked if they thought “smoking makes boys look more attractive” and if “smoking makes girls look more attractive” (table 3). Overall, the median rate of students who responded “smoking makes boys look more attractive” was 13.5%, with the highest per cent in Manipur, India (64.0%), and the lowest in Blantyre, Malawi (2.1%). More than 30% of students thought smoking makes boys look more attractive in 12 sites (North and Middle West Bank; eight states in India; Sri Lanka; and Shandong, China), whereas less than 10% responded “smoking makes boys look more attractive” in 17 of the 71 sites. Overall, the median rate who responded “smoking makes girls look more attractive” was 10.0%, with the highest in Sikkim, India (42.0%), and the lowest in Montserrat (1.6%). More than 30% of students responded “smoking makes girls look more attractive” in six sites (Gaza Strip; North West Bank; and four states in India), whereas this figure was less than 10% in 35 of the 71 sites.

Access and availability

Students were asked questions regarding their access to cigarettes and the availability of cigarettes to them.** The overall median rate of students who currently smoke cigarettes who usually smoke at home was 22.2%, with Assam, India (71.6%) reporting the highest rate, and Moscow, Russian Federation (4.8%) the lowest (table 4). Over 50% of the students usually smoke at home in seven sites (Guyana; Port-au-Prince, Haiti; Jamaica; three states in India; and Tianjin, China) and less than 10% usually smoke at home in 11 of the 65 sites.

Students who currently smoke cigarettes were also asked how they usually got their own cigarettes. Overall, the median rate for purchasing cigarettes in a store was 42.7% (table 4). Students in Manipur, India (86.3%) were the most likely to have purchased cigarettes in a store, with the least likely in the USA (9.6%). Over 50% of the students purchased their cigarettes in 30 of the 68 sites, while this figure was less than 10% only in the USA.

Students were asked if they had been refused the purchase of cigarettes in a store because of their age. Overall, for students who currently smoke cigarettes and who purchased their cigarettes in a store, the median rate for not being refused purchase was 83.0%. The site with the highest per cent not refused purchase was in Assam, India (98.1%), and the lowest was in the Philippines (46.5%). Over 70% of young people were not refused purchase because of their age in every site except eight (Monterrey, Mexico; USA; Jordan; rural Poland; Fiji; Northern Mariana Islands; Philippines; and Singapore).

Cessation

Students were asked if they wanted to stop smoking now and if they had tried to stop smoking in the year preceding the survey. The overall median per cent of current smokers who want to stop smoking now was 68.4% (table 5). Shandong and Tianjin, China had the highest per cent of current smokers who wanted to stop (86.9%), and Manipur, India (19.6%) had the lowest. Over 80% of current smokers wanted to stop smoking in 10 sites (Port-au-Prince, Haiti; Tarapoto, Peru; rural Poland; Mizoram and Nagaland, India; Jakarta, Indonesia; Shandong and Tianjin, China; Northern Mariana Islands; and the Philippines). Less than half of current smokers wanted to stop smoking in nine sites (Buenos Aires, Argentina; Barbados; Santiago and Valparaiso, Chile; Colonia, Uruguay; Jordan; and Manipur, Sikkim, and Tripura, India).

Almost two thirds of current smokers stated that they had tried to quit smoking during the year preceding the survey (median 63.1%) (table 5). Jakarta, Indonesia (91.0%) had the highest per cent of current smokers who had tried to quit during the past year, and Sikkim, India (8.4%) had the lowest. Over 80% of current smokers had tried to quit smoking in the past year in four of the 60 sites (St Vincent and the Grenadines; rural Poland; Jakarta, Indonesia; and Philippines), whereas less than 50% had tried to quit smoking in the past year in seven sites (six states in India and Sri Lanka).

Media and advertising

Anti-smoking

Students were asked two questions regarding the extent to which they had seen anti-smoking messages either in the media or at sporting/other events during the month preceding the survey. The overall median per cent who had seen any anti-smoking media message was 80.4% (table 6). The highest exposure to anti-smoking media messages was in Bihar, India (97.6%), and the lowest in Manipur, India (57.7%). Ninety per cent or more of students had seen anti-smoking media messages in 11 of the 74 sites. Less than 70% of students had seen anti-smoking media messages in only eight of the 74 sites (Ghana; Cross River State, Nigeria; Manicaland, Zimbabwe; Buenos Aires, Argentina; Port-au-Prince, Haiti; North West Bank; and Manipur and Nagaland, India).

The overall median per cent of students who had seen anti-smoking messages at sporting/other events was 76.2% (table 6). The highest exposure to anti-smoking messages at sporting/other events was in Manipur, India (95.0%) and the lowest in West Bengal, India (35.7%). Over 90% of students had seen anti-smoking messages at sporting/other events in nine of the 71 sites (six states in India; Jakarta, Indonesia; Guangdong, China; and Singapore). Less than 70% of students had seen anti-smoking messages at sporting/other events in 17 of the 71 sites.

Pro-smoking

Students were asked about their exposure to pro-cigarette ads during the month preceding the survey on billboards, in newspapers/magazines, and at sporting/other events. The overall median per cent of students who had seen ads for cigarettes on billboards was 78.3% (table 6). Exposure to ads for cigarettes on billboards was highest in Bihar, India (98.6%), and lowest in Montserrat (47.2%). Over 90% of students were exposed to ads for cigarettes on billboards in 14 of the 68 sites (Buenos Aires, Argentina; Santa Cruz, Bolivia; Costa Rica; Monterrey, Mexico; Maldonado, Motevideo, and Rivera, Uruguay; urban Poland; Moscow, Russian Federation; Bihar, Manipur, and West Bengal, India; Jakarta, Indonesia; and Nepal). Less than 70% of students were exposed to ads for cigarettes on billboards in 20 of the 68 sites.

The overall median per cent of students who had seen ads for cigarettes in newspapers or magazines during the past 30 days was 73.0% (table 6). The per cent who saw ads for cigarettes in newspapers or magazines was highest in Bihar, India (96.6%), and lowest in Montserrat (30.4%). Over 90% of students saw ads for cigarettes in newspapers or magazines in two sites (urban Poland and Bihar, India). Less than 70% of students were exposed to ads for cigarettes in newspapers or magazines in 33 of the 71 sites.

The overall median per cent of students who saw ads for cigarettes at sporting/other events in the past month was 79.7% (table 6). The per cent of students who saw ads for cigarettes at sporting/other events was highest in Costa Rica (100.0%) and lowest in Montserrat (33.6%). Ninety per cent or more of students saw ads for cigarettes at sporting/other events in 11 sites (Costa Rica; Montevideo, Uruguay; seven states in India; Jakarta, Indonesia; and Nepal). Less than 70% of students saw ads for cigarettes at sporting/other events in 20 of the 67 sites.

Receptivity

Approximately one in six students have “something” with a cigarette brand logo on it (median 16.7%) (table 6). Students in the South West Bank (34.6%) were the most likely to have an item with a cigarette brand logo, and students in Bihar, India (1.1%) were the least likely. Over one in five students had an item with a cigarette brand logo on it in 22 of the 74 sites, and less than 10% of students had an item in nine of the 74 sites. One in 10 students reported they had been offered free cigarettes by a representative of a tobacco company (median 10.6%) (table 6). This practice was most likely in Poland (urban 48.9%, rural 42.2%) and Jordan (24.8%), and least likely in Bihar, India (0.6%). Over 15% of the students had been offered free cigarettes in nine of the 66 sites.

Environmental tobacco smoke

Exposure in the home

Students were asked two questions regarding their exposure to second hand smoke (table 7). Almost half of the students reported that they were exposed to second hand smoke from others in their home (median 48.9%) (table 7). Students in Meghalay, India (79.8%) had the highest exposure and students in Lilongwe, Malawi (16.0%) the lowest. Over 70% of students were exposed to others smoking in their home in six sites, all in India. Less than 40% of the students were exposed to others smoking at their home in 31 of the 75 sites.

Exposure in public places

Overall, six in 10 students were exposed to cigarette smoking from others in public places (median: 60.9%) (table 7). Students in Buenos Aires, Argentina (86.7%) had the highest exposure, and students in Blantyre, Malawi (30.4%) the lowest. Over 70% of students were exposed to cigarette smoking from others in public places in 18 of the 75 sites, including those in South America, India, Eastern Europe, and the Pacific Islands. Less than 40% of the students were exposed to smoking from others in public places in only six sites (Blantyre and Lilongwe, Malawi; Huancayo and Tarapoto, Peru; Virgin Islands (Am.); and Goa, India).

Attitudes toward second hand smoke

Students were asked two questions regarding their attitudes toward second hand smoke. Overall, three fourths of students thought smoking should be banned from public places (median 74.9%) (table 7). The highest per cent of students desiring a ban on smoking in public places was found in Sri Lanka (91.4%), and the lowest in Manipur, India (31.4%). Over 80% thought smoking should be banned from public places in 25 of the 70 sites. Less than 40% of the students thought smoking should be banned in public places in only four sites (Manicaland, Zimbabwe; Manipur, Nagaland, and Sikkim states in India).

Overall, almost two thirds of the students “definitely” thought that smoke from others is harmful to them (median 65.5%) (table 7). The highest per cent was in the USA (90.8%), and the lowest in Manipur, India (22.8%). Over 80% of students “definitely” thought that smoke from others is harmful to them in 12 sites (Blantyre and Lilongwe, Malawi; USA; Gaza Strip, North, Middle, and South West Bank; Tamil Nadu and West Bengal, India; and Chongqing, Guangdong, and Tianjin, China). Less than 40% of students “definitely” thought smoke from others is harmful to them in five sites (Cross River, Nigeria; Manicaland, Zimbabwe; and Manipur, Nagaland, and Sikkim states in India).

School curriculum

Students were asked a series of questions about what they had been taught during the past school year concerning the harmful effects of tobacco. Overall, the median per cent of students who reported having been taught in school about the dangers of tobacco use was 50.8% (table 8). Students in Guangdong, China (83.0%) were the most likely to have been taught about the dangers of tobacco, and students in Bihar, India (2.7%) the least. Over 60% of the students had been taught about the dangers of tobacco in 15 of the 70 sites. Less than 30% of the students had been taught about the dangers of tobacco in nine sites (Santiago, Chile; and eight states in India).

Students were also asked if they had discussed, in their classes, the reasons why children their age smoke (median 34.4%) (table 8). Students in Jakarta, Indonesia (63.0%) were the most likely to have had these discussions, and students in Bihar, India (1.8%) the least. Only Jakarta, Indonesia had a rate above 60%, whereas 26 sites had rates less than 30%.

DISCUSSION

The GYTS has, for the first time, documented a serious problem in youth tobacco use that is global in nature. The problem is of equal concern in developed and developing countries. Among the 75 sites in 43 countries and the Gaza Strip/West Bank region presented in this report, not a single site had a prevalence rate of current “any tobacco use”, “cigarette smoking”, or “other tobacco use” equal to zero. Thus, from the world population of 6.2 billion people††, 186 million are estimated to be age 13–15 years and currently in school. Further, of the 186 million, an estimated 34.8 million are currently using some form of tobacco and 25.8 million are currently smoking cigarettes. The use of any form of tobacco by 13–15 year old students was greater than 10% in all but nine sites. In addition, almost one in four students who ever smoked cigarettes smoked their first cigarette before the age of 10. Thus, future health consequences of tobacco use and dependency on tobacco appear to be a significant problem facing countries throughout the world. These findings suggest that immediate attention needs to be given to the development of both global and country specific tobacco control programmes.

In addition to providing essential information for the participating governmental jurisdictions, the GYTS data also allow for the identification of important differences among sites. For example, there is an extremely wide range in responses to virtually all questions on tobacco use, with 10- and 20-fold differences between the sites with the highest and lowest rates of tobacco use common. Of particular interest is India, a country of over one billion people, which had both the highest and lowest rates for current use of any tobacco product (62.8% in Nagaland, India and 3.3% in Goa, India). These wide variations in responses within a country underscore the importance of subnational data, and how national estimates can obscure important within country differences.

Asking questions in a standardised manner also allows for comparisons among countries, revealing patterns and suggesting questions that might not otherwise be apparent. For example, in some of the Indian sites with low “ever smoking rates” of under 25% (for example, Assam, Manipur, Sikkim, and Tripura) nearly all of the students (over 80%) that have smoked, smoked their first cigarette before the age of 10.

This contrasts dramatically with countries with higher “ever smoking rates” of over 60% (for example, Chile, Poland, Russian Federation, Ukraine, Northern Mariana Islands, and Palau), but have less than one third of students that have smoked, having smoked their first cigarette before the age of 10. What is it that makes children in some parts of the world less likely to smoke, but those that do, start at an earlier age? These and other data from the GYTS should serve to guide and stimulate additional research.

In many ways the GYTS has raised as many questions as it has answered. For example, while Moscow in the Russian Federation had among the highest rates of current cigarette smoking (33.4%), students in Moscow were the least likely to smoke at home (4.8%). Of additional interest is the fact that students tended to believe that smoking provided more social benefits to boys than to girls. For instance, 28% of students thought boys who smoke have more friends, compared to 16.8% who feel that girls who smoke have more friends. Similarly, students were more likely to think that smoking made boys look more attractive than girls (13.5% v 10.0%, respectively). Further research is needed to better understand these observations.

Implications for action

The GYTS data document that in many parts of the world a serious problem of youth tobacco use already exists, and these data also provide insight into ways to shape a public health response. Because of the deadly and addicting nature of tobacco products, and the prevalence of its use among young people, it is clear that we need to change the way in which society views tobacco products and to begin to treat these products commensurate with the harm that they cause.

Specifically, the GYTS data demonstrate that in almost every site and every media, the majority of students had seen cigarette advertisements. Also, in many countries students are offered free cigarettes by tobacco company representatives at social and sporting events and 17% of the students owned an object with a cigarette brand logo on it. While cigarette advertisements are not the only factor influencing young people to smoke, there is abundant evidence that cigarette promotion and marketing efforts influence adolescent smoking behaviour, often to a greater extent than it influences the behaviour of adults.¹⁵ Given the addictiveness of tobacco products and the magnitude of harm they cause, as well as the susceptibility of young people to sophisticated tobacco advertising strategies, severe restrictions on the marketing of tobacco products are prudent public health actions.

The GYTS data show that a vast majority of students are exposed to second hand smoke in public places, and substantial proportions are exposed to tobacco in their homes. Generally, the majority of students knew that tobacco smoke from others smoking was harmful to them, and the vast majority stated that smoking should be banned from public places. These findings reinforce the need for laws which protect children from exposure to second hand smoke.

Survey results indicate that a large percentage, generally a majority of current smokers, have purchased their cigarettes from a store. A vast majority of 13–15 year old current smokers who tried purchasing cigarettes from a store were not refused the purchase because of their age. Thus, there is a need for strong laws prohibiting the sale of tobacco products to minors, and these laws must be enforced.

The GYTS data show that over two thirds of current smokers want to stop smoking. This strongly suggests the need for effective youth cessation programmes. The GYTS data on school education programmes suggest the need for development and implementation of effective tobacco prevention curricula in schools throughout the world.

Limitations

The findings in this report are subject to at least three limitations. First, these data apply only to youth age 13–15 years who attended school and, therefore, are not representative of all persons in this age group. However, in most countries, the majority of young people age 13–15 years attended regular, private, or technical schools. Data on secondary school enrolment are available for 34 of the countries included in this paper.¹⁶ These countries have an average secondary school enrolment ratio of 64, compared to 54 for the world. Second, these data apply only to youth who were in school on the day of survey administration. The median student response rate was 86.8%, and only five of the 75 sites had a student response rate less than 80%. Third, the data are all based on self reports, possibly leading to under or over reporting of behaviour. Although the extent of this under or over reporting of behaviour cannot be determined, some GYTS questions have been analysed and demonstrated good test–retest reliability.¹⁷ As GYTS expands to more countries, it is hoped that there will be more opportunity to compare GYTS findings with findings from other youth health surveys, particularly those that are conducted in multiple countries, such as the HBSC in Europe.

Conclusions

At the beginning of the 21st century, tobacco use among young people is already well established in many parts of the world. Nearly 20% of 13–15 year olds use some type of tobacco product, and among those who smoke cigarettes, nearly 25% smoked their first cigarette before the age of 10 years.

The determinants of youth tobacco use are many and varied. Cultural and religious norms, availability of different types of tobacco products, tobacco control strategies, and, perhaps most importantly, tobacco industry behaviour to promote tobacco use and undercut tobacco control strategies are determining factors. While we do not fully understand all the factors that contribute to the decision to use tobacco, which quickly leads to addiction and eventual adverse health outcomes, we do need to understand better the patterns of use, how the determinants of use interact, and how they differ among countries and cultures.

Systematic global surveillance of youth tobacco use is the essential first step in attempting to prevent the projected epidemic of death and disease that smoking will cause in the 21st century. The GYTS was developed to enhance the capacity of countries to design, develop, implement, and evaluate their tobacco prevention and control programmes. The GYTS provides data which can be used by countries to: (1) evaluate their country specific tobacco control programme; (2) monitor trends in global youth tobacco use; and (3) compare tobacco use among countries and regions.

This paper presents the basic results from 75 GYTS locations (43 countries and the Gaza Strip/West Bank region), with the major goal of making the information available, in a cross country comparison format, to tobacco control programme and policy makers throughout the world. Additional manuscripts are being prepared on sex differences in youth tobacco use, cigarette brand preference, and a multivariate analysis on the differences in tobacco use. Additional research needs to be conducted on country specific comparisons between youth and adult tobacco use rates, estimates of the incidence of initiation in order to determine the direction of tobacco use trends, and evaluation studies in sites that have conducted multiple surveys over time to assess secular trends, as well as the effectiveness of intervention programmes. Country specific data from the GYTS is available at: http://www.cdc.gov/tobacco/global/gyts/GYTS_factsheets.htm.