Tobacco use is a leading preventable cause of premature death around the world. To curtail the globalization of the tobacco epidemic and decrease the prevalence of tobacco use, in 2003, the UN World Health Organization proposed the Framework Convention on Tobacco Control (WHO-FCTC) [1
]. The WHO-FCTC outlined steps to reduce the demand and supply of tobacco. Member states agreed to show political commitment in implementing WHO-FCTC objectives: monitoring tobacco use; increase prices and tobacco taxes; restrict product content, marketing, and sales; provide education and public awareness; and protect the population from secondhand tobacco smoke.
In Latin America (LA), 12 countries signed and ratified the treaty and the region has been noted for its commitment to tobacco control [2
]. Substantial declines in smoking prevalence have been observed [4
], although the rate of decline has not been homogeneous within countries, with tobacco consumption disproportionally burdening urban areas [2
]. Within-country variation in tobacco consumption is a function of sociodemographics [5
], but also of the ability of cities to implement and enforce tobacco control policies. Efforts to monitor policy implementation have focused on the national level while efforts at the city-level have been scarce [7
]. City-level information is key, considering that governance structures for implementation and compliance are usually located at this level.
Smoking initiation is most common during adolescence [8
], a period of life highly susceptible to social influence that is exploited by tobacco companies [9
]. Consequently, monitoring youth tobacco use is a key element of the WHO-FCTC. In 1999, the WHO launched a series of repeat cross-sectional school-based surveys through the Global Youth Tobacco Survey initiative (GYTS) [10
], with ample participation from Latin American countries [9
]. Work to date that used LA GYTS data featured one to three countries [11
] or focused only on a single domain (cessation [16
]); still, a multi-country perspective considering various domains and focusing on city-level estimates has not been provided.
To date, no report has provided an assessment of how tobacco policies in Latin America have translated into perceived changes in the urban tobacco landscape as reported by adolescents. We aimed to characterize the tobacco policy environment in six Latin American counties over the past two decades using the WHO reports on tobacco control in the region. Then, we used GYTS data from 31 cities with information before and after the WHO-FCTC and selected key tobacco indicators to estimate their pre-FCTC prevalence and change observed in the post-FCTC period.
2. Materials and Methods
2.1. Policy Indicators: The MPOWER Framework
To accomplish the objectives of the FCTC, in 2008 WHO introduced the MPOWER package to support policy implementation. The MPOWER package was the first in a series of WHO reports to track the status of the tobacco epidemic and the impact of interventions to stop it [17
]. The MPOWER acronym represents a package of policies with six domains: 1. Monitor tobacco use and prevention policies; 2. Protect people from tobacco smoke (protect from exposure to second-hand tobacco smoke, mostly implemented via indoor smoking bans); 3. Offer help to quit tobacco use (offer cessation services); 4. Warn about the dangers of tobacco; 5. Enforce bans on tobacco advertising, promotion and sponsorship (tobacco marketing); 6. Raise taxes on tobacco (price of tobacco). For each of these domains, the MPOWER data provided a composite score that measures its overall strength and the detailed information on individual policies in these policy dimensions.
2.2. Policy Data
WHO’s reports on the global tobacco epidemic (published for years 2007 (the first year available), 2008, 2010, 2012, 2014) identify the strength of country-level policies according to the MPOWER framework. Data in the reports are based on the existing legislation up to December of the monitoring year, irrespective of implementation status. The reports rank each country on policies that correspond to select MPOWER domains: protection from secondhand tobacco smoke, cessation services, warnings about the dangers of tobacco, bans on advertising, and tobacco prices. WHO rated the domains on a scale of 1 to 5, with 1 representing lack of data, 2 (none/weak) represents no policy or very slight policy, 3 (good) indicates that a policy exists but is missing breadth and detail, 4 (very good) indicates that policies have good breadth but are missing important details, and 5 (excellent) indicates that policies have breadth and also include important details.
We selected five domains and three available years: 2007, 2010, and 2014. The first year available was 2007 (two years after FCTC ratification), thus, we used 2007 to proxy the presence of policies immediately following ratification. Year 2010 represented the mid-period (five years post-ratification), and year 2014 represented the later-period (nine years post-ratification). Policy progress was defined as the relative change from 2007 to 2014 ((2014–2007)/2007).
2.3. Youth Tobacco Survey
The GYTS is a series of school-based surveys that collect data on students aged 11–19 years or older using a multistage sample design, selecting schools and classes. First, public or private schools are selected with probability proportional to enrollment size. Second, classrooms are chosen randomly within schools, and all students in each selected class are eligible for participation. GYTS survey representativeness vary by country and depends on the availability of funding, in some cases it is restricted to the capital city of the country (e.g., Argentina), while in others it involves several cities to produce estimates representative of the national level (e.g., Mexico).
Between 2000 and 2015, 35 countries in Latin America participated in at least one GYTS survey. The current study is embedded in a project that aims to assess health in cities [18
], consequently, we selected countries where cities could be identified in the dataset. Further, our aim was to analyze changes in survey responses in cities from the pre- (2000–2005) to a post-FCTC (2006 onward) period. We selected six countries with data in both periods. For countries with three available surveys, we selected the oldest and the newest surveys for the analysis: Argentina (2000, 2007); Brazil (2002, 2006); Chile (2000, 2008); Colombia (2001, 2007); Mexico (2003, 2011); and Peru (2000, 2007). Thus, the post-FCTC period is represented by years 2006–2011, the most recent data with city indicators available for download is of June 2020.
All countries implemented the core items of the GYTS questionnaire: tobacco use, exposure to secondhand smoke, pro- and anti-tobacco media and advertising exposure, access and availability to tobacco and school curriculum. Surveys were slightly different by country and year, requiring harmonization. The final analytic sample included 132,065 students from 31 cities across all six countries.
2.3.1. Transformation of Youth Survey Data into MPOWER
We used 13 tobacco questions from the GYTS survey and grouped them according to the six domains. Each domain reflected one or more of the provisions of the WHO FCTC: tobacco use (“Monitor”); smoke at home and smoke in other places (“Protect”); anti-tobacco education in school (“Warn”); and media and advertising, retailer refusal to sell cigarettes and offered a free cigarette (“Enforce”) (Table 1
). A brief description of the compiled data follows, and details are shown in Supplementary Table S1
We used two survey items to assess past 30-day use of tobacco (“Monitor”). Questions focused on the number of days when tobacco was used or on the type of tobacco product used. Students were classified as using tobacco if they consumed tobacco on at least one day of the past 30 days, or if they answered ‘yes’ to the use of any type of tobacco in the last 30 days. We used two items to assess “Protection” from second-hand smoke (SHS), to estimate the proportion of students who were exposed to SHS at home or outside. We assessed “Warn” using the average of three items that capture the anti-tobacco education received at school, as presented in Table 1
(respondents were included if they answered to at least 2 out of 3 items); a higher average indicates that students received more information at school. “Enforce” was represented by three components: “media and advertising”, “offered free cigarette” and “refusal to sell”. Media and advertising was derived by calculating the average of three survey items (see Table 1
); a lower average means less exposure to tobacco media and advertising. “Offered free cigarette” was the percentage of students who had been offered free cigarettes by people related to the tobacco industry, and “refusal to sell” was the percentage of students who were unable to purchase a cigarette, among those who reported trying to buy a cigarette.
2.3.2. Survey Data Analysis
Survey data were combined into MPOWER domains and summarized as weighted percentages with 95% confidence intervals and weighted means with 95% confidence intervals, using the sampling weights calculated at each country to take into account the complex sampling design. Sampling weights adjust for non-response (by school, class, and student) and probability of selection at the school, class, and student levels. Weights were summed by grade and gender to the population of school children in the selected grades in each sample site. For each city, we calculated summary statistics (proportions (95% confidence intervals) and means (95% confidence intervals)) for the pre-FCTC ratification period (2000–2005) and the post-FCTC ratification period (2006–2011). To assess the change in survey responses, we calculated prevalence ratios (for binary variables, later:earlier period prevalence with 95% confidence intervals) and weighted mean differences (for continuous variables, later:earlier periods with 95% confidence intervals) for each city and pooled across cities to estimate a country-level weighted average. All survey analyses were conducted in IBM SPSS Statistics package (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM Corp) to take into consideration sampling weights as defined by GYTS [19
]. Forest plots—used to summarize the prevalence and mean differences across cities—were built in STATA statistical software (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP) with the “metan” command, using the expanded population on each city to estimate a country specific weighted average.
We aimed to analyze the changes in the tobacco policy landscape in six Latin American countries and to assess key indicators for tobacco control in adolescents from 31 cities within those countries. According to WHO, tobacco control policies improved between the pre- and post-FCTC period, Chile and Brazil were rated excellent, Colombia and Argentina had the largest improvements, while Peru and Mexico had the lowest scores. Based on the self-report of adolescents in the GYTS surveys, we observed improvements in tobacco use, protection against tobacco smoke at home, tobacco education and in the prevalence of refusals to sell cigarettes to adolescents. Smaller reductions were observed in secondhand smoke outside the home, and in tobacco representatives offering free cigarettes to adolescents. Exposure to media and advertising remained largely unchanged.
Tobacco use is one of the most important indicators in the MPOWER strategy, as it reflects the net impact of policies on tobacco consumption. Overall, from the pre- to the post-ratification period, we observed a net decrease in tobacco use by adolescents in the capital cities of Argentina and Chile, and in the majority of cities in Brazil, Peru and Mexico. Interestingly, according to WHO reports (2007–2014), Colombia has advanced steadily in tobacco control, yet the prevalence of adolescent tobacco use in Bogota (30.6%) remained stable between ratification periods (years 2001 to 2007). It is unknown whether adolescent smoking remained high through 2014. In contrast, Mexico and Peru were rated by WHO as lower in policy implementation, yet adolescent surveys suggested decreases in tobacco use in most of their cities.
Protection of adolescents from SHS at home and in public places is a fundamental intervention to reduce smoking initiation in this age group [20
]. Pre-ratification and change in SHS protection were higher at home than in public places, which could be related to a reduction in the prevalence of adult smoking and/or influenced by SHS regulations in public places [21
]. For SHS exposure in public places we observed reductions in Argentina, Brazil, Chile and Colombia; reductions in Mexico varied across cities, likely reflecting differences in regulation at the state level [22
]. Peru was the only country where all cities experienced an increase in SHS exposure in public places. For Peru, the last data point available was 2007, which precedes the 2011 approval of Law 29517, which banned smoking in public places.
Anti-tobacco education improved or remained high in most areas. Educational programs at school are considered to be an important part of any integral strategy to reduce tobacco use in adolescents [23
]. The prevalence of tobacco education was particularly high for Brazilian cities, relative to other countries, yet we did not observe an increase in the post-FCTC period. Brazil ratified the FCTC in 2006, which is also the last year for which GYTS was available; thus, our assessment failed to capture the 2008 Mais Saude: Direito de Todos program, which introduced important tobacco educational and legislative changes [24
]. Improvements did not occur in Bogota, Colombia. However, the last Bogota survey year was 2007, which may not have fully captured the impact of inter-sectoral policies and school-based education programs that were developed in 2006 [25
Important advances were also made in refusals to sell cigarettes to adolescents. Pre-ratification, in Argentina, Brazil and Chile refusals to sell cigarettes to minors were rare (<20%), but they increased in the post-ratification period, particularly in Chile (up to 50%). Peru and Mexico had a higher pre-ratification prevalence of refusals, yet trends were heterogeneous by city. Adolescents reported similar levels of exposure to tobacco in the media and to tobacco advertising in the pre- to post-ratification period. This is particularly worrisome, considering that smoking initiation is related to exposure to tobacco advertising [26
]. Approximately 60% of youth reported seeing pro-tobacco media and advertising, with no change in the post-ratification period. The only country that showed substantial decreases in exposure to marketing in the post-ratification period was Peru, likely related to the implementation of the General Law for the Prevention and Control of Risks Related to Tobacco Consumption, approved on April 2006, which restricted marketing to minors [29
Three countries (Brazil, Mexico, Peru) had GYTS information in more than one city and all displayed large within-country heterogeneity in prevalence and change for most indicators. This is an important finding, as it suggests that we need to reconsider the relevance of cities as key geographical and political areas to advance tobacco control. Unfortunately, we did not have multi-city information for Argentina, Chile, and Colombia. This probably reflects the chronic scarcity of funding to monitor the tobacco epidemic in the region [30
]. However, it could also reflect a paradigm of tobacco control monitoring that assumes that the most relevant level of monitoring is the country-level. This paradigm is reflected in WHO reports and across prior studies, which focus mostly on country-level or in capital cities, assuming that they are roughly representative of other urbanized areas within the country [7
]. The heterogeneity we observed suggested that tobacco control efforts face different challenges and are advancing at different speeds across cities. This should force us to reconsider the design of surveys, to cover a sample of highly diverse cities in order to inform barriers and facilitators of effective tobacco control.
Our study has several limitations. GYTS data is limited to adolescents that attend school, thus, it fails to capture the experience of adolescents who do not attend school (such as adolescents who work) and who may use tobacco more frequently [31
]. Country comparisons should consider that survey years were different across cities, being more similar for Argentina, Brazil, Chile, Colombia and Peru (all within the 2000–2008 period) than for Mexico (2000 to 2011). We tried to capture pre- to post-FCTC changes, yet countries made large changes to legislation and policy since our last year of GYTS data (2011); thus, our analysis should not be considered a current evaluation of policy, but a historical analysis of the early impact of the FCTC. We did not exclude any recent surveys. Rather the data we analyzed represent the data that were available as of June 2020. The GYTS program is informative and efforts should be made to produce a new waves of data, the lack of continuity in monitoring tobacco use in adolescents is worrisome, since such a large lag in data will negatively influence policy decision making. The effectiveness of the tobacco control response depends on having valid, representative and up to date data to keep track of changes in the epidemic, detect the impact of tobacco industry interference, maintain public awareness, and identify specific targets for new tobacco control policies [32
Between the pre- and the post-FCT ratification period, we observed advances in tobacco control for adolescents across Latin American countries and cities. Despite overall improvements in tobacco policies and declines in exposure to tobacco, policies related to media advertising and promotions and secondhand smoke were in need of strengthening. There was wide variation in adolescent exposure to tobacco between cities (within countries), which suggested major challenges in implementing policies at the local level. New GYTS data waves are needed to monitor how tobacco control is advancing in adolescents. Future studies need to further explore the reasons why the response to tobacco control is so heterogenous within countries and identify local solutions for successful implementation.