Background Knowledge about the harms of tobacco use deters initiation and is associated with cessation. Most studies on this knowledge in the general population have been in high-income countries, but the tobacco use burden is increasing in low-income and middle-income countries. We sought to estimate levels of knowledge about tobacco-related diseases in 22 countries and determine the factors associated with differences in knowledge.
Methods We used data from the Global Adult Tobacco Survey (GATS), a nationally representative survey of persons aged ≥15 years. GATSs were conducted from 2008 to 2013 in 22 low-income and middle-income countries. Information was gathered on tobacco-related knowledge and noticing of antismoking mass media messages and health warning labels on cigarette packages. We constructed a four-point knowledge scale and performed multivariate regression analyses.
Results Median country values for the proportion of adults who believed smoking causes a specific illness were 95.9% for lung cancer, 82.5% for heart attack and 74.0% for stroke. Knowledge scores ranged from 2.1 to 3.8. In multivariate regressions, adults scored significantly higher on the knowledge scale if they noticed antismoking media messages (22 countries) or health warning labels (17 countries). Significantly higher knowledge scores occurred in all 9 countries with pictorial health warning labels compared with only 8 out of 13 countries with text-only warning labels.
Conclusions Antismoking media messages appear effective for warning the public about the harms from tobacco use in all 22 countries, while warning labels are effective in the majority of these countries. Our findings suggest opportunities to motivate smoking cessation globally.
- Low/Middle income country
- Packaging and Labelling
- Public opinion
- Smoking Caused Disease
Statistics from Altmetric.com
The harms of active and passive exposure to tobacco smoke have been well established. With the release of the 2014 US Surgeon General report, 50 years after the first Surgeon General report on the health effects of smoking, the list of diseases caused by smoking has expanded even further to include diabetes and liver cancer among others.1 The report has additionally added stroke to the list of diseases presenting in non-smokers resulting from exposure to secondhand smoke. Despite this growing evidence on the harms of tobacco use to health, tobacco users continue to underestimate these risks to themselves and others.2–4 Globally, an estimated 31.1% of men and 6.2% of women currently smoke tobacco, with the burden placed mainly on low-income and middle-income countries.5 In fact, smoking prevalence from 1980 to 2012 increased in 25 out of 137 developing countries and 12 out of 50 developed countries.
Tobacco use is responsible for an estimated 6.4 million deaths annually and is expected to increase to 10 million by 2030.6 ,7 Despite this alarming figure, tobacco use is the single most preventable cause of death worldwide. Knowledge about the harms of smoking tobacco is strongly associated with decreases in smoking initiation and increases in smoking cessation.8 Additionally, tobacco control strategies to educate the public on the harms of smoking, such as antitobacco mass media messages and health warning labels on cigarette packages, have been shown to be effective in promoting behavioural change.9–13 Tobacco control messaging through mass media is particularly important in partially counterbalancing the tens of billions of US dollars that have been spent in tobacco advertising, promotion and sponsorship.14 ,15 Identifying factors associated with tobacco-related knowledge is also important for the development of tobacco control measures, in accordance with the WHO Framework Convention on Tobacco Control (WHO FCTC) and MPOWER strategies,i which include warnings about the dangers of tobacco use through the aforementioned approaches.9
Most of the studies on knowledge about the harms of tobacco use in the general population have been conducted in high-income countries,16 but this burden is increasing in low-income and middle-income countries. Some studies have examined tobacco-related knowledge for individual low-income and middle-income countries or for multiple countries within a geopolitical region, but studies with comparable data on tobacco-related knowledge across countries globally are not available.17–23 A global cross-country analysis allows for cross-country and regional comparisons, and can assist in a more targeted implementation of tobacco control strategies. The WHO FCTC, which was adopted by the World Health Assembly in 2003 and is currently ratified by 180 countries, provides a foundation for evidence-based strategies on tobacco control for countries through the MPOWER guidelines.24
More than 2.3 billion people living in 92 countries are covered by at least one tobacco control measure outlined by the WHO FCTC, with 54% of the world's population covered by effective antitobacco mass media messages and 14% covered by effective health warning labels.14
The Global Adult Tobacco Survey (GATS), a component of the Global Tobacco Surveillance System, is the gold standard for systematic cross-sectional monitoring of adult tobacco use and the tracking of key tobacco control indicators. Data from 22 countries currently account for 61.3% of the world's population of people aged 15 years or older.25 Through a standardised cross-sectional household survey, we sought to determine the levels of knowledge about the harms of smoking tobacco and the factors associated with differences in this knowledge in 22 countries representing low-income, middle-income and high-income countries.
Study design and participants
Data were gathered from all 22 countries that conducted GATS between 2008 and 2013, including Argentina, Bangladesh, Brazil, China, Egypt, Greece, India, Indonesia, Malaysia, Mexico, Nigeria, Panama, Philippines, Poland, Qatar, Romania, Russia, Thailand, Turkey, Ukraine, Uruguay and Vietnam. GATS, which has previously been described,26–28 is a nationally representative household survey of non-institutionalised persons aged ≥15 years. GATS uses a standardised questionnaire, sample design, data collection method and analysis protocol that allows for comparability across countries. A multistage geographically clustered sample design was used to produce nationally representative data for each country. Information on demographics, tobacco use behaviours, noticing of antitobacco mass media messages and health warning labels on cigarette packages, and knowledge about the harms of tobacco use was gathered from a combined total of 385 053 respondents from 22 countries. Additional details of individual country methodologies are provided in individual country reports (available online).29
Definition of variables
Knowledge about tobacco-related diseases was obtained from three questions asking respondents whether they believed smoking tobacco causes lung cancer, heart attack or stroke. They were also asked if they believed breathing other people's smoke causes serious illness in non-smokers. Response options included ‘Yes’, ‘No’, ‘Don't Know’ and ‘Refused’. Responses of ‘Yes’ were coded as 1, and ‘No’ and ‘Don't Know’ were coded as 0. A respondent who refused to answer a knowledge question was excluded from the analysis of that particular question. The proportion of the samples that refused to answer these questions was less than 0.2%. We developed a knowledge scale from the number of ‘Yes’ responses to each question, with one point assigned for each correct answer and total scores ranging from 0 (least knowledgeable) to 4 (most knowledgeable). Although individual country surveys contained additional questions on knowledge about tobacco-related diseases, we selected the four questions present in all 22 country surveys for the knowledge scale, allowing for optimal cross-country comparisons.
Selected independent variables included demographics and the noticing of antismoking mass media messages and health warning labels on cigarette packages. Demographic variables included sex, age, place of residence (urban, rural), educational attainment (low, high) and smoking status (current, former, never). We defined educational attainment as low if a respondent did not receive formal education or did not complete primary education, and as high if a respondent completed primary education or higher. We defined current smokers as respondents who currently smoke tobacco on a daily or less than daily basis, former smokers as respondents who currently do not smoke but smoked daily or less than daily in the past, and never smokers as respondents who currently do not smoke and never smoked in the past. Respondents were asked if they noticed any information about the dangers of smoking cigarettes or anything else that encouraged quitting smoking in the past 30 days. They were asked if they had seen this information in various forms of media, including newspapers/magazines, television, radio, billboards or additional media forms. Respondents were separately asked if they noticed any health warnings on cigarette packages in the past 30 days, regardless of smoking status. Response options for these questions included ‘Yes’, ‘No’, ‘Not applicable’ or ‘Refused’. ‘Not applicable’ and ‘Refused’ responses were excluded from the analysis.
Owing to the complex survey design, data were analysed using SPSS V.18.0 for complex samples. We weighted data appropriately to ensure accurate representation of the national population and analysed data from each country separately. For each country, we performed a multiple linear regression analysis for complex samples with the knowledge scale as the dependent variable. All independent variables for the multiple linear regressions were categorical, except for age. Significance was determined at p<0.05.
Demographic characteristics for the 22 countries are shown in table 1. More than one-fifth of the adults in eight countries had low educational attainment (Bangladesh, Egypt, India, Indonesia, Nigeria, Philippines, Thailand and Vietnam). Ten countries had a prevalence of current smokers of greater than 25%, with Russia being the highest at 39.1%.
Table 1 shows the proportion of adults who noticed antismoking mass media messages and health warning labels on cigarette packages. More than 80% of adults noticed any antismoking information through mass media messages in eight countries (Malaysia, Mexico, Philippines, Romania, Thailand, Turkey, Uruguay and Vietnam), while less than 50% of adults noticed this information through mass media messages in three countries (Bangladesh, Greece and Nigeria). More than 80% of adults noticed health warning labels on cigarette packages in four countries (Turkey, Philippines, Thailand and Vietnam), while less than 50% noticed health warning labels on cigarette packages in one country (Nigeria).
Table 2 and figure 1 show beliefs about the harms of smoking. The proportion of adults who believed smoking causes lung cancer ranged from 77.5% (China) to 98.6% (Argentina), causes heart attack ranged from 38.7% (China) to 95.0% (Egypt), and causes stroke ranged from 27.2% (China) to 89.2% (Romania). Median values for these beliefs across all countries were 95.9% for lung cancer, 82.5% for heart attack and 74.0% for stroke. Belief that tobacco smoke causes serious illness in non-tobacco users ranged from 64.3% (China) to 96.3% (Egypt), with a median value of 89.5% across all countries. Table 3 and figure 2 illustrate the aggregated knowledge score, which ranged from 2.1 (China) to 3.8 (Egypt), and had a median score of 3.4 out of 4. Countries in the lowest knowledge score quartile included China, India, Indonesia, Nigeria and Russia. Countries in the highest knowledge score quartile included Argentina, Egypt, Qatar, Romania, Turkey and Uruguay.
Table 3 shows results from the multiple linear regression analyses. Adults who noticed any form of anticigarette smoking information in the media scored significantly higher on the knowledge scale (adjusting for confounders) in all 22 countries. Countries with the highest coefficients included Malaysia (0.81) and Vietnam (0.80), while countries with the lowest coefficients included Egypt (0.11) and Panama (0.13). Adults who noticed warning labels on cigarette packages had significantly higher knowledge scores (adjusting for confounders) in 17 countries, with coefficients ranging from 0.07 in Uruguay to 0.39 in Vietnam.
The relationship of demographic factors to knowledge varied widely among countries. For example, men's and women's knowledge scores were significantly different in only eight countries and the direction of the relationship varied, ranging from −0.10 for Argentinian men to 0.24 for Bangladeshi men. Urban residents had significantly higher knowledge scores in 13 countries, with the maximum effect of 0.53 in China. Ten countries demonstrated a significant effect of age as a continuous variable on knowledge, although coefficients were very small. In 17 countries, low educational attainment was associated with significantly lower knowledge scores, ranging from −0.12 in Argentina to −0.97 in Ukraine. Finally, smoking status had a significant effect in 19 countries for never smokers and former smokers, with both having higher knowledge scores than current smokers.
We report on knowledge about the harms of smoking across 22 countries through a standardised, nationally representative household survey. The majority of countries displayed a step-wise trend in knowledge about tobacco-related diseases (figure 1), where the highest proportion of adults believed smoking causes lung cancer (77.5–98.6%), a lower proportion believed smoking causes heart attack (38.7–95.0%), and the lowest proportion believed smoking causes stroke (27.2–89.2%). This trend was similarly seen by Roberts et al23 in their analysis of knowledge about the harms of smoking in nine countries of the former Soviet Union using a different nationally representative household survey. Comparable proportion of adults who believed smoking causes a specific illness in their analysis ranged from 77.8% to 93.2% for lung cancer, 51.4% to 79.1% for heart attack and 11.4% to 52.5% for stroke. Knowledge about the harms of smoking in high-income countries has been shown to follow a similar step-wise pattern;16 however, levels of knowledge are typically greater for these countries. This pattern suggests that current strategies that warn about the harms of smoking have been effective in conveying the relationship between smoking and lung cancer, but less so for cardiovascular diseases (such as heart attack and stroke), which account for 29% of tobacco-related deaths globally.30
Our analysis revealed that the lowest levels of knowledge about the harms of smoking were found in Asia, where China, India and Indonesia were in the lowest knowledge quartile among the 22 countries (figure 2). These data are noteworthy since nearly half of the world's smokers reside in these three countries alone.5 For China, the International Tobacco Control (ITC) Project reported higher levels on knowledge about the harms of smoking for lung cancer (80.1% of smokers, 98.4% of non-smokers), coronary heart disease (52.1% of smokers, 71.5% of non-smokers), and stroke (26.8% of smokers, 52.1% of non-smokers).31 However, the ITC sampled populations in six Chinese cities, but not in rural areas that typically have lower levels of knowledge about the harms of smoking. The ITC data on tobacco knowledge in India varied widely among the four states surveyed (75.9–96.4% for lung cancer, 49.4–86.6% for coronary heart disease and 23.0–80.1% for stroke), and sampling was limited to one city and its surrounding rural districts in each state.32 In contrast, the India GATS was conducted in all 29 states and 2 Union Territories at the time and data were weighted to ensure national representation.
We found that adults who noticed antismoking information through mass media had significantly higher knowledge about the harms of smoking in all 22 countries (table 3). These findings are consistent with cross-sectional and longitudinal studies on media campaigns in low-income and middle-income countries.33–35 Data from Durkin et al36 suggest that the use of graphic imagery in television advertisements to communicate serious harms of smoking are effective in low-income and middle-income countries due to universally high levels of understanding and engagement of viewers. Furthermore, Zawahir et al37 found that adolescents in Malaysia and Thailand who noticed antismoking media messages had significantly greater knowledge about the health effects of smoking and the perceived health risks of smoking. Mass media messages that use graphic imagery to communicate the harms of tobacco can be particularly effective in low-income and middle-income countries, where educational attainment and literacy rates tend to be lower than that in high-income countries.38 ,39 In addition to increasing knowledge about the harms of smoking, effective mass media messaging can also influence emotions and feelings towards smoking, particularly if the content elicits negative emotions.13 ,15
Adults who noticed health warning labels on cigarette packaging were more knowledgeable about the harms of smoking in 17 countries (table 3). These findings are consistent with a number of cross-sectional and longitudinal studies that have shown knowledge about the harms of smoking increase with the introduction of health warning labels in both high-income countries as well as low-income and middle-income countries.34 ,40 ,41 All 9 countries (Brazil, Egypt, India, Malaysia, Panama, Qatar, Romania, Thailand and Uruguay) with pictorial health warning labels on cigarette packages at the time of the GATS had significantly higher knowledge in the adults who noticed warning labels compared with those who did not, while only 8 of the 13 countries with text-only health warning labels at the time had significantly higher knowledge in adults who noticed warning labels. Studies from high-income countries have demonstrated that pictorial health warning labels promote stronger beliefs about the health risks of smoking.42 Fong et al43 found that pictorial warning labels, compared with text-only labels, were more informative about the dangers of smoking, and more effective in promoting cessation and preventing smoking initiation in China, which currently has text-only health warning labels. Studies from Latin America have shown that pictorial health warning labels with bodily depictions of smoking-related diseases had a greater impact on participants’ thoughts about the health risks of smoking and quitting compared with warning labels with abstract imagery (eg, a ‘ticking time bomb’) or only text.44 Similar to antismoking media messages, graphic warning labels can allow populations with low educational attainment to be informed about the harms of smoking.45
The impact of antismoking mass media messages and health warning labels on smoking-related knowledge should be taken in context with other antitobacco measures within a country. For example, in Uruguay, a small yet significant difference in knowledge (0.07 points) between those who noticed health warning labels and those who did not notice them may be explained by a series of comprehensive antitobacco control measures launched by the country in 2005 (4 years prior to the GATS). In addition to mandating graphic health warning labels on cigarette packages, other measures included taxes on tobacco products, indoor smoking bans and the bans on tobacco advertisements, promotions and sponsorships.46 These other tobacco measures may have an indirect effect on exposure to knowledge about the harms of smoking by impacting the exposure to health warning labels.
We note some limitations of this study. First, data were obtained through self-reports and thus, are subject to information bias due to varying social acceptability of smoking in each country. For example, if a country enacts a more aggressive tobacco control policy, people in that county may be less inclined to admit to smoking or feel more pressured to respond that smoking causes a particular disease even if they had not learnt that information. These implications mean that countries with more aggressive tobacco control policies could have falsely elevated knowledge scores and these may not be related to the presence of antismoking messaging or warning labels. Second, recall bias may occur such that respondents who use tobacco or are more often surrounded by tobacco may be more likely to notice tobacco-related media and are more likely to be exposed to health warning labels on tobacco products that they purchase. This bias would mean that noticing antismoking messaging and warning labels would be more highly represented among smokers compared with non-smokers and could incorrectly show that antismoking messaging and warning labels are less effective among non-smokers. Third, self-reported responses to noticing antismoking messaging and warning labels may not be a strong proxy for measuring exposure to antismoking messaging and warning labels. It is likely that a person may have been exposed to antismoking messaging but may not have remembered this exposure at the time of the survey. On the other hand, this measure includes a component of effectiveness since a more effective message is more likely to be remembered by the respondent when compared with an ineffective message. Fourth, as a cross-sectional study, no causal relationships among variables could be determined. Countries with higher levels of knowledge about the harms of smoking may have higher levels of noticing antismoking messaging. This may be due to higher exposure of antismoking information causing increases in knowledge or conversely, it may mean that respondents who had previously learnt about the harms of smoking may be more likely to notice antismoking messaging. Fifth, this study was not able to determine the specific smoking harms included in antismoking messaging, which may vary among countries and could influence differences in knowledge about specific smoking harms. Thus, we are unable to determine if a focus on the links of particular smoking-related diseases, such as lung cancer, to cigarette smoking in antismoking messaging and warning labels is associated with differences in knowledge about the harms of smoking. Lastly, each of the four questions in our knowledge score was assumed to have equal importance, which could lead to measurement error. For example, knowing that smoking causes lung cancer may have more impact in changing smoking behaviour compared with knowing that secondhand smoke is generically dangerous. This knowledge about lung cancer may be more valuable and could have been weighted more heavily in the composite knowledge score.
Hard-hitting antitobacco advertisements and graphic warning labels on cigarette packages reduce smoking initiation and increase cessation. Our findings showed that noticing antismoking mass media messages and health warning labels on cigarette packages are significantly associated with increased knowledge about the harms of smoking in low-income and middle-income countries. Our study adds to the literature on the effectiveness of graphic health warnings in increasing people's awareness about the harms of smoking. The use of warnings with pictures or pictograms of the health consequences of tobacco use is particularly important in countries with large proportions of the population who have low educational attainment. By implementing the WHO FCTC guidelines for antismoking media campaigns and health warning labels, countries can accelerate the declines in tobacco use. Article 11 of the FCTC requires health warnings on cigarette packages that should cover at least 50% of both the front and back of the package, and should include appropriate characteristics, such as pictures or pictograms. As of 2012, only 30 countries have implemented all of the FCTC standards for health warning labels, though most of them are low-income and middle-income countries.14 Thirty-six countries have had a national antitobacco media campaign conducted with appropriate characteristics, including messages airing on television and/or radio. Messaging through television and/or radio tends to have the greatest population reach in nearly all countries around the world and have more impact due to their audio-visual nature. The WHO identifies mass media campaigns and graphic warning labels as ‘best buys’ for population-wide efforts to decrease the burden of non-communicable diseases globally because they are highly cost-effective, feasible and culturally appropriate to implement.47 Our findings suggest that opportunities exist to enhance antismoking media messages and health warning labels for improving knowledge about the harms of smoking that could promote cessation and prevent initiation, particularly in low-income and middle-income countries, which have a high tobacco use burden.
What this paper adds
Knowledge about the harms of smoking tobacco is strongly associated with decreases in smoking initiation and increases in smoking cessation.
Tobacco control strategies to educate the public on the harms of smoking, such as antitobacco mass media messages and health warning labels on cigarette packages, have been shown to be effective in promoting behavioural change.
Most of the studies on knowledge about the harms of tobacco use in the general population have been conducted in high-income countries, but the burden is increasing in low-income and middle-income countries.
Some studies have examined tobacco-related knowledge for individual low-income and middle-income countries or for multiple countries within a geopolitical region, but studies with comparable data on tobacco-related knowledge across countries globally are not available.
Our findings show that antismoking mass media messages and health warning labels on cigarette packages are significantly associated with increased knowledge about the harms of smoking in low-income and middle-income countries.
The authors thank the thousands of field workers for their contributions as well as the thousands of respondents for their cooperation, without whom this work would not have been realised.
Contributors JJC wrote the statistical analysis plan, analysed the data, drafted and revised the paper. LA wrote the statistical analysis plan, cleaned and analysed the data, and drafted and revised the paper. SA and KP helped coordinate the survey implementation in all countries, analysed the data, and drafted and revised the paper. TM facilitated the survey implementation in all countries and revised the paper. All members of the GATS Regional and Country Authors facilitated the survey design, data collection and data analysis in their respective countries, as well as revising of the paper. Roberta D Caixeta for countries in the Region of the Americas, Heba Fouad for countries in the Eastern Mediterranean Region, Rula N Khoury for countries in the European Region, Nivo Ramanandraibe for countries in the African Region, James Rarick for countries in the Western Pacific Region, Dhirendra N Sinha for countries in the South East Asian Region, Jonatan Konfino for Argentina, M Mostafa Zaman for Bangladesh, Deborah Carvalho Malta for Brazil, Qi Shi for China, Sahar Latif Labib for Egypt, Anastasia Barbouni and Eleni Antoniadou for Greece, Amal Pushp for India, Soewarta Kosen for Indonesia, Helen Tee Guat Hiong and Sukhvinder Singh for Malaysia, Luz Miriam Reynales-Shigematsu for Mexico, Isiaka Olarewaju for Nigeria, Reina Roa for Panama, Agnes Segarra for the Philippines, Dorota Kaleta for Poland, Amani Elkhatim for Qatar, Amalia Canton for Romania, Maria Shevireva for the Russian Federation, Sarunya Benjakul for Thailand, Bekir Keskinkılıç for Turkey, Konstantin Krasovsky for Ukraine, Ana Lorenzo for Uruguay, and Phan Thi Hai for Vietnam.
Funding This work was supported by the Bloomberg Initiative to Reduce Tobacco Use, a programme of Bloomberg Philanthropies; Bill and Melinda Gates Foundation for GATS implementation in China; the governments of Brazil, Greece, India, Malaysia, Panama, Qatar, Thailand and Turkey for GATS implementation in their respective countries, and External Medical Affairs of Pfizer Inc. to the CDC Foundation for funding The CDC Experience Applied Epidemiology Fellowship.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Data for this analysis can be accessed from the Global Tobacco Surveillance System Data (GTSSData) website at http://www.cdc.gov/tobacco/global/
↵i MPOWER strategies include monitoring tobacco use and prevention policies, protecting people from tobacco smoke, offering help to quit tobacco use, warning about the dangers of tobacco, enforcing bans on tobacco advertising, promotion and sponsorship, and raising taxes on tobacco.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.