Article Text
PDF
Abstract
Objective Smokeless tobacco consumption in India is a significant source of morbidity and mortality. In order to educate smokeless tobacco users about the health harms of smokeless tobacco and to denormalise tobacco usage and encourage quitting, a national television and radio mass media campaign targeted at smokeless tobacco users was aired for 6 weeks during November and December 2009.
Methods The campaign was evaluated with a nationally representative household survey of smokeless tobacco users (n=2898). The effect of campaign awareness was assessed with logistic regression analysis.
Results The campaign affected smokeless tobacco users as intended: 63% of smokeless-only users and 72% of dual users (ie, those who consumed both smoking and smokeless forms) recalled the campaign advertisement, primarily through television delivery. The vast majority (over 70%) of those aware of the campaign said that it made them stop and think, was relevant to their lives and provided new information. 75% of smokeless-only users and 77% of dual users said that it made them feel concerned about their habit. Campaign awareness was associated with better knowledge, more negative attitudes towards smokeless tobacco and greater cessation-oriented intentions and behaviours among smokeless tobacco users.
Conclusions Social marketing campaigns that utilise mass media are feasible and efficacious interventions for tobacco control in India. Implications for future mass media tobacco control programming in India are discussed.
- Smokeless tobacco
- smoking
- evaluation
- tobacco control
- social marketing
- mass media
- communications
- advertising and promotion
- public policy
- smokeless tobacco products
Statistics from Altmetric.com
- Smokeless tobacco
- smoking
- evaluation
- tobacco control
- social marketing
- mass media
- communications
- advertising and promotion
- public policy
- smokeless tobacco products
Introduction
A global health epidemic, tobacco use prevalence is of significant concern in a populous nation like India where it is projected that tobacco-related deaths may exceed 1.5 million annually by 2020.1 While cigarette smoking is primarily responsible for the burden of disease in most countries, in India smokeless tobacco usage is more prevalent than smoking, particularly among women. According to the latest National Family Health Survey,2 36.5% of men and 8.4% of women chew tobacco; 32.7% men and 1.4% women smoke; and 57% men and 10.8% women use either form. Smokeless tobacco usage is also of growing concern among youth with an estimated 8% of children in 8th–10th grades currently using smokeless tobacco.3 Smokeless tobacco consumption has been linked to India's high oral cancer rates, poor reproductive outcomes and high rates of cardiovascular diseases.4–7 Reduction of tobacco usage, including the use of mass media public education campaigns, has thus emerged as a critical public health priority.
Tobacco control social marketing campaigns
Tobacco control social marketing campaigns, which primarily utilise mass media to educate and inform the public, have been effective in denormalising smoking and in reducing tobacco prevalence in high-income countries.8–10 Similar evidence from low and middle-income countries is limited but emerging.10–12 The relative lack of such campaigns in developing countries until recently may have largely been due to concerns about the ability of mass media to reach the vast majority of the population that lives in rural areas.13 14 However, economic development in countries such as India and China has also led to a rapidly changing media landscape providing social marketers with greater opportunities for achieving effective dissemination of tobacco control and other health campaign messages to these previously difficult to reach populations.
India has a diverse and dispersed population15 16 and a complex media environment.17–19 Although rural areas were isolated until recently from most external media influences,20 the government's commitment to electrification and the provision of televisions in every village,21 coupled with changes in media content, have restructured the media landscape. Given the potential influence of this development on the vast majority of India's people, mass media, and in particular television, may be a powerful tool to set agendas and motivate behaviour change in relation to tobacco usage. Indeed, a recent survey of approximately 200 000 Indians found that higher rates of viewing movies and television were associated with higher prevalence,22 suggesting that the likely pro-tobacco imagery in visual media may impact tobacco usage. It is no surprise then that the Indian tobacco industry has successfully leveraged mass media and alternative promotional strategies to promote smokeless tobacco products.23
A solid body of evidence in the tobacco control literature indicates that an effective strategy for countering the effects of tobacco promotion is to use anti-tobacco health messaging that is graphic and negatively emotive, eliciting fear, anger, disgust and concern.8 10 Such graphic messages are not only effective in engaging viewers and prompting thoughts of quitting, but they are also an efficient strategy for maximising campaign budgets. Graphic emotional advertisements require less media expenditure because they have a higher immediate impact in terms of recall,24 and several studies demonstrate sustained use of negatively emotive campaigns using a combination of graphic imagery and personal stories can reduce smoking prevalence at the population level.25–27
Government of India smokeless tobacco campaign
Guided by this evidence on effective anti-tobacco messaging, the Government of India embarked in 2009 on what was, to our knowledge, the first mass media campaign in India on the harmful effects of smokeless tobacco use. The Government of India, in collaboration with World Lung Foundation and other stakeholders, pretested and developed a 30-second documentary style public service announcement (PSA). It features an oral cancer surgeon from a tertiary care cancer hospital, the Tata Memorial Hospital, Mumbai, with one of the busiest cancer wards in the world, as he describes and presents the serious illnesses and disfigurement of his patients, caused by cancers resulting from use of smokeless tobacco.i The smokeless campaign had a media spend of approximately US$1 million and was aired for 6 weeks from November to December 2009. There were a total of 3136 television exposures with 81% occurring during ‘primetime’ (56% on weekday evenings, 20% on weekend evenings and 5% on weekend mornings) and 19% during non-primetime periods (11% weekday mornings and 8% weekend afternoons). A third of the campaign spend was on state-owned national and regional channels and the rest was allocated to privately owned, national cable and satellite channels. The campaign was supplemented by 11 460 planned radio spot exposures for the duration of the campaign. Targeted at smokeless tobacco users aged between 16 and 50 years in urban and rural areas, the campaign objectives were to raise awareness, increase knowledge and build risk perceptions of the health consequences of smokeless tobacco use.
Methods
To evaluate the campaign's impact, a nationally representative household survey of tobacco users was conducted by ORG Centre for Social Research soon after the campaign's conclusion, from 20 December 2009 to 10 January 2010. Owing to time constraints, it was not feasible to conduct a pre-campaign survey.
Sampling method and sample size
A three-stage stratified random sampling procedure was used with the relevant sampling frame from the 2001 census. ‘Wards’ for urban areas and ‘revenue villages’ for rural areas formed the primary sampling units. In the first stage of sampling, districts—the third level administrative units in India after states—were selected. Wards/revenue villages were selected in the second stage. Households and respondents within households were selected in the third stage. A modified cluster sampling approach was followed to obtain the required number of households. A Kish grid was used to select one from among the multiple eligible respondents in a household. To enable regional comparisons, the national sample was divided equally across the north, south, east and west, and the final data were weighted to arrive at regional and national estimates.
Respondent eligibility criteria included: (1) current smoking or smokeless tobacco usage; (2) aged 16–50 years; and (3) access to mass media, which was defined as watching television or listening to the radio at home or elsewhere in the community during the 2 months prior to the survey (November–December 2009).
Of the 15 927 households sampled across India, 14 906 were identified as having household members between 16 and 50 years of age. Screening interviews were successfully completed with 92% or 13 686 age-qualifying respondents. A third of the screened respondents (n=4509) were identified as current tobacco users, and comprised 2714 (or 60% of current tobacco users) who only use smokeless forms of tobacco (referred to as ‘smokeless-only users’); 779 (17%) who only smoke (‘smokers-only’); and 1016 (23%) who use both smokeless and smoking tobacco forms (‘dual users’). The full survey was administered to current tobacco users who had used TV or radio in the previous 2 months, including 2108 smokeless-only users, 622 smokers-only and 790 dual users. In this paper we present data for the smokeless-only and dual user groups, the primary target audience of the smokeless campaign.
Measures
The following were the key indicators measured in the questionnaire (but this does not reflect their order in the questionnaire). Tobacco usage was measured with a series of 10 questions on past and current frequency of smoking or smokeless tobacco use (daily, less than daily or not at all) and daily/weekly rates of consumption.
Campaign awareness was determined through two sets of questions. First, respondents were asked to freely recall and describe all the advertisements about smokeless tobacco harms that they had encountered in the previous 2 months; where they came across these advertisements; and what they recalled as the advertisements' messages. All those who recalled the campaign PSA through either television or radio and correctly described its key messages were defined as having ‘category-cued’ recall, which is an unprompted measure of an advertisement's salience. Later in the questionnaire, respondents were shown images from the television spot and read a few sentences from the radio script and asked if they recognised the PSA in either form. All those who recognised the PSA with these prompts and were able to accurately recall its messages were defined as having ‘prompted recognition’. Campaign awareness was defined to collectively include all those with category-cued recall and recognition of the campaign PSA; ‘unaware’ respondents included all others.
Campaign-aware respondents then appraised the campaign PSA on four proximal indicators that have been found to be indicative of a campaign's effectiveness: specifically, they indicated how strongly they agreed or disagreed that the PSA made them ‘stop and think’, ‘was relevant to (their) lives’, ‘provided new information’ and ‘made (them) feel concerned about their use of smokeless tobacco’. They also indicated whether after viewing the PSA they ‘discussed it with others’ or tried ‘to persuade others to quit’.
Knowledge and attitudes towards smokeless tobacco were measured, the former using true-false rating scales and the latter using five-point agreement scales. To assess campaign effect, four of the seven knowledge and attitudinal items directly pertained to the messages communicated in the PSA (eg, ‘smokeless tobacco causes mouth cancer’; ‘It would improve my health if I quit using smokeless tobacco’). The remaining items represented knowledge and attitudes that were not directly mentioned in the PSA (eg, ‘Smokeless tobacco causes the birth of low birthweight babies among pregnant women’; ‘The people important to me believe that I should not use smokeless tobacco’).
Interpersonal communications about quitting in the 2 months prior to the survey were measured with two items by asking respondents, in the past 2 months ‘how often (they) had discussed smokeless tobacco and health at home’ (response options: never/once or twice/more often), and whether or not ‘someone in (their) house had tried to get them to quit’.
Cessation-related intentions and behaviours were measured with four items by asking respondents, ‘how often (they) had seriously considered quitting in the past 2 months’ (never/once or twice/more often), how strongly they agreed (or disagreed) that ‘they had the skills and ability to quit smokeless tobacco’, whether or not ‘they had tried to quit in the previous 2 months’ and whether ‘they consume more, less or the same amount of tobacco now than they did 2 months ago’.
Data analysis
Data were weighted to adjust for regional representation, sampling design and response rates, and then analysed using SPSS 18.0. Comparisons between proportions were conducted using χ2 tests. Logistic regressions to assess campaign impact were performed separately for the smokeless-only and dual user groups. Campaign awareness was regressed on dichotomised knowledge, attitudes and cessation-related items. Covariates included sex, age, socioeconomic status (SES),ii settlement type (rural vs urban), region, weekly television and radio usage and intentions to quit tobacco in the next 12 months. Interpersonal communication about smokeless tobacco harms in the previous year, defined as education by anyone in the community (eg, doctor, health worker), was also included as a covariate.
Results
Sample characteristics
Demographic characteristics of the final sample of smokeless-only and dual users are presented in table 1. As expected, the majority of smokeless-only users (75%) and dual users (82%) resided in rural areas. The majority was also men (particularly among dual users) and from the lowest SES category. Both groups of smokeless users were more likely to have used TV than radio in the previous 2 months.
Demographic characteristics of the final study sample
Campaign awareness and message appraisal
Sixty-five per cent of all smokeless users were aware of the campaign PSA. Among smokeless-only users, 58% had category-cued recall and an additional 5% accurately recognised the PSA when prompted with audio/visual cues. Among dual users, 69% had category-cued recall and an additional 3% recognised the PSA with the audio/visual cues. Among all the campaign-aware respondents, 62% recalled the campaign on television only, 16% on radio only and 21% on both.
As indicated in table 2, campaign-aware smokeless-only users were more likely than unaware users to reside in rural areas and be of a higher SES category. They were also more likely to reside in the north and west of the country. Campaign-aware dual users were somewhat less likely to be male and more likely to reside in the northern and western parts of India.
Campaign awareness among smokeless-only and dual users in the final study sample
The campaign message was appraised as intended: the vast majority of campaign-aware smokeless-only and dual users agreed that the PSA made them stop and think, was relevant to their lives and provided new information (see table 3). Additionally, 75% of smokeless-only and 77% of dual users said that it made them feel concerned about their smokeless tobacco usage. Campaign awareness led to greater interpersonal communication among dual users than smokeless-only users: 50% of campaign-aware dual users compared to 43% of campaign-aware smokeless-only users discussed the PSA with others; 41% of aware dual users compared to 26% of aware smokeless-only users tried to persuade others to quit using smokeless tobacco.
Appraisal of campaign messages and actions taken as a result of campaign awareness by tobacco user types
Impact of campaign awareness on knowledge and attitudes towards tobacco
Among smokeless-only users, there was a significant relation between campaign awareness and campaign-relevant knowledge and attitudes towards smokeless tobacco (table 4). Campaign-aware smokeless-only users had greater knowledge that smokeless tobacco causes mouth and throat cancer; they were more likely to agree that it causes serious illnesses and that quitting would improve their health. There was no such significant association on non-campaign-relevant knowledge and attitudinal items.
Knowledge and attitudes about tobacco usage among smokeless-only users and dual users
Although the association was more limited, among dual users too, campaign awareness was significantly associated with campaign-related knowledge and attitudes (table 4). Campaign-aware dual users were more likely than unaware users to know that smokeless tobacco causes throat cancer and to agree that quitting would improve their health. There was no such association between campaign awareness and non-campaign relevant knowledge/attitudes.
Cessation and intentions to quit
Among smokeless-only users, campaign awareness was significantly associated with greater interpersonal communication about smokeless tobacco harms (table 5): campaign-aware smokeless-only users were more likely to have discussed smokeless tobacco and health at home in the previous 2 months; and were more likely to say that somebody in their house had been trying to get them to quit in the previous 2 months. Campaign awareness was also associated with cessation-oriented behaviours: aware smokeless-only users were significantly more likely than unaware users to seriously consider quitting smokeless tobacco; to have stronger beliefs in their own ability to quit; and to have tried to quit in the previous 2 months.
Tobacco cessation intentions and behaviours among smokeless-only users and dual users
Among dual users, the association between campaign awareness and cessation behaviours was more limited. Campaign-aware dual users were significantly more likely to have experienced familial pressure to quit and were more likely to believe in their ability to quit smokeless tobacco (table 5). However, campaign awareness was not associated with greater attempts to quit, and campaign-unaware dual users were more likely than campaign-aware users to say that they use less tobacco now than they did 2 months ago.
Discussion
The study findings indicate that the smokeless campaign performed as intended, particularly among smokeless-only users. The campaign PSA was recalled by over two-thirds of all smokeless tobacco users. Despite their rural residence and low SES background, smokeless users recalled the campaign more often through television than radio, bolstering the idea that mass media, and in particular television, can be an effective tool for reaching more marginalised populations and facilitating behaviour change.
The campaign PSA was memorable and resonated with the intended audience. The vast majority of respondents who recalled the PSA did so with minimal prompting—the visual and audio memory aids only helped an additional 3–5% recall the PSA, indicating the powerful effect of its graphic imagery.24 The PSA was rated as persuasive and causing concern about the dangers of smokeless tobacco, and significant numbers reported discussing it with others.
There was a significant association between campaign awareness and campaign-relevant knowledge, attitudes and behaviours among smokeless-only users: Campaign-aware users had better knowledge, more negative attitudes towards smokeless tobacco and greater engagement in cessation intentions and behaviours than unaware users. The fact that this association was present only on items that were specifically targeted by the campaign and not in areas unrelated to the campaign message, further bolsters the idea that campaign exposure was probably responsible for the change in knowledge, attitudes and behaviours of smokeless-only users. Among dual users, there was a similar pattern of results in that campaign awareness was associated with better knowledge and more anti-tobacco attitudes only on items that were related to the campaign's messages. However, the effect of campaign awareness was weaker among dual users with the association occurring in fewer outcomes.
This comparison between the smokeless-only and dual users is an important one. Our data suggest that while there are many similarities between these groups, there are also some meaningful differences that need to be considered in future campaigns. Both groups were similar in their appraisal of the campaign, rating it as thought provoking, personally relevant, informative and worrying. However, while dual users were more likely than smokeless-only users to talk about the campaign and to attempt to persuade others to quit, suggesting the significant impression that the campaign made on them, they were less likely to demonstrate shifts in knowledge, attitudes or behaviours as a consequence of campaign awareness. This finding suggests that dual users' motivations regarding tobacco use may be quite distinct and possibly more complex than those of smokeless-only users. It will be important to monitor dual users for a potentially serious fallout of campaigns that focus on one form of tobacco or the other—namely, it will be important to ensure that dual users are not reacting to these messages by substituting one form of tobacco for another. Our own data suggest that this is not a concern: when probed in subsequent sections of the survey on their smoking habits in the previous 2 months, campaign-aware dual users were not more likely than unaware respondents to have increased their cigarette consumption over this time period. However, future research utilising longitudinal designs or follow-ups will be important to study the consequences of different types of tobacco control messaging on this group.
Study limitations
First, a true baseline survey could not be conducted prior to campaign launch owing to time and resource constraints. However, to make the most of the post-intervention survey, it was designed to be rigorous with large and representative samples, including hard-to-reach tobacco users from rural areas and low SES groups, in order to enable both a meaningful evaluation and serve as a baseline for the future. Second, as in any campaign evaluation, ‘selective attention’ may have been a potential confounder in this study. It is possible that tobacco users who were more concerned about tobacco harms were more attentive to, and likely to recall, the campaign. That said, even if the campaign were reaching those predisposed towards the campaign's message, the ‘low hanging fruit’ as it were, the very movement of these individuals from mere contemplation to actual attitudinal or behavioural change is an important achievement. Moreover, our finding that these changes occurred primarily on campaign-related messages (and not on indicators that were unrelated to the campaign) lends support for a campaign-specific effect. Third, particularly in the absence of baseline data, this evaluation would have been strengthened by a ‘dose-response’ analysis of the campaign effects. Unfortunately, there are currently a number of inhibitors to establishing gross rating pointsiii as a measure of campaign reach and frequency, including the rudimentary data based on gross rating points available outside of urban programming areas. As media planning capacity is developed in these programmes, it is expected that future campaign evaluations will be able to utilise such an analysis. Finally, it is possible that the regression analyses that assessed the effect of campaign awareness did not control for all potential confounders. Nevertheless, the analyses were designed to provide a stringent test of campaign effect: For example, ‘campaign awareness’ was defined conservatively to include not only recall and recognition of the PSA but also accurate recall of the campaign's messages; the regression analyses accounted for a comprehensive set of potential confounders, including future intentions to quit.
Conclusion
In conclusion, this study lends support to the feasibility and efficacy of running social marketing campaigns utilising mass media in India. Despite their largely rural and low SES composition, most smokeless tobacco consumers were effectively reached through mass media, especially television. The campaign's messages were found to be relevant and persuasive, and campaign awareness was linked to changes in knowledge, attitudes and behaviours—particularly among smokeless-only users, the campaign's primary target audience. Our data demonstrate the efficiency of using graphic emotional appeals in PSAs for tobacco control in India. Such PSAs are not only effective in eliciting the desired concern about tobacco usage, but as indicated by their memorable nature, they offer a valuable strategy for optimising media delivery and developing more efficient media budgets. Further rigorous field evaluations with no or low campaign exposure comparison groups are needed to track the impact of tobacco control mass media programming in India.
What this paper adds
Tobacco usage, particularly of smokeless tobacco, is an alarming concern in India where it is projected that tobacco-related deaths may exceed 1.5 million annually by 2020. While tobacco control mass media campaigns have been effective in changing social norms and reducing tobacco prevalence in high-income countries, there is limited evidence of their efficacy from low and middle-income countries.
This study assessed the impact of a seminal Government of India campaign that graphically portrayed the horrifying consequences of using smokeless tobacco. The evaluation survey was conducted among smokeless tobacco users across India and found high levels of campaign awareness, even among rural and low SES smokeless tobacco users. Campaign awareness was associated with better knowledge, more negative attitudes and greater cessation-related behaviours towards smokeless tobacco, suggesting that social marketing campaigns utilising mass media are an effective strategy for addressing the tobacco epidemic in India.
Acknowledgments
The authors gratefully acknowledge senior staff at the Ministry of Health and Family Welfare, Government of India, Dr Pankaj Chaturvedi (head and neck surgeon), management of Tata Memorial Hospital, India, and Dr Vinayak Prasad (formerly at GOI and currently with WHO), for their vision in supporting the Surgeon campaign; Bloomberg Philanthropies and the Bill and Melinda Gates Foundation, as part of the Bloomberg Initiative to Reduce Tobacco Use, for their generous support that enabled several aspects of the Surgeon campaign, including pretesting research, ad production and the impact evaluation study; the dedicated researchers and field staff at ORG Centre for Social Research, including Mukesh Chawla, Alpa Puri, Umesh Pathak, Sudhir Shrivastav and Vishnu Tiwari; and World Lung Foundation communications staff, particularly Alexey Kotov and Rebecca Perl, for their input and review through the campaign and publication process.
References
Footnotes
Funding The impact evaluation was supported by a grant from the Bloomberg Philanthropies to World Lung Foundation. However, Bloomberg Philanthropies was not involved in any aspect of the evaluation study or the writing of this manuscript. The authors have not been paid to write this article.
Competing interests None.
Ethics approval Prior to participation, the study was described to all participants and their formal consent to participate was sought. The questionnaire was administered only to respondents who agreed to participate in this research. As an ESOMAR member, ORG-Nielsen complies with the professional and ethical standards of ESOMAR International Code of Marketing and Social Research Practice.
Provenance and peer review Not commissioned; externally peer reviewed.
↵i The PSA may be viewed at: http://www.worldlungfoundation.org/ht/d/sp/i/7218/pid/7218.
↵ii We used the modern SES definition,28 which uses education and occupation of the chief wage earner to classify households in urban areas. The chief wage earner is defined as the person who contributes the most to the total family income. Based on various combinations of these variables, the households are classified into different socioeconomic classes from A–E (in that order, from affluent to deprived). Households in rural areas are classified using the chief wage earner's education and type of house (‘pucca’ houses are those built concrete and steel; ‘kaccha’ include thatched houses or huts; and, ‘semi pucca’ are those that are a mix of concrete and thatching). Occupation as a variable is not used in rural areas because this can be similar in most households across most Indian villages. (The definition can be viewed at: http://www.mruc.net/images/stories/Glossary_KeyIRSDefinitions.pdf.)
↵iii Gross rating points are a measure of the percentage of the target audience reached by an advertisement, by the frequency to which they are expected to see the advertisement in a given campaign.