METHODS

Data

The data were obtained between April and June 2000 in a province wide, population based survey (“Q2000”) undertaken in Ontario, Canada. The primary objective of this survey, conducted under the auspices of the Ontario Tobacco Research Unit and fully described elsewhere,³ was to obtain information on the public’s knowledge about the health effects of smoking and their attitudes regarding tobacco and tobacco control measures. Briefly, the English speaking population aged 18 years and older and who resided in households, was sampled using a two stage, probability based selection process. First, households were identified by random digit dialling, and then respondents were randomly selected from all eligible adults in the household, based on the most recent birthday method.⁴ Interviews were conducted by the Institute for Social Research at York University in Toronto, Canada, using computer assisted telephone interviewing (CATI). Up to 12 call attempts were made during the day and evening on weekdays and weekends, before a number was considered non-responding.

Interviews were completed with 1607 respondents (response rate 60%, calculated by dividing the number of completed interviews by the estimated number of eligible households). Interviews lasted about 20 minutes on average and addressed the respondent’s sociodemographic characteristics, knowledge about smoking and health, smoking history, attitudes toward the tobacco industry and tobacco products, and toward a wide range of tobacco control measures, including TID measures (the questionnaire is available from the first author). The sociodemographic characteristics of the sample were compared with those of Ontario residents based on the 1996 Census. The sample was representative with respect to age, sex, and marital status, but those with low levels of education were under-represented, a common finding in survey research.^5,6

Tobacco industry denormalisation

Eight questions directly concerned selected facets of TID. Four dealt with attitudes toward the tobacco industry and its products and four concerned TID strategies. Specifically, questions concerning the tobacco industry and its products addressed the dishonesty of the industry about the effect of smoking on health, its responsibility for the health problems smokers have because of their smoking, and for youth starting to smoke, and the dangerousness of cigarettes. Questions concerning TID strategies addressed regulation of tobacco as a hazardous product, fining of the tobacco industry by government for earnings from underage smoking, and suing of the industry by the Ontario government for health care costs caused by smoking and for taxes lost because of smuggling.

In order to identify predictors of support for TID, the responses to the eight TID questions were incorporated into a denormalisation scale. This scale had a reliability (α) coefficient of 0.70. Scale values ranged from 0 to 1, with a mean of 0.55 and a median of 0.63. Respondents whose answers to six or more of the eight items comprising the scale indicated support for denormalisation were classified as “high scorers” on the scale.

Independent variables

Support for TID, as indicated by scores on the denormalisation scale, was considered in relation to the sociodemographic characteristics of respondents (sex, age, and educational attainment) and their smoking status (smoker, non-smoker). Because former smokers and never smokers were virtually identical in their support for TID, they were considered in the analysis as one group, non-smokers.

In previous work we showed that knowledge about the health effects of active and passive smoking,^7,8 and attitudes toward the role of government in health promotion,^8,9 were independent predictors of attitudes toward tobacco control measures. To examine the relation between the TID measures and both knowledge concerning health effects and views about the responsibility of government for health promotion, two scales were created. One, a health effects knowledge scale, was based on 10 items in the questionnaire that dealt with the specific effects of active and passive smoking (smoking as a cause of: heart attacks in smokers, lung cancer in smokers, chronic bronchitis in smokers, pregnancy complications in smokers, strokes in smokers, and male impotence in smokers; and other people’s smoking as a cause of: lung cancer in non-smokers, heart attacks in non-smokers, chest problems in children, and ear problems in children). The other scale, a government responsibility scale, was based on five items (government’s responsibility regarding programmes and policies to: encourage healthy eating habits, encourage people to quit smoking, encourage people to be physically active, prevent alcohol abuse, and discourage young people from smoking). The reliability coefficients of the scales were 0.86 and 0.79, respectively.

Analysis

The data were analysed using Stata.¹⁰ Observations were weighted according to the probability of adults in the selected household being selected into the sample.³ Support for each of the eight facets of TID was determined (table 1). Bivariate analyses examined the relations between each of the six independent variables (sex, age, education, smoking status, health knowledge score, and government responsibility score) and the denormalisation score (table 2), and each facet of TID (table 3). Respondents whose views were supportive of TID on six or more items of the scale (that is, their overall score was above the median score) were classified as “high scorers” on the denormalisation scale. Multiple linear regression, using the denormalisation scale as the dependent variable, was used to examine the independent roles of the six variables as predictors of support for TID (table 3).

RESULTS

Overall attitudes toward tobacco industry denormalisation

The attitudes of the population to the individual TID items (table 1) showed wide variability. Substantial support was found for three items. Four fifths of respondents thought that tobacco should be regulated as a hazardous product, three quarters viewed the tobacco industry as dishonest, and two thirds thought that the industry should be fined for earnings from underage smoking. A majority of respondents also thought that cigarettes are too dangerous to be sold at all and that the Ontario government should sue tobacco companies for health care costs arising from tobacco. However, only a minority thought that the tobacco industry was mostly or completely responsible for the health problems that smokers have because of their smoking, that the tobacco industry is most responsible for young people starting to smoke, and that the Ontario government should sue the tobacco companies for taxes lost from smuggling.

Bivariate analysis

The percentage of respondents with a high score (above the median) on the denormalisation scale did not vary by sex, age, or educational attainment (table 2). However, respondents who did not smoke, respondents who were most knowledgeable about the health effects of active and passive smoking (above the median on the health effects knowledge scale), and respondents who thought that government was responsible for health promotion programmes and policies (above the median on the government responsibility score) were much more likely to score high on the denormalisation scale (p < 0.001).

A largely similar pattern was observed when the TID items were individually considered (table 3). Support for all of the TID items was significantly related to smoking status, knowledge about the health effects of smoking, and views about the government’s responsibility for health promotion programmes and policies. There was only one exception—views about the dishonesty of the industry were not related to the government responsibility score. Most of the items were not significantly related to education or age. However, those with higher education were more likely to favour suing for health care costs, and younger persons were more likely to think that the industry should be fined for earnings from underage smoking. With regard to sex, females were more likely than males to agree that tobacco products are too dangerous to be sold at all, and they were more supportive of fining for underage smoking and for regulating tobacco as a hazardous product. On the other hand, males were more likely to view the tobacco industry as dishonest.

Multivariate analysis

In the multiple linear regression, sex, age, and education were not found to be independently related to the denormalisation score; however, support for TID was higher among respondents who were non-smokers, more knowledgeable about the health effects of active and passive smoking, and had strong beliefs that government should be responsible for health promotion programmes and policies (table 3).

DISCUSSION

There is now strong evidence that TID can be an effective tobacco control strategy, particularly in the context of a comprehensive tobacco control programme. Much of this evidence comes from California,^11,12 Massachusetts,^13,14 and Florida.^15–17 Recently, the effectiveness of TID in the American Legacy Foundation’s “Truth” campaign was documented at a national level.¹⁸ The campaign was shown to change youth attitudes toward tobacco, in particular, the attitudes of youth at risk of smoking.

In a recent review of TID conducted for Health Canada, Lavack¹⁹ recommended that Canada’s national anti-tobacco media campaign incorporate TID. This view has also been endorsed by the federal Minster of Health’s Ministerial Advisory Council on Tobacco Control,²⁰ and by 12 national and provincial health/tobacco control groups in an open letter to the minister, published in the Globe and Mail national newspaper.²¹

Although the TID strategy is often considered in the context of mass media campaigns, the intent of TID goes well beyond this component of tobacco control programmes. TID mass media campaigns, however, appear to be critical in preparing the ground for other measures aimed at fundamentally changing how the tobacco industry and its products are regulated, and how the industry is held responsible for the effects of its products and the way it conducts its business.

Our data indicate that there is a need for TID campaigns in Ontario. In particular, less than half of Ontarians (43%) consider the tobacco industry to be primarily responsible for the health problems caused by their products, and only 55% favour holding the tobacco industry responsible for the health effects of its products. Even fewer Ontarians (22%) appreciate the pivotal role of tobacco industry in youth smoking. Other data, although sparse, lend credence to this finding. In a recent survey of Canadians 15 years of age and older (CTUMS), conducted by Health Canada, only 13% of respondents identified the tobacco industry as the main source of influence in initiating smoking.²² In developing TID campaigns, consideration may need to be given to targeting specific subgroups of the population, such as youth, or particular aspects of tobacco control, such as public policies.²³

Our data also show that most adult Ontarians view the tobacco industry as dishonest and most support the regulation of tobacco as a hazardous product, suggesting that some TID measures would be acceptable to residents of this province now and could be implemented. In her report, Lavack¹⁹ quoted findings from a national 1999 Environics Survey commissioned by Health Canada that also suggest receptivity to TID initiatives.

Our findings indicate that adults who are non-smokers and more knowledgeable about the health effects of smoking are more supportive of TID. This is in keeping with the findings of 1996 surveys of adult Ontarians⁷ and Canadian legislators,^8,9 which showed that both knowledge of health effects and smoking status are independently associated with support for more restrictions on smoking in a variety of venues and other tobacco control measures. As Canadians become better informed about the health effects of smoking, we can expect growing support for TID initiatives. Given the key roles that legislators play in implementing tobacco control measures, particular efforts should be made to assure that they are fully informed about tobacco’s public health toll.

Our finding that respondents who viewed government as having responsibility for health promotion programmes and policy were more supportive of TID is also consistent with findings from the study of Canadian legislators; in this study it was shown that legislators’ views about the role of government in health promotion were associated with support for various public policy interventions,⁹ including tobacco control measures.⁸ However, such views may not be as amenable to change as is knowledge about the health effects of tobacco use.

This study is not without limitations. Only 60% of eligible respondents took part in the survey and it is not known if or how the attitudes of non-respondents might have differed. As well, other potential predictors of TID were not explored—for example, living in a tobacco growing or manufacturing area of Ontario, or living in a community where smoking is highly regulated in public places and workplaces.

We conclude that the Ontario population is ambivalent toward TID, and that interventions (for example, mass media campaigns) are needed to ensure that the public is better informed about the underlying role of the tobacco industry in the tobacco epidemic. However, some TID strategies are now supported by a majority of the population and could be implemented with reasonable certainty of popular support. TID attitudes should continue to be monitored in provincial and national surveys, in order to gauge changing support for TID strategies. Finally, research to understand better the theoretical basis of how TID leads to reductions in smoking prevalence and increased support for policy interventions could be undertaken.