Against a background of declining tobacco use and generally positive changes in other heart disease risk factors, a systematic review of 14 multiple risk factor intervention trials for preventing coronary heart disease1 concluded that reductions in mortality in the intervention groups were insignificant and changes in risk factors only modest, when compared with the reductions also seen in control groups. The Minnesota Heart Health Program reported similar outcomes2 3 and the major multi-community smoking cessation trial, COMMIT4 5 had a similar modest effect on smoking. Compared with typical community health promotion initiatives which operate on token budgets, all of these interventions were large scale, although still were funded with petty cash when compared with the promotional budgets used by the tobacco industry. Favourable improvements in the secular trend for risk factors such as smoking, and programme contamination of control groups have generally been cited as putative explanations of the lack of difference between intervention and control groups, with media leakage—being the most uncontrollable factor—deemed responsible.6 Doubtless, some of this leakage involved news coverage of specific interventions intended only for the eyes and ears of the experimental populations. However it is myopic to assume that it is only discrete efforts orchestrated by health agencies leaking into control areas which, in aggregate, constitute the possible forces generating positive secular trends in the wider population.

In a recent issue of Tobacco Control, Melanie Wakefield and Frank Chaloupka called for more attention to the description and quantification of tobacco control “inputs”.7 They noted that the preoccupation with outcomes in evaluation research is often accompanied by overly casual accounts of the policy and intervention variables that are assumed to be the causative factors potentially producing change. In their editorial they argued for the further development of a range of indices to measure the comprehensiveness of tobacco control policies and programmes. They also noted the importance of quantifying and accounting for “environmental” issues such as unpaid media coverage of tobacco issues. This importance cannot be over-emphasised: were it possible to quantify all media coverage of tobacco in societies with 24-hour access to a multitude of radio, television, internet, and print media, it would almost certainly be the case that, in aggregate, this coverage would routinely eclipse even the most intensive coverage gained through formal public health “campaigns”. The current evaluation of ASSIST (American Stop Smoking Intervention Study for Cancer Prevention), for example, has apparently incorporated over 100 000 American newspaper clips on tobacco issues from 1993 to 1999—radio and television, which are more elusive and expensive to capture in research, would have added hundreds of thousands more. And much of this massive reportage is not easily dismissed as inconsequential ephemera: some of the most potent and recalled episodes in the history of tobacco control have been powerful prime-time television documentaries, prolonged episodes such as presidential candidate Bob Dole's foot-in-mouth saga over tobacco policy8 and the coverage of legal cases.

Tobacco control has long been highly newsworthy.9 10 In a recent year, 38% of all front pages of the Sydney Morning Herald carried at least one health story.11Of these, tobacco stories ranked second after those about health services. From a journalist's perspective, tobacco control offers rich pickings that are likely often to conform with editors' notions of newsworthiness. Our field is resplendent with stories of conflict, corruption, moral rectitude, and reprobation. To the endless fascination of the media, practically every organ of the body can be afflicted by tobacco use and tobacco's stratospheric toll on health lends itself to numerous excursions into quantification rhetoric.12 Celebrities' efforts to quit or criticism directed at the influence of their smoking on young people are now routine news events. As the tobacco industry is fond of noting with its unique brand of cynicism, the tobacco epidemic has generated an epidemic of research, much of which is covered by the news media. The media's appetite for villains finds bountiful sustenance in the form of the tobacco industry, which has entered journalistic lexicons as a shorthanding index case device for referencing ethical low life.13 Repeated framing of the tobacco industry in such an unfavourable light seems likely to be associated with the community's ranking of tobacco industry representatives' trustworthiness as lower than that of used car salesmen, traditionally the low-water mark in expressions of ethical behaviour.14This in turn may be a critical factor in generating political antipathy toward the tobacco industry in some political circles—as a former president of the American Cancer Society noted: “Most politicians know that you don't stand too close to a pariah in the next photo op.”

In addition to news coverage of tobacco, communities are exposed to a wide range of other discourses about smoking in the mass media. Advertisements for shared accommodation commonly list non-smoking as an essential attribute in would-be housemates,15 as do “lonely heart” advertisements. Employment advertisements frequently state the employers' policies on smokefree workplaces, as does the voluminous advertising for life insurance which almost universally offers large reductions for non-smokers.16 Restaurant reviews and guides now often note the smoking policy of each restaurant, as do notices for an as yet small, but increasing, number of music venues.

There is some salutary historical evidence that coverage of smoking in the news media can be influential in promoting smoking cessation. For example, Reid et al 17identified unpaid media publicity as the main cause of the 30% decline in smoking prevalence among British males in the 20 years after the publication of the first report on smoking and health by the Royal College of Physicians of London in 1962. During this time, anti-smoking information and publicity was almost wholly disseminated via news reportage and commentary, with major government health promotion campaigns not starting until the 1980s. On a smaller scale, an American evaluation of a week-long local newspaper series on smoking cessation, which caused around 4% of readers to quit for at least one week, was calculated as having an impact equivalent to that which would have resulted from the establishment of 380 dedicated cessation clinics.18

The way in which the media frame issues for public consumption influences public perceptions, if not behaviour. A good example of this is the growth of public and political concern about illicit drug use. In observing that the issue of drugs being considered by the public as the most important problem facing the United States increased from 5% in 1985 to 60% in 1994, Fan19 undertook an ingenious study of news coverage from the New York Times, Washington Post, and United Press International newswire from 1985 through 1994. Fan's time series analysis showed that change in this opinion was explained by increases in the news media describing drugs as “a crisis”, whereas other types of discussion about drugs contributed negligibly. In the reverse direction, public opinion about drugs was a weak but significant contributor to press coverage. In tobacco control, there is evidence that the extent of news coverage about tobacco issues influences tobacco consumption. Laugesen and Meads20 in New Zealand showed that consumption, measured by weekly purchases of tobacco from a selected number of tobacco outlets, was significantly related to the weekly number of news stories about tobacco, although the effect was short-lived. They estimated that news coverage had approximately the same impact as a 10% increase in price.

Two papers in this issue of the journal admirably extend the research field of media studies in tobacco control. Kennedy and Bero document the emerging reportage of passive smoking in the American press over 14 years.21 Part of their interest lay in examining the process of patronage of industry consultants through which passive smoking continues to be reported as “controversial”—an appellation now seldom used to describe the relationship of active smoking and illness outcomes. Poyant and Siegel's study of theWashington Post's reportage of the American tobacco settlement22 uses frame analysis to consider the dominant ways in which the reportage effected closures around what was thereby defined to be at issue in the debate surrounding the settlement. Both papers not only quantify one dimension of the “grey”, largely unresearched “inputs” that characterise modern cultural discourse on smoking, but they also offer strategic insights to tobacco control advocates seeking to become more active participants in such discourse.

It is worth reflecting on the intense and prolonged preparation and development that characterises the research development process in public health, when at the end of the day the great majority of published studies resulting are never cited even once, and library copies of the journals in which they are published are seldom, if ever, taken off the shelf.23 By contrast, given the opportunity to appear on a national or international television news programme, tobacco control advocates and what they have to say can often be exposed to tens if not hundreds of millions of people. The corresponding preparation for such opportunities is typically fleeting. Studies of media reportage such as the two published in this issue can serve as invaluable background in the project of preparing for framing and reframing in tobacco control debates.

Despite the depressingly common conclusion that many health promotion interventions do not “work”,1-5 there are too few signs that the tobacco research community has yet taken seriously the implications of the downward secular trend in smoking. In many communities, smoking appears to be declining sometimes desspite many formal interventions. There can be few more important questions than understanding the nature of dominant forces driving this secular trend, and few more perplexing ones than why its study does not command the same or greater research attention as the interest that the study of discrete, sponsored interventions typically generates. The research privileging of sponsored interventions and the comparative neglect of the study of background media “noise” about smoking, such as the examples given earlier, is probably best explained by the political imperatives of evaluation: funding agencies want to know what their investment has achieved and may be less concerned about issues they feel they do not control, or for which they cannot take credit.24

Analysis of this vast and relatively neglected potential source of influence on public and political views about tobacco presents many research challenges. Methodological reductionists seeking neat categories to fit seamlessly into their modelling efforts will mostly settle for content analytical approaches, where frequencies can be calculated and hypotheses put to empirical test. But media discourse is inherently qualitative, reflexive and contextual, reflecting the fluid and interactive nature of media discourse. If we wish to understand the process through which the tobacco industry's efforts to frame tobacco control as health fascism jostles for public and political ascendancy with our efforts to frame the industry as run by those with the ethics of a cash register, it will not be sufficient to count just their respective media “hits” and declare a winner. Discourse analysis,25 26 studies of the ways in which key audiences decode meaning from media performances, and studies in the encoding of news by media workers will all be required. Tobacco Control will welcome submissions that bring these hitherto largely background issues more into the foreground of core research questions being addressed in our field.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1-1

Single digit pseudoclubbing caused by smoking History: A smoker since childhood, now aged 43 years, presenting for elective surgery, was discovered by his anaesthetist to have “pseudoclubbing' of one thumb. Many years ago he had inserted a coin into a cigarette vending machine. The machine did not keep to its part of the bargain. With some vigour he pressed the refund button, and was somewhat alarmed as his thumb became stuck in the machine. He was much more alarmed when a painful thud signalled the descent of Something Very Heavy inside the machine onto his thumb. He managed to extract his rapidly swelling digit, and without medical (or legal) advice nursed it for three weeks till the pain subsided. The swelling remains to this day.

Comment: With a careful history taking and examination doctors can explain the connection between smoking and almost anything.

Submitted by John Curran, Department of Anaesthesia, Nottingham City Hospital, Nottingham, UK.