Population tobacco control interventions and their effects on social inequalities in smoking: placing an equity lens on existing systematic reviews

We searched electronic databases and library catalogues, bibliographies and reference lists for published and unpublished systematic reviews and "borderline systematic" reviews of the effects of any type of intervention to prevent or reduce smoking, access to tobacco products, or exposure to environmental tobacco smoke. The search syntax included groups of terms for "smoking" and "review". No language restrictions were applied. Further details of the search strategy are available from the authors.

We included reviews of the effects of population-level tobacco control interventions which reported characteristics of the participants in at least some of the included primary studies in terms of at least one socio-demographic variable (sex, race or ethnicity), socio-economic status (occupation, educational level or income), religion, place of residence or area-level index of deprivation. Age was also included as a socio-demographic factor if the intervention targeted vulnerable age groups (adolescents or young adults).

We defined population-level tobacco control interventions as those applied to populations, groups, areas, jurisdictions or institutions with the aim of changing the social, physical, economic or legislative environment to make them less conducive to smoking. These are approaches that mainly rely on state or institutional control, either of a link in the supply chain or of smokers' behaviour in the presence of others. Examples include tobacco crop substitution or diversification, removing subsidies on tobacco production, restricting trade in tobacco products, measures to prevent smuggling, measures to reduce illicit cross-border shopping, restricting advertising of tobacco products, (enforcing) restrictions on selling tobacco products to minors, mandatory health warning labels on tobacco products, increasing the price of tobacco products, restricting access to cigarette vending machines, restricting smoking in the workplace, and restricting smoking in public places. Such approaches could also form part of wider, multifaceted interventions in schools, workplaces or communities. We did not include interventions whose main aim was to strengthen the capacity of individuals to stop smoking or to resist taking up smoking, even if these interventions were applied to whole groups or populations (e.g. mass media health education campaigns). These are approaches that mainly rely on individuals engaging voluntarily with measures intended to help them.

To be eligible for inclusion in our analysis, reviews had to meet the two mandatory criteria for admission to the Database of Abstracts of Reviews of Effects (DARE) [15]: they had to address a clearly defined question, and to have made an effort to identify all relevant literature by searching at least one named database combined with either checking references, hand-searching journals, citation searching, or contacting authors in the field. We defined reviews as "systematic" if at least two components (interventions, participants, outcomes, or study designs) of the review question were explicitly defined and the search criteria were met, and as "borderline systematic" if two or more components of the review question could be inferred from the title or text.

Two reviewers independently screened abstracts for inclusion, extracted data, and quality assessed all included reviews using a checklist adapted from the seven criteria used for DARE [15] (Figure 1). We classified reviews into five categories of intervention: youth access restrictions, increasing the unit price of tobacco products, multi-component community-based programmes, smoking bans and restrictions, and interventions to prevent exposure to environmental tobacco smoke. Where reviews covered more than one type of intervention, the dominant area determined its classification in our summary. An overview of the key characteristics of each review is available as an additional file 1. We then grouped reviews according to the intervention type, and the participants and combined the results in a narrative synthesis.

We found 176 systematic and "borderline systematic" reviews of the effects of interventions on smoking, access to tobacco products or exposure to environmental tobacco smoke. Of these, only 25 (14%) addressed the effects of population-level tobacco control interventions [16‐40]. Nineteen of these reviews, ranging in quality (Figure 1) from four "borderline" reviews to four high-quality Cochrane reviews, reported socio-demographic data of some kind and were included in our analysis [16‐34]. These reviews included the results of 581 primary studies in total of which 82 had been included in more than one review. No study was included in more than five reviews, with the majority of the 82 studies included in two reviews. Some reviews were focused on a specific type of intervention, others had a broader focus such as community-based interventions or reducing exposure to environmental tobacco smoke. Only three of the 19 reviews explicitly aimed to assess if effects of interventions varied between socio-demographic groups [29, 32, 33]. The others focused on specific at-risk socio-demographic groups or reported that some of their included primary studies had such a focus.

Across the 19 included reviews, we found evidence about the effects of three types of population-level tobacco control intervention.

We found two reviews, both dealing specifically with young people, one based exclusively on US data [20] and one from the UK [28]. The US review found evidence that higher prices for tobacco products were associated with lower overall levels of smoking uptake and tobacco consumption by both adolescents and young adults. Four primary studies in this review included stratified analyses showing differential effects by ethnic group or sex. Two of these studies provided tentative evidence that young black Americans were more responsive to price than their white counterparts [41, 42]. The review also concluded that males were more responsive to price than females [41‐44]. In contrast, the UK data showed that females in all socio-economic groups were more responsive to price than their male counterparts [45]. In the lowest socio-economic group, however, smoking prevalence in both males and females was significantly associated with price.

We found six reviews dealing with education, law enforcement, community mobilisation, or combinations of these approaches to deter retailers from selling tobacco to minors or allowing them access to vending machines [16‐19, 28, 30]. These reviews found that enforced controls on retailers could reduce illegal under-age sales, but evidence of any effect on actual smoking behaviour was equivocal both within and between reviews [16‐19, 28, 30]. The majority of voluntary agreements with retailers had no effect even on sales. Although five reviews reported differences in the age, sex or ethnicity of participants between studies [16, 18, 19, 28, 30], none reported whether the effects of interventions varied according to these individual characteristics; nor could we deduce from these reviews whether the effects of access controls varied according to area-level socio-economic characteristics.

We found 11 reviews examining the effects of smoking bans or restrictions in a variety of population groups including adolescents, students in higher education, employees, Indigenous Australians, and people being treated for mental illness or substance misuse [21‐25, 28‐30, 33, 34]. Bans or restrictions were associated with reduced cigarette consumption at work or school [24‐26, 28, 29], but evidence of a reduction in overall consumption was less clear. Two primary studies indicated more comprehensive policies were associated with lower consumption by students both in and outside school and college [28]. Four reviews examined the effects of bans or restrictions on exposure to environmental tobacco smoke [24, 26, 33, 34] and found significant improvements in nicotine vapour levels, smoke exposure and air quality in both workplaces and public places [24, 26, 34]. Two reviews aimed to produce stratified estimates of effects [29, 33]. One included a primary study with results stratified by sex [29]. This provided tentative evidence that girls were more responsive to school-wide smoking policies than boys [29]. The second review failed to find any differential effects for the population-level tobacco control interventions [33]. Although a further six reviews reported differences in the age, sex, ethnicity or occupational status of participants between studies [22‐25, 28, 30], none reported whether the effects of interventions varied according to these individual characteristics.