Objective A systematic review to assess the equity impact of interventions/policies on youth smoking.
Data sources Biosis, Cinahl, Cochrane Library, Conference Proceedings Citation Index, Embase, Eric, Medline, Psycinfo, Science Citation Index Expanded, Social Sciences Citation Index and tobacco control experts. Published January 1995 to October 2013.
Study selection Primary studies of interventions/policies reporting smoking-related outcomes in youth (11–25 years) of lower compared to higher socioeconomic status (SES).
Data extraction References were screened and independently checked. Studies were quality assessed; characteristics and outcomes were extracted.
Data synthesis A narrative synthesis by intervention/policy type. Equity impact was assessed as: positive (reduced inequity), neutral (no difference by SES), negative (increased inequity), mixed (equity impact varied) or unclear.Thirty-eight studies of 40 interventions/policies were included: smokefree (12); price/tax (7); mass media campaigns (1); advertising controls (4); access controls (5); school-based programmes (5); multiple policies (3), individual-level cessation support (2), individual-level support for smokefree homes (1). The distribution of equity effects was: 7 positive, 16 neutral, 12 negative, 4 mixed, 1 unclear. All 7 positive equity studies were US-based: price/tax (4), age-of-sales laws (2) and text-messaging cessation support (1). A British school-based intervention (A Stop Smoking in Schools Trial (ASSIST)) showed mixed equity effects (neutral and positive). Most neutral equity studies benefited all SES groups.
Conclusions Very few studies have assessed the equity impact of tobacco control interventions/policies on young people. Price/tax increases had the most consistent positive equity impact. There is a need to strengthen the evidence base for the equity impact of youth tobacco control interventions.
Statistics from Altmetric.com
While adult smoking prevalence is declining in many high-income countries, smoking is becoming increasingly concentrated in socioeconomically disadvantaged groups, as measured by educational level, occupational class and income level. The patterning of smoking by socioeconomic status (SES) reflects the stage of the tobacco epidemic in that country.1 In countries characterised as being in the fourth (last) stage of the epidemic, which include the USA, Canada, Australia and most countries in the European Union, lower SES groups have higher levels of smoking prevalence and cigarette consumption, and lower rates of quitting compared to higher SES groups.2 ,3
SES is an important determinant of smoking uptake in young people.2 ,4 However, the relationship between SES and smoking uptake is generally less clear than that for adult smoking, reflecting the difficulty of assessing SES among adolescents. The 2005/2006 Health Behaviour in School-aged Children survey (HBSC), carried out in 39 countries, found that the relationship between youth smoking and SES varied between countries depending on their stage of the tobacco epidemic and gender.5 Low family affluence was significantly associated with weekly smoking among girls in nearly half the countries, but in only a few countries among boys. This pattern was strongest for girls in countries in stage 4 of the tobacco epidemic (North and Western Europe, Canada, USA). Fifteen-year-old girls from less affluent families in North Europe were also more likely to have started smoking earlier.5
There is good evidence on what reduces adult smoking,6 ,7 that is, cigarette price increases, comprehensive smokefree public places, antitobacco mass media campaigns, bans on advertising, health warnings and cessation support. These policies form the basis of the Framework Convention on Tobacco Control (FCTC).8 Reviews on youth smoking prevention endorse the importance of these measures for preventing smoking uptake, although the evidence on effective youth cessation support is less strong than that for adults.9 The 2012 US Surgeon General's report on Preventing Tobacco Use Among Youth and Young Adults4 stated that the evidence is sufficient to conclude that mass media campaigns, comprehensive community programmes, comprehensive statewide tobacco control programmes and increases in cigarette prices reduce smoking initiation and prevalence in youth (and taxes also reduce prevalence among young adults), and certain types of school programmes can produce at least short-term reductions in youth smoking prevalence. However, little is known about equity impact of tobacco control interventions/policies on youth smoking. We found only one review10 which had examined the equity effect of tobacco control interventions in youth. This included studies of population-level interventions published up to 2006 but not health promotion or mass media campaigns. They found no studies that had reported the effects of tobacco control interventions among young people by income, occupation or educational level.
The social gradient in smoking is not declining, and tackling inequalities in smoking is crucial to reducing health inequalities.11 ,12 It is therefore important to identify which types of policies/interventions may be effective in reducing inequalities in smoking, that is, have a greater impact on low SES young people. Health equity is defined as the absence of avoidable and unfair inequalities in health.13 This paper reports the findings of a systematic review which assessed the equity impact, in terms of SES, of interventions/policies on smoking among youth published since 1995. This review goes beyond the only previous published review12 by including all types of interventions/policies aimed at reducing smoking uptake in young people. This review forms part of the European project14 ‘Tackling socioeconomic inequalities in smoking’ (SILNE). The research question was ‘What is the equity impact of interventions/policies to reduce youth smoking?’ This was addressed by assessing primary studies of any intervention/policy that reported differential effects on a smoking-related outcome in at least two socioeconomic groups.
The study protocol is available from the corresponding author. The review followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses-equity reporting guidelines: PRISMA-E 2012 (see online supplementary file 1).15 The search strategy identified studies for this review and two other reviews for the SILNE project which covered adult targeted population-level interventions/policies and adult individual-level cessation support interventions (figure 1).
Search strategy and study selection
A comprehensive search strategy was undertaken in 10 electronic databases: Biosis, Cinahl, Cochrane Library, Conference Proceedings Citation Index, Embase, Eric, Medline, Psycinfo, Science Citation Index Expanded, and Social Sciences Citation Index (October 2013). Terms for smoking, smoking policies and outcomes, and SES were combined using database-specific terms and keywords (see online supplementary file 2).
Papers in press in four key journals (Addiction, Nicotine and Tobacco Research, Social Science and Medicine, and Tobacco Control) were identified through hand-searching (October 2013). Three key reviews were searched for relevant primary studies.5 ,16 ,17 Reference lists from included studies were also searched. Members of SILNE and the European Network for Smoking Prevention (ENSP) were contacted to identify possible additional studies. No restrictions were set on the type of intervention/policy.
Reference Manager 12 was used to produce a database of references generated from the search. A sample of the initial 200 references was independently screened by title and abstract by two reviewers (AA and TB) to establish screening consistency. One reviewer (TB) screened the remaining references which were independently checked by another (AA). Any disagreements between reviewers were resolved by discussion and, if necessary, by consulting a third reviewer (SP).
Primary studies conducted in a country in the WHO European Region or non-European country at stage 41 ,18 of the tobacco epidemic were eligible for inclusion. Experimental designs are often not appropriate or feasible for evaluating certain types of tobacco control interventions/policies, such as national media campaigns and smokefree legislation. Also, a previous review16 found no studies reporting the effects of such interventions on young people by parental SES (income, occupation or educational level). All primary study designs were deemed eligible, including randomised controlled trials (RCT), non-randomised trials, cohort studies (controlled and uncontrolled), cross-sectional and qualitative studies. The inclusion ages for study participants for this review were from birth to 25 years. This age range was chosen because smoking uptake continues until around 25 years of age.5 ,19 It also enables comparisons across studies in different countries where age of leaving secondary education varies.
Population-level interventions/policies and individual-level cessation support interventions were included. Population-level interventions were ‘those applied to populations, groups, areas, jurisdictions or institutions with the aim of changing the social, physical, economic or legislative environments to make them less conducive to smoking.’16 Any type of tobacco control intervention/policy or other type of policy (eg, educational, social), of any follow-up length, with any smoking-related outcome was included. A range of smoking-related outcomes, either self-reported or validated, was included such as: intentions/attitudes/perceptions, exposure to secondhand smoke (SHS), smoking behaviour, sensitivity to price, initiation, relapse and cessation rates, smoking prevalence, and morbidity. Other smoking-related outcomes included compliance with age-of-sale legislation, density of advertising and vending machines, brand appeal, and awareness and receptivity to antismoking campaigns. This broad range of smoking-related outcomes was included in order to encompass the diverse ways in which tobacco control policies can influence youth smoking-related outcomes.
Only studies which reported differential smoking-related outcomes for two or more socioeconomic groups were included. However, an explicit inequality or equity focus was not an inclusion criterion. Studies had to be published since 1995 in full text and in English. No settings were excluded. To be included, a measure of SES had to be reported in the abstract of the electronic references. Studies identified through hand-searching, searching of key reviews, or contacting experts, could be included if a measure of SES was reported in the main text but not in the abstract.
Data extraction and quality assessment
Data from each study were extracted by one reviewer (TB) and checked by another (AA) using a piloted data extraction form. Data were extracted on study design, population characteristics, intervention details and outcomes by SES. Details of SES variables were extracted, including how SES was measured, and data sources. All smoking-related outcomes by SES were extracted, including absolute and relative differences (the data extraction sheets for each study are available in online supplementary file 3).
All studies were quality assessed by one reviewer (TB) and checked by another (SP). Quality was assessed by adapting the method used in a previous review.16 Each study was assessed using the Six Item Checklist Of Quality Of Execution adapted from the criteria developed for the Effective Public Health Practice Project in Hamilton, Ontario,20 including representativeness of study samples; randomisation; comparability of baseline groups; credibility of data collection tools; attrition rate; and attributability to the intervention. An additional criterion of ‘generalisability’ assessed whether findings were likely to be transferable at a regional or national level. While some sources of potential bias were not applicable to all study designs, attrition and confounding issues were always considered in the quality assessment form. Particular attention was paid to internal and external validity; important quality and validity issues are discussed alongside study results.
A meta-analysis was not possible due to the variations in study designs, intervention types and outcome measures. Results are presented as a narrative synthesis according to intervention type. The equity effect of each intervention/policy was summarised by adapting a model used in a previous review (box 1).16
Box 1 Definitions of equity impact of each intervention/policy
Positive equity impact—evidence that lower SES groups were relatively more responsive to the intervention/policy.
Neutral equity impact— no social gradient in the effectiveness of the intervention by level of SES, that is, same impact on high and low SES groups.
Negative equity impact—evidence that higher SES groups were relatively more responsive to the intervention/policy.
Mixed equity impact— effect of intervention/policy varied by SES measure and/or other variables, such as gender, setting, country and/or outcome measure.
Unclear equity impact— not possible to assess the equity impact, for example, no statistical analysis by SES group.
SES, socioeconomic status.
The electronic search produced 30 381 references. Nineteen papers were identified through hand-searching, searching grey literature and key reviews, and contacting experts. After removing duplicate references and articles published before 1995, 14 798 titles and abstracts were screened. Four hundred and eleven full-text articles were assessed; 39 papers were included and 372 were excluded. The 39 included papers covered 38 separate studies (figure 1).
Methodological characteristics and quality of included studies
Thirty-eight studies of 40 interventions/policies (two studies each assessed two types of policies) were assessed for quality (see online supplementary file 4). Twenty-eight studies were population-based observational studies: 10 single cross-sectional studies, 13 repeat cross-sectional studies, three uncontrolled cohort studies and two econometric studies. Two studies were intervention-based and observational, with the same participants studied before and after the intervention. Eight studies were intervention-based experimental studies: six RCTs and two quasiexperimental trials. Of the eight experimental studies, five were school-based interventions, two were cessation support interventions and one was an online cigarette packaging experiment.
Eighteen studies had representative study samples, of which 13 were generalisable on a national scale and five on a regional scale. Four of the eight experimental trials met criteria for baseline comparability between groups. Thirty-three studies used credible data collection tools. Four out of 25 relevant studies had unacceptably high attrition rates, that is, 30% or more, which was the attrition rate threshold in the Six Item Checklist Of Quality Of Execution20 which we used. In 23 studies, it was not possible to be confident that the observed effects were attributable to the intervention, largely because of study design limitations or other concurrently implemented tobacco control legislation or policies.
Sixteen studies were based in the USA, eight in the UK, two in Germany, three in New Zealand and one each in Australia, Canada, Finland, France, Israel, The Netherlands, Portugal, Spain and Sweden. One study of smokefree policies pooled data from Scotland, Wales and Northern Ireland,21 and so, individual data from the Scottish22 and Welsh23 studies are described separately and are also included in the pooled analyses.21 Three school-based studies24–26 were identified in a paper by Mercken et al27 which included secondary analyses of effect by SES. No study included data from multiple countries outside the UK.
Participants included smokers and non-smokers in the general population aged up to 25 years. Settings included primary and secondary schools, university, child healthcare centres, paediatric practices, hospitals, a US Navy recruitment centre, and vehicles. SES variables included education,28–35 parental education,24 ,36–41 household income,33 ,42–44 Family Affluence Scale (FAS),21–23 ,25 ,45 family socioeconomic classification,22 free school meal (FSM) eligibility,25 ,46 ,47 income/spending money,26 ,41 ,48 and occupation.49 Measures of area deprivation included the Index of Multiple Deprivation/area deprivation,25 ,50–54 Index of Relative Socio-Economic Disadvantage,54 household education and income level at ‘zip’ code level,55 ,56 percentage of population with a college education, 32 ,33 ,55 percentage of youth unemployment, percentage receiving social welfare and attending ‘low-qualifying’ schools, 35 ,53 percentage below 150% poverty level,32 ,57 and school census tract data.58
Participant numbers ranged from 86 to over 641 000. Data collection periods ranged from a single time point to 24 years. The types of interventions/policies included were: smokefree (12 studies); price/tax (7 studies); mass media campaigns (1 study); controls on advertising, promotion and marketing of tobacco (4 studies); controls on access to tobacco products (5 studies); school-based prevention programmes (5 studies); multiple policies (3 studies); individual cessation support (2 studies) and individual-level support for smokefree homes (1 study). Two studies28 ,36 examined more than one type of policy. Two school-based interventions were drug prevention programmes which included elements of smoking prevention.46 ,45
Types of outcomes included intentions/attitudes/perceptions, SHS exposure, smoking behaviour, sensitivity to price, initiation, relapse and cessation rates, smoking prevalence, morbidity, brand appeal, advertising density, awareness/receptivity, and compliance.
The findings of the equity impact of each type of intervention/policy are summarised in table 1 (for summary equity of each study see online supplementary file 5).
Seven studies28 ,29 ,36 ,37 ,42 ,48,59 examined the equity impact of cigarette price/tax increases. Five were US-based studies using retrospective survey data. Four studies demonstrated a positive equity impact,36 ,37 ,42 ,48 one study was neutral29 and two negative.28 ,59 Four of the five US studies had a positive equity impact including the most recent study which analysed data from 1991 to 2010 in 14–18-year-olds. Low SES youth were more responsive to price/tax increases than high SES youth.36 The evidence suggests that there is variation in the impact on smoking behaviour among youth of different ages within different SES groups.
Smoking restrictions in cars, schools, workplaces and other public places
Twelve studies21–23 ,30 ,36 ,38–40 ,50–52 ,58 examined the equity impact of smoking restrictions in public places: 11 cross-sectional studies and 1 intervention study. Six studies assessed comprehensive national/state smokefree legislation, and six studies assessed policies that were voluntary, including one intervention study that assessed a worksite total smoking ban enforced for 8 weeks military training.30
The equity impact of comprehensive national/state smokefree policies was: two neutral50 ,51 and four negative.21–50 Five UK studies explored the equity impact of comprehensive national smokefree legislation on primary school children's exposure to SHS,21–23 parental smoking in cars,21 ,23 and hospital admissions for childhood asthma.50 ,51 In Scotland, there was a significant reduction in SHS exposure for all school children. However, while the greatest absolute reduction was in low SES children, the relative equity impact was negative.22 In Wales and Northern Ireland,21 ,23 declines in exposure were restricted to high SES children (negative equity impact). One US study36 showed that comprehensive smokefree air laws reduced smoking prevalence in the total sample of 14–18-year-olds and for high SES youth, but there was no significant effect for low SES youth (negative equity impact).
National smokefree legislation did not significantly impact on car and home-based smoking restrictions: socioeconomic patterning remained stable. English and Scottish national smokefree legislation was associated with a significant reduction in childhood asthma admissions which did not differ by SES50 ,51 (neutral equity impact). The equity effect appeared to vary according to how exposure was measured (absolute levels or relative levels); on the preban levels of exposure; and the balance between sources of exposure, that is, public places versus home.
The equity impact of voluntary smokefree policies was: four neutral,30 ,38 ,39 ,58 one negative36 and one mixed (neutral and negative).40 Three of these studies investigated smoking behaviour in cars with children, in the context of no smokefree vehicle legislation (one negative equity impact and two neutral for equity impact).38 ,52 ,58 It was unclear whether any of these studies38 ,52 ,58 had representative study samples or produced results which are generalisable on a national scale. National New Zealand surveys across 7 years among 25 000 14–15-year-olds showed a significant decrease in ‘past 7 days’ exposure to in-vehicle SHS.58 However, marked differences by SES persisted, with lowest SES students reporting the highest in-vehicle SHS exposure, and the highest SES students reporting lowest SHS exposure. Two studies explored the impact of voluntary compliance with smoking restrictions on smoking behaviour in secondary school children in Spain40 and Israel.39 One study40 found that parental education level influences adolescents’ smoking behaviour in school; however, at the school level SES, there was no significant impact (mixed equity impact). The other study39 found no association with parental educational level (neutral equity impact).
A 24 h 8-week workplace smoking ban in female US Navy volunteers reduced the proportion of women smoking immediately post-ban, but most had relapsed by 3-month follow-up.30 Educational level did not significantly predict smoking relapse (neutral equity impact); however, loss to follow-up was high and non-respondents were more likely to be smoking.
Mass media campaigns
One study55 evaluated the impact of the US truth campaign on awareness and receptivity among youth aged 12–17 years. The truth campaign is a branded counter-tobacco marketing campaign designed to prevent smoking among at-risk youth. There were inconsistencies in results by outcome (campaign awareness/receptivity) and also by SES variable (income/education) for campaign awareness. Overall, the equity impact of the mass media campaign was mixed.
Controls on advertising, promotion and marketing of tobacco
Four US studies28 ,31 ,32 ,56 examined the equity impact of controls on advertising, promotion and marketing of tobacco. Two studies31 ,32 had a neutral equity impact and two studies demonstrated a negative equity impact.28 ,56 The limited evidence suggests that, when there is no enforced control of advertising, promotion or marketing of tobacco, there is the potential for an increase in inequality in youth smoking.
Controls on access to tobacco products
Five studies33 ,41 ,47 ,53 ,57 examined the equity impact of controls on access to tobacco, including age-of-sale legislation33 ,57 and electronic locking devices on cigarette vending machines.53 A US study41 examined the impact of several policy elements including: statewide enforcement, random inspections, graduated penalties, photo identification, free distribution, minimum age, packaging, vending machines, and clerk intervention. Two studies demonstrated a positive equity impact,33 ,41 two studies a neutral a neutral equity impact,33 ,47 ,57 and one a negative impact.53 Stronger (ie comprehensive and enforced) state-level tobacco policies on age-of-sale were associated with lower smoking initiation and adverse transition among low SES adolescent girls (positive equity impact), though the effect sizes were small.41
Five studies24–26 ,45 ,46 examined the equity impact of school-based interventions in 11–14-year-olds. Two of these studies were drug prevention programmes which included smoking prevention.46 ,45 Two had a neutral equity impact,46 ,45 one a negative equity impact,24 one a mixed equity impact,25 and one study's equity impact was unclear.26 The results varied by type of SES measure, over time and by gender. The ‘A Stop Smoking in Schools Trial’ (ASSIST) assessed the effectiveness of a peer-led intervention using a social network approach, that aimed to prevent smoking uptake in secondary schools in England and Wales.25 No significant main effect of the intervention was found for FAS or FSM entitlement. The intervention was significant among schools located in the valleys which are areas of economic deprivation, but not in schools in other locations. While this intervention showed mixed equity results depending on the SES indicator used, it was effective in reducing smoking uptake at 1 year and most effective for adolescents in deprived areas, particularly low SES girls (positive equity impact). However, the intervention effect seemed to attenuate after 2 years.25
Three studies43 ,49 ,54 examined the equity impact of multiple policies, including ‘high tobacco-control activity’ in Australia54 (smoking restrictions and increased tax on cigarettes), ‘smoking intolerance’ in Canada43 (access and restrictions), and the Finnish Tobacco Control Act49 (smokefree, age-of-sale and health warnings). Two studies showed a neutral equity impact,43 ,54 and one study had a mixed equity impact49 (positive equity impact for low SES men and neutral impact for low SES women).
Individual-level cessation support interventions
Two studies34 ,44 examined the equity impact of cessation support. Both used text-messaging interventions. Participants in both studies were mobile phone owners in their late teens to early twenties, who were motivated to quit smoking. A New Zealand study44 showed personalised mobile phone text-messaging support could significantly increase quit rates at 6 weeks and at 12 weeks but not at 26 weeks, irrespective of income level (neutral equity impact). A US study34 of tailored text-messaging compared to a non-tailored text-messaging showed a significant increase in quit rates in the intervention group compared with control participants at 4 weeks postquit, but not at 12 weeks. However, low SES intervention participants had greater odds of quitting smoking at 12 weeks compared to high SES intervention participants (positive equity impact).
Individual-level smokefree homes interventions
A Swedish intervention35 to reduce children's SHS exposure in their homes showed a significant decrease in the proportion of children exposed to SHS (measured by cotinine levels) at 1-year follow-up. Significantly more mothers with a higher level of education (>12 years education) participated in the intervention compared with non-participating families. Seven of nine quitters had lower level of education (12 years or less). No statistical associations between nurses’ actions, outcomes, and parents’ sociodemographic data were found. Overall, the equity impact was negative.
The review identified 38 studies which evaluated the equity impact of 40 interventions/policies. Thirty-five studies were population-level tobacco control interventions/policies, and three were individual-level cessation support and/or smokefree homes interventions. Overall equity effects for all types of interventions/policies were: seven positive equity impact, 16 neutral equity impact, 12 negative equity impact, four mixed equity impact and one unclear equity impact.
The only consistent effect was for price/tax interventions/policies, which had the most consistent positive equity impact. Seventeen interventions/policies had a neutral equity impact and indicate that these interventions/policies have benefits for youth across all SES groups. Only seven studies, all in the USA, showed the potential to produce a positive equity impact. These included four studies of increasing the price/tax of tobacco products,36 ,37 ,42 ,48 two studies on control of youth access to cigarettes,33 ,41 and one text-messaging cessation intervention.44
Four US studies36 ,37 ,42 ,48 demonstrated a positive equity impact of increasing price/tax in reducing smoking: low SES youth were more responsive to price/tax increases than high SES youth. The evidence suggests there is variation in the impact on smoking behaviour among youth of different ages within different SES groups. National US survey data36 of over 15 000 14–18-year olds from 1991 to 2010 showed differential policy effects by SES: price policies had a positive equity impact, and smokefree laws had a negative equity impact.
UK comprehensive smokefree legislation was associated with significant declines in SHS exposure in all primary school children and significant reductions in childhood asthma hospital admissions for all children, regardless of SES. The evidence suggests that the equity effect of smokefree legislation may vary according to how SHS exposure is measured (absolute levels or relative levels), by preban levels of exposure and by setting (home/public places). Evidence shows significant variation by SES in levels of SHS exposure prior to smokefree legislation, with higher levels of SHS exposure among lower SES school children. Declines in SHS exposure occurred predominantly among children who had low SHS exposure prelegislation, and who were from more affluent families, indicating a negative equity impact. Substantial SES gradients in children's SHS exposure levels remained unchanged. As SES increased, the likelihood of full smoking restrictions in the home and/or car increased significantly, and this socioeconomic patterning remained stable. Welsh data showed that children from low SES households continued to perceive adult smoking as the norm, but this perception was less marked among high SES children.23
Evidence suggests that strict control of youth access to cigarettes could benefit all youth and may have potential equity benefits for low SES youth. A US prospective cohort study41 showed that stronger state-level tobacco policies on age of sale were associated with lower smoking initiation and transitioning to smoking among low SES adolescent girls, although the effect sizes were small.
Five school-based smoking prevention programmes showed mixed equity results. The UK ASSIST programme25 was effective at 1 year and most effective for low SES girls. However, none of the five school-based interventions demonstrated an overall positive equity impact. Changing social norms is a long-term process which may also be more challenging among low SES youth and may partially account for lack of positive equity impact among the school-based interventions.
The evidence on advertising controls and multiple tobacco control policies was so disparate and limited that it was impossible to draw any conclusions about the equity impact of these two types of policy. One study31 indicated that controlling the promotion of cigarettes through plain packaging might have a positive effect on all young women.
Two text-messaging interventions increased quit rates among young adults in the short term, and one of these showed potential equity benefits for low SES youth. The representativeness of both study samples was unclear as they were motivated young adults who owned a mobile phone. However, text messaging interventions have the potential to reach large numbers of young smokers.
Strengths and limitations
This review systematically assessed all the available evidence on the impact of tobacco control interventions/policies on socioeconomic inequalities in youth smoking, which is relevant to European countries and other countries at stage 4 of the tobacco epidemic. This review goes beyond the previous review16 in including all types of youth interventions/policies (prevention and cessation), at the population and individual level; and also searching for non-tobacco control interventions/policies (eg, education, social policy). Considerable attempts were made to include as much evidence as possible, including articles not published in the peer-reviewed literature, and articles recently published online but not yet indexed on electronic databases. However, it is possible that studies which undertook analyses by SES but did not report this in their abstract were missed.
There are major limitations in the available evidence, most importantly the very small number of studies which have considered the equity impact of tobacco control interventions aimed at young people. Several studies were pilot or feasibility studies, and/or involved small numbers of participants. Thus, their findings may not be replicable. Most of the primary study designs included in this review would not meet the criteria used by other systematic reviews. However, search results confirm that very few included studies were experimental in design and these mainly relate to school-based programmes. The vast majority of evidence was derived from studies using research designs which fail to deal with typical threats to internal validity, especially the problem of causal attribution. Nearly half the studies were from North America and a quarter from the UK, which raises concerns about the generalisability, transferability and relevance of findings for other stage 4 countries which have different sociocultural contexts and/or different levels of tobacco control. Additionally, a range of indicators of SES was used in studies, which made comparisons between studies difficult.
It should also be borne in mind that studies used different outcome measures, reflecting differences in study designs and intervention/policy aims. While some looked at the impact on smoking incidence and prevalence such as smoking uptake in school pupils (eg, school interventions), others looked at more intermediate outcomes in terms of youth smoking uptake, such as changes in attitudes (eg, mass media campaigns) and perceived social norms (eg, smokefree legislation), or included both, such as reduced prevalence and/or reduced consumption (eg, tax/price increases). In this review, all outcomes were treated as being of equal value. However, these outcomes may not be of equal importance in terms of reducing youth smoking or the immediacy of intervention/policy effects.
For several important areas of youth tobacco control, for example, social marketing, multifaceted community programmes, mass media approaches using social media, combating smuggling/reducing the black market, and most forms of cessation support, we found no evidence on equity impact.
Only 38 studies assessed the equity impact of interventions/policies on smoking prevention or cessation in youth. The distribution of equity effects across interventions was: 7 positive, 16 neutral, 12 negative, 4 mixed and 1 unclear. Most of the neutral equity studies were beneficial for all SES groups. There was variation in the equity impact of each type of tobacco control policy/intervention. Only seven of the intervention/policies showed the potential to reduce inequalities in youth smoking and all were US-based: four studies of increasing the price/tax of cigarettes, two studies of enforcing strong policies on age-of-sale, and one study of smoking cessation support through text-messaging. One school-based prevention study, ASSIST, showed potential for a positive equity impact in adolescent girls from deprived areas in Wales.
The limited nature and extent of the evidence constrains what conclusions can be drawn about which types of tobacco control interventions/policies are likely to reduce inequalities in youth smoking. The review provides very little evidence to suggest that any specific policy reduces inequalities in smoking initiation, with the exception of price/tax policies which had the most consistent positive equity impact. Overall, more interventions/policies were found to have a negative equity impact (12), and therefore to potentially exacerbate inequalities in youth smoking, than had a positive equity impact (7). There is an urgent need to strengthen the evidence-base for the equity impact of tobacco control interventions aimed at young people. This should include evaluation studies which explore the mechanisms of change of interventions and policies, as well as outcomes, in order to understand why these are relatively more or less effective with low SES groups, and thereby inform the development of equity-positive tobacco control interventions and policies.
What is already known on this subject
While overall smoking prevalence in young people has fallen in many high-income countries, there are increasing inequalities in youth smoking uptake by socioeconomic status.
There is a lack of evidence on the equity impact of tobacco control interventions and policies on inequalities in smoking in young people.
What this paper adds
Very few studies have assessed the equity impact of tobacco control interventions and policies on smoking in young people.
Increased tobacco price is the only tobacco control intervention that had a consistent pro-equity impact on youth smoking.
More research is needed to identify interventions and strategies with the potential to reduce socioeconomic inequalities in youth smoking uptake.
We would like to thank Thomas Tjelta for his help in acquiring some of the papers, Anton Kunst for his helpful comments on the review and ENSP for contacting members about relevant grey literature. This study is part of the project ‘Tackling socio-economic inequalities in smoking (SILNE)’, which is funded by the European Commission, Directorate-General for Research and Innovation, under the FP7-Health-2011 programme.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Contributors TB, SP and AA designed the study. TB wrote the first draft of the paper and undertook the search, screening, data extraction and quality assessment. AA oversaw the whole study and was involved in the screening and data extraction. SP was involved in the screening and quality assessment.
Contributors TB, SP and AA developed the initial strategy for the review; TB undertook the literature search with support from AA and SP: TB, SP and AA reviewed individual articles; TB wrote the first draft of this paper; AA and SP contributed to the writing of the manuscript and agree with its results and conclusions.
Funding This research was supported by the European Commission, Directorate-General for Research and Innovation, under the FP7-Health-2011 programme, with grant agreement number 278273.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
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.