The tobacco endgame concept reorients discussion away from the persistent control of tobacco toward plans for ending the tobacco epidemic, and envisions a tobacco-free future. A variety of policy approaches have been proposed, with many offered prior to the introduction of the unifying term ‘endgame’. We conducted a qualitative synthesis of the literature on tobacco control endgames, and drew on media accounts and discussion of analogous ideas for illustrative purposes. We identified proposals focused on the product, user, market/supply or larger institutional structures. Research on public support for these proposals was limited, but suggestive of some public appetite for endgame ideas. Advocates should be encouraged to explore new policy options and consider the goal of a tobacco-free future.
Statistics from Altmetric.com
The tobacco ‘endgame’1–3 concept suggests moving beyond tobacco control (which assumes the continued presence of tobacco as a common, widely-available, ordinary consumer product) toward a tobacco-free future wherein commercial tobacco products would be phased out or their use and availability significantly restricted.4 No single definition of the endgame (ie, the method or policy approach), or the end point (ie, the specific, measurable outcome) has emerged and the word ‘endgame’ carries translational challenges. Nevertheless, the idea—of seeking to end the tobacco epidemic, rather than control it—has become a focal point for national and international meetings and has spurred longer term planning. Identifying the full range of endgame thinking is challenging, as many relevant proposals were published prior to introduction of the unifying term ‘endgame’. We conducted a qualitative synthesis of the literature on tobacco control endgame strategies, defining them as: Initiatives designed to change/eliminate permanently the structural, political and social dynamics that sustain the tobacco epidemic, in order to end it within a specific time.5
The lack of index terms and variety of literature encompassed under the endgame term necessitated a ‘snowball’ retrieval approach. We started our search with the May 2013 Tobacco Control supplement on the tobacco endgame. The supplement contained 20 articles, including editorials, endgame proposals, commentaries and perspectives. We included those in our review, and searched their reference lists for additional relevant materials, finding 38 articles.
For articles published since the supplement went to press in February 2013 through March 2015, we searched the PubMed database using the search term ‘tobacco AND endgame’. Exclusion criteria included non-English language and focus on a tobacco control issue other than endgame. This added eight articles. These articles suggested other search terms, including ‘nicotine reduction’, and ‘e-cigarettes AND endgame’. Those searches added 7 articles, for a total of 73 articles. We included articles describing or discussing endgame-related ideas or proposals, and research related to such proposals, for example, measuring levels of public support for or modelling policy impacts of endgame scenarios. Reviewers suggested five additional sources of relevant public opinion data.
The endgame idea is still developing; much of the literature discussed here is exploratory. Thus, this is not a ‘systematic review’ that collates all empirical evidence in order to answer a particular research question;6 instead, it represents the most thorough qualitative synthesis to date of tobacco endgame scholarship. In synthesising the literature, we categorised the articles as focusing on the product, the user, the market/supply, or larger institutional structures. We also drew on media accounts of endgame proposals and on discussions of analogous ideas for illustrative purposes. We integrated some material from published critiques (indicated via citations) and added our own assessments.
Product-focused endgame proposals
Regulate nicotine levels to make cigarettes non-addictive or less addictive
Nicotine levels in cigarettes or all combustible tobacco products could be regulated to reduce tobacco use among current smokers and prevent new smokers from becoming addicted.7–9 Nicotine levels could be reduced gradually, over 10–15 years, to wean addicted smokers off nicotine,7 ,10 or more abruptly.11 Theoretically, nicotine levels could be established that maintained addiction in current smokers but minimised addiction among new smokers.12 Simultaneously, cleaner nicotine products such as patches and gum could be promoted by reducing their cost and increasing their availability.10 This approach would require determining the appropriate level and method and speed of administration of nicotine, and establishing regulatory authority over the tobacco industry. More stringent limits on marketing and availability of tobacco could ensure that ‘clean’ nicotine products dominated the market.10
Potential risks include the tobacco industry marketing reduced nicotine tobacco products as ‘safer’, or ‘government approved’, thereby promoting uptake and sustained use of these products among people who might otherwise not become addicted or quit smoking.12 Using low nicotine combustible tobacco as ‘starter’ products and then switching to other products having higher nicotine levels would be another potential problem if nicotine levels were not reduced in all tobacco products.13 A ‘starter’ effect could potentially occur with pharmaceutical nicotine products as well. In addition, individuals might find ways to add nicotine to tobacco products, and the tobacco industry might offer new, companion products to sustain nicotine addiction.13 Robust product testing and enforcement would be needed to ensure manufacturer compliance.14 Prohibition of products that added nicotine to cigarettes (eg, spray-on or injectable nicotine) would also be necessary.
Compensation is another concern. Smokers of reduced nicotine tobacco might smoke more or more intensively in order to maintain their usual nicotine level, increasing their exposure to toxic combustion products.7 Research on reduced nicotine content cigarettes is limited, but shows that those who smoked such cigarettes in laboratory conditions (over two 8 h sessions), for 6 weeks and over 6 months did not compensate for the reduced nicotine yield by smoking more cigarettes or smoking more intensively.15–20 Their level of exposure to toxic combustion products also remained stable,15 ,19 ,20 or, in some cases, was reduced.16 ,18 However, heavily addicted smokers may compensate more than others.21 A comparison of two strengths of reduced nicotine content cigarettes showed that greater reductions in nicotine were associated with higher rates of abstinence 1 month after a 6-week trial.9
Redesign the cigarette to make it unappealing
Raising the pH of cigarettes to 8 or more (its level prior to 20th century methods of cigarette manufacturing) would make cigarettes harder to inhale.22, p.553 A more acrid smoke that cannot easily be drawn deep into the lungs could reduce both smoking uptake and the risk of lung cancer.23 Banning all non-tobacco cigarette ingredients would make cigarettes distasteful, discouraging uptake and encouraging cessation.24 A ban on menthol, which masks the harshness of cigarettes and facilitates exposure to nicotine, has also been proposed,25–27 as has banning filters, which provide no health benefit but reduce irritation and may make it easier for those experimenting with smoking to become regular smokers.28 ,29
There have been several attempts to ban menthol in cigarettes (eg, Brazil in 2012, Chile in 2013).30–32 In February 2014, the European Parliament approved the European Union Tobacco Products Directive, which included a ban on characterising flavours in cigarettes and roll-your-own tobacco, including menthol; it will be phased in over 4 years.33 In May 2015, the Canadian province of Nova Scotia became the first jurisdiction in the world to implement a ban on menthol-flavored tobacco.34 It will be followed in September by the province of Alberta.34 Banning particular constituents could lead the industry to add other ingredients to counter the effects of their removal, creating the need for additional scientific evidence about the effects of such new ingredients, which would be challenging to obtain.35
The popularisation of electronic cigarettes (e-cigarettes) and the rapid innovation in the category caused some endgame commenters to suggest that these may be part of an endgame scenario in which combustible cigarettes are made less appealing through nicotine reduction and addicted cigarette smokers switch to or use e-cigarettes to quit.36–38 The intense controversy over these products’ marketing and use, lack of long-term research on their health effects,39 ,40 variability among the products themselves, introduction of novel next-generation products and the political dynamics of rapid acquisition by cigarette companies of e-cigarette companies and products complicate consideration of how they might figure in various endgame scenarios. A full consideration of the implications of e-cigarettes for endgame purposes is beyond the scope of this review. However, it is clear that jurisdictions undertaking endgame planning must consider the role of this rapidly growing market segment, and must do so in the context of limited scientific knowledge (particularly about longer term effects, dual use and the potential for these products to lead to use of tobacco products).
User-focused endgame proposals
Under a smoker licensing scheme41 those seeking to purchase tobacco from licensed retailers would obtain a ‘smart swipecard’ smoker's license, renewable annually, with purchase limits established by the user. Financial incentives to permanently relinquish the license could be offered, and new smokers seeking a license would demonstrate knowledge of tobacco's health and financial costs. The legal smoking age could be raised annually by 1 year; since smoking initiation usually occurs among those under 23 years of age, new requests for smokers’ licenses would presumably decline rapidly after several years.
Critics of this proposal argue that it is financially burdensome to smokers.42 Moreover, a knowledge test of smoking's dangers could reinforce the tobacco industry's framing of smoking as a choice made by fully informed consumers, deflecting attention from industry behavior;42 new smokers could also have difficulty passing a test that assessed more than superficial levels of knowledge.43 Given that smoking is increasingly concentrated among the poor, stigmatising them further as ‘registered addicts’ has social justice implications.44 However, supporters consider the license to be akin to a prescription for access to pharmaceuticals,41 or licenses to obtain medical marijuana.45 As with driver's licenses, the issuing agency or other parties would likely provide study material and practice tests that would enable most applicants to pass the knowledge test. Requiring a license before accessing the products would send a stronger social message about their dangerousness than is conveyed through current practices.
Prescription to purchase tobacco
In Iceland in 2011, a former health minister sponsored a bill to limit cigarette sales to pharmacies and require purchasers (aged 20 and over) to obtain a prescription. A prescription would only be given after cessation efforts had failed.46 ,47 Similarly, in the USA, an Oregon lawmaker proposed in 2013 that the state classify all nicotine products as controlled substances, available only by prescription.48 It is unknown whether doctors would write prescriptions for cigarettes, or if pharmacists would fill them; however, clinicians would most likely have ethical objections. In addition, in jurisdictions where tobacco is still sold in such establishments, such a plan would directly conflict with policy efforts to end sales of tobacco products in pharmacies.49–52
Restrict sales by year born
Researchers in Singapore have proposed prohibiting the sale of tobacco to citizens born in or after a certain year (eg, the year 2000), thereby creating ‘tobacco free generations’ legally barred from purchasing tobacco at any age.53 ,54 In effect, this would phase out the sale of tobacco: when the population is composed entirely of those born after the chosen year, tobacco would cease to be sold legally. Although there would presumably be some demand for illegal sales initially among those barred from purchasing tobacco, this might decline as the smoking population aged and smokers became undesirable youth role models.53 Media campaigns could portray smoking as ‘a ‘last century’ phenomenon'.54, p i23 Jurisdictions might choose to sanction underage smokers (although focusing attention on smokers rather than the tobacco industry lets the industry off the hook).53 A tobacco-free generation bill was introduced to the Australian state of Tasmania's parliament in November 2014; as of May 2015, a vote on the bill had been delayed.55
One objection to the proposal is that it denies adults the ability to ‘take informed risks’.41 However, consumer choices are frequently constrained, particularly regarding toxic products.54 Moreover, because the majority of smokers begin smoking before age 18, a ban on the sale of cigarettes would constrain the choices of a small minority of adults.54 An analysis of how the proposal would comport with the Universal Declaration of Human Rights, the International Covenant on Civil and Political Rights, and the Convention on the Rights of the Child concluded that it was consistent with all three; its support of rights to life, health and a healthy environment outweighed its relatively minor restrictions.3
Precedent for the licensing and tobacco free generation proposals occurred in Taiwan and British Ceylon in the early part of the 20th century.54 Smoking opium was phased out by requiring smokers to display a license in order to purchase opium; after an initial registration period, no further licenses were granted.54 The incrementalism of this proposal may appeal to governments, as it would have less immediate impact than, for example, a ban on sales of tobacco products; however, for the same reason, achieving an end point would likely take decades.
Market/supply-focused endgame proposals
Licensing, outlet restrictions, display bans and price controls
Researchers have proposed a variety of restrictions at the retail level that could be employed in a broader endgame strategy, starting with licensing of tobacco retailers.56 Although existing licensing schemes are designed primarily to limit tobacco sales to adults, they could be designed to discourage them. For example, the number, location and opening hours of tobacco retailers could be limited, including prohibiting new outlets, barring outlets near schools or limiting sales to non-school hours, banning duty free sales or restricting all sales to government-controlled outlets (as is carried out with certain types of alcohol in some US states) or to one type of outlet.1 ,14 ,56–59 The cost of licenses, typically low, could be raised,58 along with the cost of violating licensure provisions (eg, the sanctions associated with underage or other illegal sales).56 Retailers could also be incentivised to give up tobacco licenses.58 Product display and point-of-sale advertising bans, already in force in numerous countries, could be a condition of licensing.56 Minimum prices could be set (already the case in at least 24 US states and the District of Columbia)60 to counteract manufacturer discounting.56 Research in New Zealand suggests that various tobacco outlet reduction strategies, including the elimination of 95% of current outlets, could help reduce smoking rates, but would not, on their own, achieve dramatic reductions in the near term.61 Opposition to such plans from the retail sector would be vigorous in the absence of appropriate incentives. However, considerable evidence suggests that ubiquitous availability is a factor in smoking initiation and relapse after quit attempts.62–67
The commercial sale of cigarettes (or all combustible tobacco) could be prohibited.23 ,68 ,69 The ban could be announced well in advance of implementation, giving smokers time to quit.24 Pre-existing bans on smokeless tobacco or other nicotine product alternatives could be lifted to offer smokers cleaner nicotine alternatives.70 Opponents of a sales ban point to the failure of alcohol prohibition in the USA in the 1920s to suggest that a cigarette sales ban would be unpopular and ineffective.71 However, Proctor argues that tobacco and alcohol are different: nicotine, unlike alcohol, is not a recreational drug, and most smokers do so to satisfy their addiction rather than for pleasure.22, p 557 There is also precedent for governments to ban sales of products that are exceptionally harmful when used as intended. The majority of smokers want to quit;72 ending sales could help them do so. Some might argue that this measure would create more hardship for less-educated and low-income people, among whom smoking is increasingly concentrated. However, communities could institute targeted cessation aid and provide other resources to offset impacts; easy access to deadly products cannot be regarded as a benefit.
Advantage cleaner nicotine products over combustibles
Combustible tobacco could be placed at a regulatory or market disadvantage compared to ‘cleaner’ nicotine products such as low nitrosamine smokeless tobacco, pharmaceutical nicotine and e-cigarettes.73 ,74 For example, combustible tobacco could be subject to higher taxes, restricted availability and enhanced warning labels. To eliminate negative outcomes associated with wider availability and use of non-combustible products (eg, youth uptake, increased or sustained nicotine addiction among smokers who might otherwise quit, undermining public smokefree laws, and remodelling smoking as a desirable activity), the marketing, design, distribution and use of such products could be regulated.74 In addition, e-cigarettes could be required to look less like cigarettes and their use prohibited in places where cigarette smoking is banned.74 Alternatively, tobacco companies could be permitted to market clean nicotine products, but only if they agreed to phase out manufacture and sale of combustible products.75 This latter approach would require new legislation in most jurisdictions, which, given the political resources of the tobacco industry, would be challenging to pass. In addition, there remains some concern that nicotine itself may have negative effects on health,76 although these could be far less harmful than the effects of using ‘dirty’ combustible tobacco.
A quota on tobacco manufacture and imports, to be regularly reduced under a ‘sinking lid’, is another approach.77 ,78 Manufacturers and importers would bid at government-run auctions for shares of the market. As quotas were reduced, prices for the shares and consequently prices for tobacco products, would rise, until demand shrank. Revenues from share auctions (as well as tax revenues) could be applied to complementary tobacco control programmes. The sinking lid explicitly lays out a timeline for the cessation of all tobacco product sales. This idea is based on similar schemes in use to reduce carbon emissions (ie, ‘cap and trade’),79 and to control the catch taken from fisheries.80 US Senator Mike Enzi (Republican, Wyoming) introduced legislation in 2007 to create a cap and trade system for tobacco products, but it failed to pass.81 The proposal has been critiqued as too complex and opponents have argued it would be easier to simply raise taxes.82
Under the tobacco price cap proposal, a tobacco regulatory body would set the maximum wholesale price for cigarettes,83 ,84 taking into account production costs and reasonable profit. The retail price would be the wholesale price, plus retailer-added costs, plus excise tax and sales tax/value added tax. Thus, tobacco manufacturers’ profits would be reduced, while the consumer's price could be established by the government through excise taxes. This system prevents the industry from using tax increases as an opportunity to increase its profit; gives the government more control over the consumer price, allowing it, for example, to increase the excise tax while maintaining a price ceiling; ensures that price increases benefit the government (via tax revenues) rather than the industry; and ultimately reduces industry influence by reducing its lobbying fund. Price caps would also reduce the price differentials among brands (actual production cost differences are likely to be minimal), thus reducing the tendency of smokers to shift to lower-priced brands rather than quit in the face of price increases. This system does not necessarily imply an end to commercial tobacco sales; however, it gives government the ability to ultimately cause profit margins to shrink to levels unacceptable to the industry. The price cap system has been used in the UK to regulate utility prices.85
Tobacco control agency
Many aspects of tobacco production, marketing and sales are weakly regulated, in contrast to the controls put on other potentially harmful consumer products, such as alcohol and pharmaceuticals. For example, there are no restrictions on cigarette ingredients, no ingredient label requirements and varied regulations about sale (including many jurisdictions where no retail license is required) and promotion. The tobacco industry operates under a ‘perverse incentive'86, p 463 whereby “the more people it addicts and kills, the more money it makes”. Owing to the unique qualities of tobacco, a new agency may be necessary to reverse the perverse incentive. Such an agency would manage products, marketing, development of less harmful/addictive products, price, sales and monitoring of the regulatory system. Thomson et al87 endorsed this model for New Zealand, suggesting that it be funded by taxes on tobacco companies.
Regulated market model
Borland88 suggested a regulated market model (RMM) under which an agency would be both regulator and sole purchaser of tobacco from manufacturers and importers. This agency could set standards for manufacturers (from whom it would buy) as well as for retailers (to whom it would supply products). This system could permit innovation (eg, the agency would buy demonstrably safer products) while controlling price, packaging and promotion. It could introduce plain packaging, ratchet down nicotine levels, raise prices or restrict outlets, to reduce tobacco use prevalence to near zero. Others have specifically suggested this proposal as a way to maintain the cigarette industry but compel production of a less harmful product, to be distributed in a better-controlled way.89 A system like the RMM was established to regulate sales of marijuana in Uruguay in 2013;90 however, implementation has been delayed.91
State takeover of tobacco companies
A similar scheme92 ,93 suggests that tobacco companies be purchased and managed by a not-for-profit entity with a health promotion mandate, which could then use multiple strategies to meet mandated tobacco use reduction goals. The ‘voluntary or legislated’ purchase ‘could be financed by industry assets and future revenue streams'.92, p 282 This entity would be akin to public water systems or state-run alcohol distribution systems.93 These parallels are not perfect, as they do not aim to eliminate usage of the relevant resource, whereas the goal here is ‘the phasing out of tobacco use or its reduction to levels of minimal use'.92, p 280
A simpler approach suggests that a public agency set goals for reductions in smoking prevalence that tobacco companies would be required to meet within a certain time frame, and measure whether those goals were met.94–96 Failing to meet a target would result in substantial fines. Tobacco companies, not the state, would decide how to proceed—higher prices, media campaigns, plain packaging, etc. In the USA, performance-based regulation to reduce youth smoking rates was included in proposed legislation to end multiple state lawsuits against tobacco companies, and in suggested remedies in a federal civil suit.95 In neither case were such regulations enacted.
Recently, a similar approach has been advocated in the UK, where a proposed ‘Tobacco Companies Obligation’ would legally require tobacco companies to pay a levy based on sales volume, which in turn would be managed by the Department of Health to fund tobacco control initiatives.97 Based on a ‘polluter pays’ principle, such an approach would provide a consistent source of resources for tobacco control efforts, similar to a dedicated tobacco tax.
Each of these more structural solutions could have great advantages in allowing increased control of the supply-side apparatus—if implemented well and supported by a government committed to the endgame goal. However, given the US experience with tobacco products regulation at the national level—weak, ineffective and subject to multiple legal challenges98—the feasibility of implementing such a plan in the near term may vary widely across countries.
Integrated endgame strategies
Countries likeliest to adopt any of the strategies discussed here already have created the needed context with extensive tobacco control programmes. For example, comprehensively implementing the WHO Framework Convention on Tobacco Control (FCTC) recommendations would create an environment conducive to a combination of the Tobacco Free Generation proposal and a phasing out of combustible cigarettes, while allowing a strongly regulated market in electronic nicotine delivery devices and/or low-nitrosamine smokeless products, such as Swedish snus.3 Other combinations and modifications of the proposals discussed here will likely be adopted in different jurisdictions.1 ,14 ,99–104 Table 1 summarises these integrated proposals, as well as other endgame proposals discussed here.
Public support for endgame proposals
There is limited research on public perceptions of endgame proposals. In 2004, Canadians were asked whether ‘governments should develop new ways to phase out smoking in 25 years'.93, p 139 Seventy-six per cent of smokers strongly or somewhat agreed. Since then, two endgame proposals, enacting a tobacco sales ban and reducing nicotine in cigarettes, have received the most research attention. Despite the absence of organised engagement of the public on these proposals, studies find some public support for both, with variations by smoking status, question wording and, in some cases, race or ethnicity and education level (tables 2 and 3). American public opinion on banning menthol cigarettes has also been explored;106 ,107 perhaps due to differences in question wording, one study found majority support for a ban,107 while the other found that the majority neither supported or opposed a ban.106
Research in New Zealand, one of a handful of countries with a deadline (2025) for becoming a smoke-free nation (smoking prevalence of 5% or less), has shown high levels of public approval of the smoke-free goal,108–110 with support higher when respondents fully understood the goal, highlighting the importance of clear communication by governments.111 Likewise, a majority of Singapore residents surveyed in 2007 backed the smoke-free generation proposal, although support was higher among non-smokers (72.7%) than smokers (60%).53
Critiques of endgame proposals
Most of these proposals have not been implemented, making it difficult to evaluate their practicality or legality. However, some overarching critiques of endgame thinking have emerged. There is concern that too much focus on novel approaches will diminish effort toward policies that have proven successful but may seem less exciting, such as cigarette tax increases.128 ,129 Focusing on novel approaches may also leave behind regions that have not achieved baseline successes (eg, reductions in smoking prevalence or establishment of smoke-free laws) that might make endgame plans acceptable to the public.4 Other critiques suggest that endgame proposals threaten fundamental values by empowering the state to take property from tobacco companies, or restrict the freedom of adults to purchase chosen products.82 Other hazards include inadvertently increasing lethality to tobacco users through product changes, and creating unenforceable regimes which reduce confidence in law and government.82
Any proposal which reduces supply, substantially changes the product or restricts access also raises the potential for smuggling and black markets. In Bhutan, the only nation to ban tobacco sales (in 2004), cigarette smuggling is reportedly ‘significant’.130 However, currently, the most problematic black markets rely on the tobacco industry for their product;131–133 thus, proposals that increase government authority over the industry or reduce or closely monitor production would likely have less potential to create such markets. In addition, although black markets are certainly a downside, endgame proponents should consider the likely size and specific consequences of their proposals in order to estimate whether they outweigh the potential benefits. The mere presence of a black market is not necessarily an argument against a particular policy.134 Indeed, in 2009, only 2.8% of Bhutanese used combustible tobacco products, men (4.2%) more so than women (1%).120
This paper has offered a synthesis of the ‘endgame’ literature to date. The idea of the ‘endgame’ as such is still emergent and perhaps most useful as an organising concept to push governments toward setting, widely publicising and engaging the public in efforts to achieve specific, date-linked goals to end the epidemic. Until recently, most tobacco control goals were modest and expressed solely in terms of a foreseeably endless process of reducing uptake and aiding cessation, with little sense of an identifiable end point. Tobacco industry rhetoric about the failures of the American attempt at alcohol prohibition may have muted such discussion, as many tobacco control proponents were reluctant to be identified as ‘prohibitionists’ or more recently, ‘nanny staters’.135 However, as novel policy approaches are advanced, there appears to be a growing recognition that ‘prohibition’ is not the only model. Recent articles on point of sale interventions,136 FCTC implementation137 and the US Surgeon General's 50th Anniversary report138 have been framed in terms of a tobacco endgame, suggesting that even in the absence of a unifying definition of endgame (or end point), the popularisation of endgame thinking has the potential to spur innovation. Advocates should be encouraged to explore new policy options and embrace the goal of a tobacco-free future.
Doing so requires leaders and governments willing to risk the political wrath of powerful tobacco industry interests. Encouragingly, more governments appear willing to do so, as evidenced by Australia's leadership in defending plain packaging for tobacco, Uruguay's strong tobacco control stance in the face of industry attacks, and leaders in Africa who supported public health in the face of tobacco industry threats.139–142 The FCTC has stimulated healthy ‘competition’ among countries and regions; its full and rapid implementation lays the groundwork for true endgame planning and allows a different vision of the future to emerge. Ideas discussed in this review, in combination with one another or with innovations not yet envisioned, should be considered as part of that planning.
What this paper adds
The tobacco control endgame discourse is relatively new; the term ‘endgame’ can include models first proposed before the term itself was widely adopted.
We identified and synthesised the literature on tobacco control endgames from multiple nations.
There is no single endgame solution, but endgame thinking opens up novel policy approaches that could be useful for jurisdictions around the world.
Contributors PAM retrieved and analysed articles, helped write the first draft of the paper and edited all subsequent drafts. EAS retrieved and analysed articles, helped write the first draft of the paper and edited all subsequent drafts. REM conceived the study, helped with the analysis, wrote parts of and edited all drafts of the paper.
Funding This work was supported by Cancer Research UK (no grant number assigned) and the Mary Harms/UCSF Nursing Alumni Endowed Chair Fund.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All of the data on which this study is based are publicly available.
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.