Background It is well-evidenced that environmental or human rights advocacy comes with risks for those involved. Much less is known about the risks of tobacco control advocacy despite the emphasis on tobacco industry conduct. This study explores the experiences and perceptions of intimidation among members of the tobacco control community in low- and middle-income countries (LMICs).
Methods We interviewed six experts representing each of the World Health Organization (WHO) regions to inform an online survey conducted among the LMIC-based tobacco control community. Thematic analysis was used for analysing qualitative data while quantitative data were analysed descriptively.
Results Twenty-three participants from five WHO regions completed the survey. Almost three-quarters of survey participants reported that they, or another member of the tobacco control community in their country, had experienced intimidation. The most frequently reported forms of intimidation were discreditation on social or traditional media, legal threats/action and threatening messages. Physical intimidation, theft/burglary, cyberattacks and surveillance were individually rare but reported collectively by over 40% of participants. Results suggested intimidation might be increasing and changing in nature, and undermines tobacco control efforts: it affects organisations’ capacity and agenda, their ability to build relationships with those needed to advance policy, as well as the well-being of individuals. Participants used a range of strategies in response to intimidation and reported that they would benefit from better support.
Conclusion This exploratory study suggests that intimidation is a crucial challenge for the tobacco control community. There are several measures the global tobacco control community could take, including establishing an international support mechanism and building capacity to deal with intimidation and draw attention to it.
- tobacco industry
- public policy
- low/middle income country
Data availability statement
No data are available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
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.
WHAT IS ALREADY KNOWN ON THIS TOPIC
The tobacco industry uses a wide range of strategies to oppose tobacco control policies and those in favour of them. Its efforts have included threats against policymakers and governments, including legal threats.
Less is known about industry attempts to intimidate members of the tobacco control community, including advocates and researchers.
WHAT THIS STUDY ADDS
This initial study on the topic suggests that intimidation is common, takes many forms and has significant detrimental effects on organisations’ and individuals’ ability to function effectively in tobacco control. It impacts well-being and has driven individuals out of tobacco control.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE AND/OR POLICY
The study identifies an urgent need to study this issue in more detail and find solutions.
It draws attention to measures the global tobacco control community could take to address intimidation and support those affected by it.
Environmental and human rights advocacy have been recognised as potentially dangerous undertakings, and cases of advocates who have lost their lives are well documented.1–5 Organisations such as Human Rights Watch6 and Amnesty International7 routinely expose cases of state repression against advocates, and threats and violence against them from powerful corporations.8 Less attention has been paid to risks faced by the tobacco control community despite evidence of the tobacco industry’s willingness to discredit and silence its opponents.9–13 A Philip Morris memorandum from 1981 suggested dealing with troublesome research findings by ‘attacking researchers themselves, where vulnerable’.14
Tobacco control researchers have exposed aggressive and covert industry surveillance of public health groups, and attacks directed against governments15–19 and prominent researchers.20–24 Peer-reviewed articles also mention instances in which advocates were attacked or ridiculed in the media or public statements,11 12 14 25 26 were subject to legal threats or action14 27 28 or received intimidating anonymous messages.29 Most of these studies focus on high-income countries. A recent study on industry interference in eight low- and middle-income countries (LMICs)30 reported examples where advocates were publicly discredited, received threatening messages, believed that they were being followed, experienced cyberattacks, and attempts at cancelling the funding of tobacco control organisations, showing that intimidation is a concern shared by advocates globally.
More evidence on the risks of tobacco control work is available in the media. Examples include the break-in at tobacco control organisations in Brussels31 and anonymous messages and calls received by tobacco control researchers from the University of Bath.32 A New York Times article reported instances of intimidation of tobacco control advocates in LMICs.33 It documented, among others, public discreditation and physical intimidation against advocates in Indonesia and an armed attack on an advocate’s home, leading to fatalities, in Nigeria.33 Moreover, intimidation of public health advocates working in other areas has been documented, arguably the most notable case being the use of sophisticated spyware against advocates for a sugar-sweetened beverage tax in Mexico.34
The WHO Framework Convention on Tobacco Control (FCTC)35 and its implementation guidelines36 only require parties to monitor the industry, not to provide support to individuals intimidated by the industry. Furthermore, none of the publicly available tobacco control toolkits provided by international tobacco control organisations are dedicated to or offer detailed advice regarding the risks associated with advocacy.
This paper seeks to enhance our understanding of the lived experiences, and perceptions, of intimidation in the LMIC-based tobacco control community. Focusing on the experiences and perceptions of members of the tobacco control community, it seeks to explore the forms and impact of (and changes over time in) intimidation, responses to intimidation and whether the tobacco control community needs to better address intimidation.
Given the lack of an agreed definition of the concept, our starting point was to consider intimidation as an umbrella term based on the dictionary definition ‘action[s] of frightening or threatening someone, usually in order to persuade them to do something that [one] want[s] them to do’,37 which allowed for both overt (eg, public discreditation in the media) and covert (eg, private messages) intimidations. However, we deliberately explored meanings of intimidation as part of the work. Members of the tobacco control community typically include advocates and researchers, but we allowed for a broader conception to also include, for example, journalists and civil servants by inviting participants to self-identify as part of the tobacco control community while also specifying their background.
It was beyond the scope of the project to capture intimidation in its entirety. Instead, the study aims to act as a starting point to provide initial insights which will help determine whether further research and attention should be directed to this topic.
Due to the limited literature and primary data on experiences of intimidation among members of the tobacco control community, we used a two-step approach to data collection using interviews to inform a survey.
Step 1 (expert interviews)
Sampling and recruitment
We purposely selected interview participants (IP) with several years of experience in tobacco control in multiple countries and at the regional or global level. To ensure that we covered all geographical areas, we sought to recruit one participant per WHO region. The participants needed to be fluent in English or Spanish so that the lead researcher could conduct the interviews. Interviewees were identified through our networks with snowball sampling used to identify further participants.
The interview schedule was informed by the existing literature11 14 20 21 25–30 and developed in a series of author meetings. It explored the participants’ understanding of the concept, their insights into forms of intimidation members of the tobacco control community face and how they respond to it, specific examples they were aware of, and their perception of the drivers of intimidation, its impacts and existing or potential support mechanisms. We piloted the guide with two experienced members of the tobacco control community. All interviews were conducted in November and December 2020, recorded and transcribed by the lead author.
Step 2 (survey)
Sampling and recruitment
The survey collected primary data from members of the LMIC-based tobacco control community. To participate, a member had to have internet access, read and write English, French or Spanish and self-identify as a member of the tobacco control community in an LMIC. People with tobacco industry links were not eligible.
We shared the surveys through our networks but avoided distributing to mailing lists with unknown recipients to avoid attracting the attention of actors outside the tobacco control community.
The questionnaire was based on the existing literature and the interview findings. The survey’s purpose, scope and content were discussed in a series of author meetings. It was piloted in English with four non-native speakers of English, including two researchers and two advocates, then translated into French and Spanish by professional translators. In light of both the interviews and piloting, our initial definition was broadened further as follows: ‘We understand intimidation very broadly as actions that make you feel frightened or threatened. It also includes unsuccessful attempts at intimidation’. Given the lack of prior research on the topic, we sought to gain an overview of participants’ experiences and perceptions. We included open and closed questions and ensured that participants could add unique answers in sections with closed questions. For example, participants were asked whether they identified as advocates, researchers, civil servants, etc, but also given the option to select ‘other’ and describe their role differently. The questionnaires are available in the online supplemental file 1.
We used Online Surveys (https://www.onlinesurveys.ac.uk/) to collect the survey data. The surveys were password protected and were open for six weeks between April and June 2021. The results were downloaded into a Microsoft Excel sheet.
Demographic data and closed questions were analysed descriptively. Thematic analysis38 39 was used for the open-ended survey questions and expert interviews. NVivo V.12 was used to facilitate the analysis. To ensure participants’ anonymity, direct quotes only indicate if they came from an IP or survey participant (SP) and, for the latter, whether they had self-identified as advocate, researcher, etc.
Given the sensitivity of the topic and that data collection had to be remote due to the COVID-19 pandemic, we chose a survey as the most appropriate method to collect data from individuals who may experience intimidation. The drawback was that the survey data were not as detailed as, for example, the interview data might have been.40 Not collecting personal data, which was a means of protecting participants’ identities, meant that we could not ask participants for further clarification or more information. The research team benefited from advice given by colleagues in all country income groups and WHO regions. However, due to concerns around data security and collaborator safety during a time when travel was not possible, the research team was limited to UK-based researchers. The team was however diverse with participants from three WHO regions.
There was no patient involvement.
Between November and December 2020 we conducted remote interviews with five tobacco control experts from five WHO regions. The duration ranged from 53 to 70 minutes. A participant from the sixth WHO region submitted a six-page written response. Four of the six interviewees were from LMICs and five had extensive experience working in LMICs.
Twenty-three participants completed the survey. The majority identified as advocates and researchers, had less than 10 years of experience in tobacco control, and all WHO regions bar the Western Pacific region were represented (see table 1 for sample characteristics).
Types of intimidating actions: perceived commonality and frequency
Almost three out of four SPs (17 out of 23; 74%) reported they or another member of the tobacco control community in their country had experienced some form of intimidation. Overt forms of intimidation aiming to discredit their target publicly—public discreditation via various routes—were reported and experienced more frequently than less public facing forms of intimidation (table 2). Legal threats and actions were also commonly reported. More covert forms of intimidation took numerous forms including threatening messages (also common) or cyberattacks, physical intimidation/violence, burglaries/theft, surveillance and formal complaints, which while individually uncommon were collectively reported by over two-fifths of participants (10 out of 23; 43%).
Changes over time
Two out of three participants (15 out of 23; 65%) covering all five regions reported that intimidation has changed over time. Of those, two-thirds (10 out of 15; 67%) covering all regions reported that intimidation has intensified and almost all (14 out of 15; 93%) found that the forms of intimidations have changed over time (see online supplemental file 2, graph 1). Participants attributed these changes to two processes. First, the emergence of newer nicotine and tobacco products which led to ‘growth of tobacco industry’s business’ (SP, advocate), and brought new actors into the arena. Second, an increase in channels of intimidation, including ‘paid journalists [and] media companies’ (SP, advocate). Most importantly, online platforms, especially social media, have become crucial spaces through which members of the tobacco control community are targeted. This shift was reported to have resulted in a higher frequency and wider reach of attacks, with attackers often remaining anonymous—‘online, there is no control of authorship or any responsibility’ (SP, advocate/researcher).
The responses also illustrate that experiences over time varied: while some participants found that industry attacks have become more subtle, one pointed out that there are now ‘more blatant threats against civil society organisations’ (SP, advocate). As an interviewee noted, ‘[the tobacco industry] chooses its intimidation tactics depending on what works effectively in local settings’ (IP, expert).
Perceived drivers of intimidation
Over a third of the SPs (8 out of 23; 35%) from four regions (African Region (AFR), South-East Asian Region (SEAR), European Region (EUR), Eastern Mediterranean Region) reported that intimidation happens because the tobacco industry was fearful of tobacco control: ‘adequate tobacco control policies mean less business for the tobacco industry’ (SP, advocate) and, therefore, the tobacco control community is ‘a thorn in the side of the tobacco industry’ (SP, researcher). Through intimidation, the industry would also try to prevent ‘new voices [to] come up in the future’ (SP, researcher). One interviewee who could not recall examples of intimidation noted that civil society had not been heavily engaged in tobacco control, suspecting that the industry might not perceive it as a threat.
The government’s role was also important for explaining intimidation. ‘The industry intimidates more advocates in LMICs because there is less support from the governments and fewer consequences to face’ (SP, advocate/researcher/civil servant). Interviewees also linked intimidation, especially its more dramatic forms like physical intimidation, and burglary, to contexts of weak institutions, poor governance and high levels of corruption, also implying that people in such contexts have less expectation of being protected by the state.
Knowing who is behind the attacks
Participants tended to be more confident about the originator of the attack where the intimidation was public facing: at least half of the participants reported being at least ‘quite sure’ that such attacks could be attributed to the tobacco industry (see online supplemental file 2, table 1). In some cases, evidence on who is the aggressor was available: in one instance, there was a legal case filed against a tobacco control organisation, in which the plaintiff was known. In another, a participant reported having evidence that money had been offered to journalists for publishing denigratory information about the tobacco control community. Some participants stated that obtaining evidence on the source of the attack was at times challenging although the content was revealing: ‘They use exactly the same narrative as those furthered by [tobacco company] against tobacco control advocates’ (SP, advocate).
Participants were less confident in identifying the aggressor where the intimidation was more covert, including anonymous messages, cyberattacks, physical intimidation and theft (see online supplemental file 2, table 1). Between 20% (physical intimidation/violence) and 45% (non-anonymous messages) of participants reported being at least ‘quite sure’ that the tobacco industry was behind the attack. Advocates considered multiple factors in attempting to understand the events: ‘the circumstances, the timing and the methods and the targeting were more what let us… to have the firm conviction that this was not just a random job’ (IP, expert). While some participants reported they would retain messages or letters, many participants reported difficulties in obtaining evidence on the origins of covert attacks.
Impact of intimidation
More than two in three participants (17 out of 23; 70%) covering all five regions reported that intimidation affected them and their work. The analysis identified three impacts. First, it affects the organisational level, ‘eating up time and resources’ (IP, expert).
I have to spend more time preparing for possible industry attacks when I am proposing an action… There is a huge quantity of energy spent on dismantling the allegations of the industry. (SP, advocate/researcher)
Second, intimidation impacts collaboration: ‘[i]t creates obstacles for building relationships with decisionmakers, media, non-governmental organisations, WHO and other international organizations, government, opinion leaders’ (SP, advocate); ‘Some legislators might hesitate to work with us given the suggestion that we are “foreign agents”’ (SP, advocate); ‘Some […] stakeholders fear being attacked by the [tobacco] industry if they partner with [tobacco control] organisations’ (SP, advocate).
Finally, it affects individuals, making them uncomfortable or fearful. ‘[It] sometimes scares me to go public with the information I may uncover’ (SP, advocate/researcher/civil servant). Also, an interviewee suggested that ‘when you receive a legal challenge or many people attack you on social media (…) it’s understandable if you maintain quiet because of it’ (IP, expert). Participants from three regions (AFR, SEAR, EUR) shared examples of people who had quit advocacy as a result of intimidation. One reported: ‘Many medical doctors and young people stopped because they were not expecting to receive accusations of unethical intentions’ (SP, advocate/researcher).
Over a third of the participants (8 out of 23; 35%) covering all regions said that they do not engage in certain activities or do not say certain things because they fear consequences. Talking about a country with very limited tobacco control progress, an interviewee shared, ‘I would personally want to have a strong voice and share what I’m sharing in other countries… But I can’t… I’m scared’ (IP, expert). An SP commented, ‘sometimes, you feel threatened and find yourself carefully selecting the words… or you feel the need to protect your family members from getting hit by the assaults indirectly’ (SP, advocate).
Where close links between policymakers and industry existed, advocates would try ‘to avoid getting into a direct conflict with some of these very influential people or institutions’ (IP, expert). They would instead focus on areas that were deemed safe, for example, raising awareness about the dangers of smoking. ‘They don’t mention names [of those with industry links]. This is a red border; they have to stop there’ (IP, expert). It is worth noting that in such context, the state is the feared oppressor.
Responding to intimidating actions
Participants reported a range of techniques they employ in response to attacks, most prominently ‘correcting’ industry false claims and exposing industry conduct through public statements or the media. One participant reported that when false information was shared through the media, they would respond by ‘sending official letters to the media with proofs of incorrect information’ (SP, advocate). Another participant mentioned that the standard response was to file a legal complaint.
Several participants mentioned that their organisation or funder had a non-engagement policy, either generally or for a specific area such as social media. Yet, a participant reported,
it’s difficult to not engage when you are being attacked personally in so many platforms at the same time. It scares you and it would be great to […] expose these situations so they stop happening. (SP, advocate/researcher)
Looking at whether their responses worked well, most participants found that they were at least partly successful (see online supplemental file 2, graph 2). Responses were found to create awareness among policymakers and the public. In addition, responding was effective in showing the industry that ‘advocates will not be scared easily’ (SP, advocate). A few participants reported direct effects when responding to intimidation (eg, an attacker backing off and a legal case being withdrawn).
Some participants mentioned responses to intimidation that were disruptive or had a negative impact on the targets: a participant shared that they had changed their contact details and became very concerned about digital security. Another participant reported that following an attack from a tobacco company, their boss asked them to stop working on that company. An important concern regarding responding to intimidation was that exposing industry conduct could raise the visibility of the industry which would not necessarily facilitate its discreditation.
What is needed to better handle intimidation?
While most participants covering all regions commented that there is awareness and evidence that intimidation happens, most saw room for improvement: ‘it’s not taken seriously (by the) tobacco control community’ (SP, researcher/journalist). Furthermore, an increased understanding that it is not a matter of individual cases was important: ‘[we need] more evidence on how this is a concerted effort’ (SP, advocate).
It would be helpful for everyone to know that they are part of a community facing similar challenges. It would reassure them, it would encourage them because when you then face intimidation, you feel you're not alone. (IP, expert)
Most participants saw the need for more and more timely support, both from the government and the international tobacco control community ‘to reach more people and fight back better’ (SP, advocate). Types of support included legal and financial support as well as psychological support for those affected by industry attacks. Government support was perceived as crucial and often lacking. When advocates filed complaints, the response was slow. Another concern was related to the government’s role in protecting and not constraining the freedom of speech of civil society organisations and individuals more generally.
Regional or international tobacco control organisations were also perceived as an important source of support. An interviewee explained these organisations could, and to some extent already do, facilitate responding to threats or discrediting efforts by ‘helping with legal support and, with letters to media, letters to authorities or even sometimes putting some media communication’ (IP, expert). While it was known that organisations provide help for governments receiving legal threats, less was known about existing support for civil society organisations or individuals: ‘because most of the intimidation is not really documented… I can’t say that there is that direct form of assistance’ (IP, expert).
SPs stated they would also benefit from more collective action, which was perceived as less risky than speaking up alone. Also, having another organisation expose industry conduct for the local tobacco control community would make it less frightening for local advocates.
Finally, to pre-empt attacks and improve the handling of future intimidation, creating and sustaining strong networks and building capacity were considered crucial. Networks should provide support when responding to concrete cases, and could facilitate learning, for example: ‘how to respond to various forms of intimidation through case studies from around the world’ (IP, expert). In addition, training initiatives could help build capacity in the tobacco control community, covering, among others, how to collect evidence of and expose intimidation. Yet, for this to work, ‘you need to build the trust first’ (IP, expert) and, hence, such sessions should not only cover the risks and dangers of tobacco control work. Furthermore, ‘there has to be some way forward and people have to go with heartening ideas’ (IP, expert).
To our knowledge, this is the first paper on the experiences and perceptions of intimidation among members of the LMIC-based tobacco control community. It suggests that intimidation is widespread and increasing, targets both organisations and individuals and significantly threatens tobacco control. Some participants suggested the industry was deliberately seeking to deter advocates and may have had some success given reports that some individuals had left the field as a result and others limited what they said or did. A further route to impact is that, by discrediting individuals and organisations, the relationship building needed for tobacco control becomes difficult. Participants suggested that intimidation was more egregious in contexts where institutions and governance were weak and corruption was high.
Intimidation was found to take several forms, some overt and some covert. Public discreditation via mainstream or social media and other public platforms, including by paying journalists, was common and appears to have intensified with the advent of social media. This is consistent with the grey literature33 and reflects broader social concerns around the ‘dark side’ of social media, including cyberbullying, trolling and fake news.41 Previous research has shown the tobacco industry’s increased use of social media and the growing online harassment faced by public health advocates and researchers more broadly.23 42–45 We also found that tobacco control advocates, researchers and organisations are confronted with legal threats or actions. This is consistent with documented evidence, for example, in the cases of Australia46 and Uruguay,15 that industry uses such techniques against governments or public officials to impede, weaken and undermine tobacco control policies, even when claims are unfounded and chances of winning the case are low.47
In addition to these public threats, participants experienced more covert actions. These included threatening messages, cyberattacks, false information being disseminated about them or their organisations and formal complaints. Arguably, more worryingly they also included physical intimidation or violence, burglary or theft, and surveillance which were collectively reported by five out of 23 SPs (22%). In line with the grey literature,33 it was often difficult to obtain evidence and know with certainty who the originator of such actions was and participants reported greater certainty in ascribing overt actions to the tobacco industry.
The needs participants identified to help deal with intimidation include more support from governments and the global tobacco control community when reacting to immediate threats, and learning from each other and capacity building. These overlap with needs identified in a previous study on countering industry interference more generally.48
The findings of the study should be considered preliminary. The survey sample is small and does not cover the Western Pacific region. The survey relies on the limited information provided by participants, in which recall and social desirability biases cannot be excluded. It is also possible that those who perceive intimidation as an issue felt more inclined to fill the survey, thus skewing the sample.49 Conversely, what might be considered intimidation by some could be perceived as normal industry conduct by others leading to possible under-reporting. The small sample size could, among others, link to language barriers (the survey was only available in English, French and Spanish and not the other official United Nations languages (Arabic, Chinese and Russian)), our cautious approach to survey distribution and fear among potential participants. We received informal feedback that some potential participants were too scared to share their experiences. Another potential explanation is the length of the survey—it took around 30 to 45 minutes. A shorter survey might have led to higher completion rates, but it would have yielded less rich data. The study was also undertaken during the COVID-19 pandemic when the public health community was overwhelmed. This limited our ability to conduct in-depth in-person interviews.
Although care must therefore be taken in generalising our findings, this paper nevertheless offers insight into the scale of intimidation, suggesting it is not limited to a few extreme cases but is widespread. It also identifies the diversity of experiences in both receiving and dealing with intimidation.
Given that few tobacco control advocates and researchers have opened up about experiences of intimidation before, we hope that this research facilitates dialogue and encourages members of the tobacco control community to share their experiences and needs. This study also invites the global tobacco control community to reflect on its approach to exposing and addressing intimidation, including the policies of funding bodies. It suggests that both proactive and reactive support is needed. This might include integrating this topic into curricula of capacity-building initiatives, creating new mechanisms for collecting data on and enabling support and action on intimidation. The role of media and social media must also be carefully scrutinised given the role it can play both in enacting and addressing intimidation.
A crucial step could be a decision in the WHO FCTC Conference of the Parties, calling governments to protect those advocating for tobacco control in their countries from industry threats and attacks and supporting those who experience intimidation. International organisations, both intergovernmental and non-governmental, could implement policies offering protection, including funded communications and legal support, to staff and member organisations in order to increase their ability to respond and deter intimidation.
This study has laid the foundation for further research on intimidation in the tobacco control community. Further in-depth work including detailed country or regional case studies is needed to explore intimidation and its contextual variation in greater detail; larger surveys might help give a clearer picture of scale.
This study suggests that intimidation is an important issue for tobacco control in LMICs which goes beyond the attacks on governments and policy makers and may be limiting tobacco control progress in LMICs.
Data availability statement
No data are available.
Patient consent for publication
This study involves human participants and ethical approval was obtained from the University of Bath’s Research Ethics Approval Committee for Health (REACH) (reference: EP 19/20070). Participants gave informed consent to participate in the study before taking part.
This article is dedicated to our friend, colleague and co-author Dr. Mateusz Zatoński (1987-2022). He is sorely missed. We wish to thank our colleagues within STOP and beyond who provided advice and all study participants for sharing their experiences. We would also like to thank Adam Bertscher for his comments on a draft of this paper and the anonymous reviewers for their helpful feedback.
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.
Contributors ABG gained project funding. BKM, MZ and ABG conceptualised the project and designed the interview questionnaire. BKM gained ethics approval and conducted the interviews. The survey tool was developed by BKM and MZ, all other authors provided feedback. BKM coded the data and prepared the first draft of the manuscript. MZ, RA, MC and ABG provided feedback and edits during the manuscript preparation. BKM is the acting guarantor.
Funding This project and the authors’ time were funded by Bloomberg Philanthropies Stopping Tobacco Organisations and Products project funding (www.bloomberg.org).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.