Article Text
PDF
Abstract
Introduction Reunion Island, a French overseas department, is located in the southern hemisphere, close to the Capricorn tropic. This island has a multicultural and multiethnic population of 860 000 inhabitants, a quarter of whom are at high risk of developing skin cancer due to intense ultraviolet radiation. Melanoma is responsible for the majority of skin cancer deaths. The early prevention of melanoma is based on sun protection in childhood, but French regulations are not adapted to the environmental conditions of this tropical region.
The main objective of our study is to evaluate the effectiveness of three sun protection programs conducted in Reunionese primary schools through a pupil knowledge questionnaire.
Methods and analysis PRESOLRE is an interventional, open-label, cluster-randomised controlled trial, in four parallel arms, that is being conducted throughout 2022–2023 on Reunion Island. The trial design assumes an escalation interventional effect using: first, a control arm without proposed intervention (arm 1); second, an arm whose classes are encouraged to use the validated educational programme ‘Living With the Sun’ (LWS) (arm 2); third, an arm whose classes are encouraged to use both ‘LWS’ combined with ‘Mission Soleil Réunion’s sun protection awareness programme (arm 3); fourth, an arm benefiting from an intervention similar to arm 3, combined with the distribution of hats, sunglasses and sun creams (arm 4). In all, 1780 pupils from 18 classes of 20 pupils, on average, will be included. Randomisation applies to the classes of pupils (so defined as clusters). The primary outcome is based on the proportion of correct answers to the knowledge questions after the awareness programme, compared between the four arms using a linear mixed model with random intercept.
Ethics and dissemination The study obtained ethics approval in 2022 (ID: 2022-A00350-43). Results will be published in peer-reviewed journals.
Trial registration number NCT05367180.
- Primary Prevention
- Child protection
- Dermatological tumours
- Health Education
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
Strengths and limitations of this study
This study focuses on primary prevention by targeting the period of lifetime that is the most critical and the most at risk for photo exposure: the childhood.
This is an interventional escalation study that involves multiple levels of primary prevention: strengthening individuals by health education methods, strengthening community by peer-pupils education methods (ambassador classes) and improving access to sun protection materials.
This study is helping to create an intra/extra hospital partnership by building bridges between the hospital environment, the voluntary sector and the education authority.
The main limitation is the difficulty for the project team to work and evaluate within the constraints of the school calendar.
Introduction
Background and rationale
Reunion Island is located in the southern hemisphere in a tropical zone, the South-West Indian Ocean region, close to the Tropic of Capricorn. This French overseas department has a multicultural and multiethnic population of 860 0001 characterised by a high level of social inequalities. It is estimated that a quarter of the population has phototypes I, II or III, representing a high-risk population exposed to very intense ultraviolet radiation (UV), which ranges from UV 6 to UV 13 on the coast and up to UV 20 on the mountains resulting in sunburn even at phototypes V. Although it represents less than 5% of all cutaneous malignancies, melanoma is responsible for the majority of skin cancer deaths.2 3 Near to Reunion Island, in Australia in 2018, the age-standardised incidence rate was 54 cases per 100 000 persons.4 However, multiple sun protection programmes, including SunSmart created in 1988, have stabilised melanoma incidence since 2013 and decreased it for certain age categories.5–7 In Reunion Island, the mean estimated crude incidence rate of invasive cutaneous malignant melanoma for skin phototypes ≤ III was 27.8/100 000 person-years (PY) in men and 23.4/100 000 PY in women.8
The main risk factors for melanoma are physical characteristics such as phototype, personal history and environmental factors related to UV exposure. Prevention of melanoma relies heavily on sun protection, especially in childhood.9 10 However, there is a lack of sun protection in primary school children.11 12
Reunion Island, as well as the other French Overseas Departments and Regions (DROM), comply with the regulations and administration of mainland France, but these are not adapted to the environmental conditions prevailing in these tropical regions. Thus, school programmes and education on photoprotection are taught very little in schools, even though an educational programme ‘Living With the Sun’ (LWS) is already available.13 In this context, the ‘Société Réunionnaise de Dermatologie’ created the association ‘Mission Soleil Réunion’ (MiSolRé) in 2017.14 Its few members carry out prevention actions/steps/activities in elementary schools throughout the region by raising awareness among headteachers, teachers, pupils and parents about solar risks.
It is essential to be able to evaluate any public health intervention implemented in order to ensure its effectiveness and its real impact on the health of the population in the short, medium and long term. Two pilot studies carried out in voluntarily participating schools in Reunion Island show encouraging results regarding the feasibility and effectiveness of solar protection programmes aimed at pupils.15–17 However, methodological limitations weaken the validity and extrapolation of the results. To improve both evaluation and generalisability of the results of local public health interventions, cluster randomised controlled trial design can be used. In particular, this methodological framework allows questions to be addressed on the relevant intervention level(s) and their combination for tackling social inequalities in health.18 For example, strengthening individuals by the educational programme LWS; strengthening communities by involving peer-pupils to inform their counterparts; improving access to prevention by offering photoprotection material to the pupils.
Objectives
The primary objective of this study is to evaluate the effectiveness of three awareness-raising programmes about the risks linked to sun exposure, delivered in Reunionese primary school classes, based on pupils’ knowledge of sun protection.
The secondary objectives are as follows: (1) to evaluate the effectiveness of three awareness programmes about the risks related to sun exposure, delivered in Reunionese primary school classes, on the attitudes and behaviours of pupils in terms of sun protection in a declarative and observational way; (2) to determine the phenotypic characteristics of the pupils associated with their level of knowledge and their behaviour in terms of solar protection before the awareness programme; (3) to compare the level of knowledge between different grade levels before and after the awareness-raising programme; (4) to compare the level of knowledge of pupils according to the education policy to which the school belongs (priority education zone/non-priority education zone) before and after the programme; (5) to assess the processes and feasibility of rolling out each programme in primary schools.
Trial design
This article describes the protocol of the PRESOLRE trial, which is a prospective, multisite, cluster-randomised, open-label, interventional study using four parallel arms. Randomisation applies to the classes of pupils so defined as clusters. This research protocol was built with reference to the 15 recommendations of the Ottawa Statement dedicated to clustered randomised controlled trials19 (see online supplemental additional file 1).
Supplemental material
Methods and analysis
Study setting
PRESOLRE is conducted in 20 primary schools of Reunion Island, from September 2022 to June 2023.
Eligibility criteria
The inclusion criteria for this study are as follows: pupils must be enrolled in a Reunion Island primary school selected at random for the study and whose headmaster agrees to participate in the sun protection programme. The primary school must have more than 10 classes in total and the pupil must be enrolled in a class from the first to the third cycle, that is, CE1 (comprising pupils of 7 years old on average), CE2 (8 years old on average), CM1 (9 years old on average) or CM2 (10 years old on average). Finally, consent must have been given by both holders of parental authority.
The exclusion criteria are as follows: pupils enrolled in a school that has benefited from the MiSolRé association’s awareness-raising programme in 2016/2017, 2017/2018, 2019/2020 and 2020/2021; a school that does not depend on the board of education (public school without contract); pupils enrolled in the first grade or a multi-grade class.
Interventions
Three sun protection programmes are compared.
The control arm is an arm without proposed intervention (arm 1).
The intervention in arm 2 is to encourage teachers to use the LWS. LWS is a multidisciplinary teaching guide on sun safety for teachers, created in 2006 by an astrophysicist. It offers teachers to run 10 workshops for primary school pupils over a period of 3 months.13 This simple intervention is based on health education techniques aimed at strengthening individuals.18
The intervention in arm 3 consists of carrying out the LWS programme by the teachers in the same way as in arm 2, combined with the MiSolRé awareness programme on solar protection. The MiSolRé awareness intervention consists of training a so-called ambassador class in solar protection by presenting them with a 30-minute slideshow. This ambassador class will then be responsible for disseminating this information to the other four classes in their school within the next 15 days. The tool used to transmit this information is free. Its use depends on the class and the teacher preferences. This composite intervention is based on health education techniques aimed at strengthening individuals and peer education aimed at strengthening communities.18
The intervention in arm 4 consists of the implementation of the LWS programme by the teachers according to the same modalities as in arms 2 and 3, combined with the MiSolRé sun protection awareness programme according to the same modalities as in arm 3, plus the distribution of sun caps, sunglasses and creams at the beginning of the school year. This composite intervention is based on health education techniques aimed at strengthening individuals and peer education aimed at strengthening the community and offers photoprotection material to the pupils to improve access to protection measures.18
In summation, the trial design assumes an escalation interventional effect from arm 1 to arm 4.
For ethical purposes, at the end of the study, participants who belong to arm 1 or 2 will benefit from the MiSolRé outreach programme, and participants who belong to arm 1, 2 and 3 will be given photoprotection material (caps, sunglasses and sunscreens).
Study duration
The duration of the inclusion period is 7 months and the duration of follow-up per participant is also 7 months. The total duration of the study is therefore 14 months. The recruitment for the qualitative part of the study is still in progress.
Primary outcome measures
The primary outcome measure is a comparison of the percentage of correct answers to knowledge questions among pupils, measured via a self-questionnaire, after the outreach programme, between the four arms.
The knowledge questionnaire used is a self-questionnaire consisting of 45 items, divided into 13 parts (see online supplemental additional file 2). Each child answers true or false or yes or no to each question. The rate of correct answers will be calculated. This questionnaire has been used in previous publications by a team from Montpellier and was designed by experts in the field of sun protection. This questionnaire has good internal validity20 and has been pretested with Reunionese children before. However, we have adapted it in form by simplifying some of the wording for better understanding by the pupils.
We have the authorisation to use this questionnaire and we will be able to compare our results with the literature.
Secondary outcomes measures
Comparison of change in pupils’ attitudes and behaviour: (a) measured declaratively via a questionnaire and (b) measured by observation (ie, the ratio of the number of pupils wearing a cap in the schoolyard to the total number of pupils present) before, during and after the outreach programme, between the four arms. The behavioural questionnaire is a declarative self-administered questionnaire consisting of 40 items (see online supplemental additional file 3). These are single-choice questions. This questionnaire has been used in previous publications by a team from Montpellier and was designed by experts in the field of sun protection.20 This questionnaire has good internal validity.20 We also have adapted it in form by simplifying some of the wording for better understanding by the pupils. Photographs of the playground are taken before, during and after the study to observe the behaviour of pupils, including the wearing of caps and hats.
Phenotypic characteristics, according to the classification of phototypes ranging from I to VI (Fitzpatrick), associated with (a) knowledge level and (b) behaviour, measured via a self-questionnaire, in terms of sun protection, before the awareness programme. A visual aid with diagrams corresponding to the different phototypes is available for the pupils to identify their correct phototype themselves.
Comparison of the percentage of correct answers to knowledge questions among pupils, measured via a self-questionnaire, before and after the awareness-raising programme, between the different grade levels.
Comparison of the percentage of correct answers to knowledge questions among pupils, measured via a self-questionnaire, before and after the awareness-raising programme, according to the priority education policy in which the child’s school is located (priority education zone or not).
Description of the efforts made by different stakeholders to mobilise resources, carry out solar protection activities and reach out to the pupil audience, through data collection, observation and qualitative interviews. The fourth visit (T3) consists of assessing the processes by consulting data archived in the minutes of monitoring meetings and conducting qualitative interviews. They will be conducted by a trained interviewer with a minimum of two representatives of each trade involved in the study: school director, referent teacher, field coordinator of the study, school welfare assistant, parent, pupil and associative actor (MiSolRé). They will be invited to express feelings about the study and organisation of the programmes. Depending on their point of view, they will also be invited to describe the elements that have worked and have not worked well in the course of the programmes, to bring avenues for improvement in the organisation. A questionnaire will also be distributed to teachers in arms 3 and 4 to measure compliance with the LWS programme (number of sessions completed during the year).
Participant timeline
Figure 1 and table 1 show the course of the study.
Course of the study. CPP, Personal Protection Committee; CRF, case report form; LWS, Living With the Sun; MiSolRé, Mission Soleil Réunion; QR, questionnaire.
Schedule of enrolment, interventions and assessments of the study
Information about the study was given to school headmasters in June 2022, who were contacted again at the beginning of the school year, in August 2022, to pass information on to the teachers concerned. The inclusion of pupils, with the completion of knowledge and behaviour questionnaires and the observation of playgrounds at time T0, took place from September 2022 to February 2023. In March 2023, the second school playground observations took place, with the taking of photographs (T1). The completion of the questionnaires after the various sun protection programmes (T2) took place from March to June 2023. The sun protection programmes are carried out between the time of T0 and the time of T2 in each school.
Sample size
According to a previous study, the percentage of correct answers to the knowledge questionnaire used before the awareness programme was 60 points out of 100.20 We assume that the expected increase in the percentage of correct answers to knowledge questions after the awareness programme will be 10, 15 and 20 points in the arm 2, 3 and 4, respectively.20 Furthermore, we assume that the intracluster correlation coefficient is constant and takes the value 0.10.20 As for cluster randomised trial, we can calculate a design effect.21 Thus, with an average of 21 pupils per class, the design effect would be 3. Under these assumptions, the inclusion of 448 pupils multiplied by the design effect equals 1344 pupils in total (16 classes per arm, that is, 64 classes in total). This will show a minimum detectable difference of 10 points on the percentage of correct answers to the knowledge questions, respectively, in the intervention arms, compared with the reference arm, at an overall alpha risk of 5% and with a power of 90%. This calculation was done taking into account the multiple comparisons to be made (Dunnett test, see online supplemental additional file 4). In order to allow for possible refusals and non-response (estimated at 25%), we included 20 classes per arm (80 classes in total, that is, a maximum of about 1780 pupils, factoring in the real number of pupils in each class).
Recruitment
The MiSolRé association is approved by Reunion Island’s board of education and has already obtained an agreement for this study protocol. The links between the association, the investigators of this study (who are members of this association) and the schools are high since they have already carried out interventions in previous years in voluntarily participating school.
According to the list provided by the Reunion Island government office, a total of 73 632 pupils were enrolled in 3446 classes in 364 schools in 2018–2019. After excluding schools with less than 10 classes, classes that had already received training from the MISOLRE association on solar protection in previous years and classes with mixed levels and first-year classes, there were a total of 59 041 pupils in 2791 classes in 315 schools remaining.
Due to the growing notoriety of the association and its already pre-established network, it was feasible to include enough classes to obtain the necessary number of subjects. By respecting the school calendar and including classes from the beginning of the school year, the number of lost follow-ups should be limited.
After obtaining the agreement of the Personal Protection Committee (CPP), which is a French ethics committee, an initial agreement in principle was sought from all head teachers with more than 10 classes in their school to ensure that they could be drawn. Before the draw, the board of education sent a letter of information about the study to all schools in Reunion Island. Once the draw had been made, two field coordinators and medical interns contacted the school headmasters to present the study and obtain their permission to intervene in their schools.
Then, an information letter was sent to each headmaster whose school was drawn to explain in more detail the objectives of the research and its modalities. For arms 2, 3 and 4, an email was sent to each headmaster at the beginning of the school year to remind them of the LWS programme.
Assignment of interventions
The random draw was performed by the statistician via the SAS software, just after the information stage for headteachers. Indeed, for this study, it is necessary to obtain the list of schools drawn at random and to know the randomisation arm of the clusters (classes) from the start.
The allocation ratio was 1:1:1:1 and the randomisation was stratified by grade level. In order to minimise the imbalance between the different schools, a constrained randomisation was proposed (proc plan; SAS Institute, Cary, North Carolina, USA). This constrained randomisation took into account two covariates: the geographical area and the level of priority education zone of the schools (schools with more social difficulties). All pupils enrolled in the randomly selected classes were invited to participate in the research.
Data collection methods
A knowledge and behaviour questionnaire was distributed to pupils at the beginning of the study (T0) and the end of the school year (T2) in the same manner.
Responses to the questionnaires were collected on an electronic case report form (e-CRF) available on tablets. Each pupil had their own tablet. Paper versions of CRF were also available in case of tablet failure or lack of internet access in the school.
Pupils’ behaviour is assessed by observation through photographs of the recreation yard taken at T0, T1 and T2 by the field coordinators. The visit is unannounced so that the pupils are not warned. The pupils’ faces will not be made visible. The photographs are taken using the same procedure with a reproducible method and a similar angle of view. The ratio of the number of pupils wearing caps in the schoolyard to the total number of pupils in the photograph is recorded on an observation grid.
Trial participants, assessors and data analysts are not blinded.
Data management
The e-CRF is carried out by the data manager of the Methodology and Data Management Centre using Ennov Clinical software. Access to the database is secured (access by personal login and password) with different levels of security depending on the roles assigned. Inclusion data are entered into the e-CRF (Ennov Clinical) by the field coordinators and interns. Questionnaire data are completed in the e-CRF by the pupils or by the field coordinators and the interns if the pupils have completed the questionnaires on paper (second-grade class with school welfare assistant or no internet). Data are saved daily. Once entered, the data will be checked for consistency. Errors, omissions or inconsistencies will be reported in the CS Test module of Ennov Clinical.
Statistical methods
The statistical method used for the primary outcome is based on a linear mixed model with random intercept. The autocorrelation of individual observations within the class of pupils will be modelled by random intercept. The ‘randomisation arm’ factor will be coded by three dummies (one per intervention arm; none for the control arm). A first analysis (not adjusted for covariates other than the ‘randomisation arm’ factor) will consist of a global test of the equality of averages in the proportion of correct answers between the four interest arms, by a test of the overall effect of the factor. Subsequently, if at least one of the four means differs significantly from the others, so each intervention arm will be compared with the reference arm, in a post hoc analysis, according to the Dunnett method.
A secondary analysis will consist of adjusting the previous model using covariates collected at T0. The aim is to control for any imbalance between the arms of pretest values (baseline knowledge) and other characteristics predictive of knowledge (such as grade level) or cohort attrition (such as socioeconomic status of the school or pupils), and thus improve statistical power.22 23
These analyses will be based on the principle of intention-to-treat, that is, according to the arm assigned by the randomisation of the classes and independently of the deviations from the research protocol.
Management of missing data from the knowledge questionnaire: the proportion of correct answers will be calculated from the completed items that are not missing. In other words, missing responses will be excluded from this calculation. No multiple imputation of missing data is planned.
Regarding the secondary outcomes:
The analysis of pupils’ behaviours, collected in a declarative way, will use a non-linear mixed model for binary dependent variables with random intercept. The comparison between the four arms of the ratio, calculated from the number of pupils with a cap in the yard and the total number of pupils, will be done using a Kruskal-Wallis test.
A random intercept logistic regression model will be used to find out which of the pupil’s phenotypic characteristics measured at T0 are associated with the individual’s level of knowledge and individual behaviour in terms of solar protection before the awareness programme.
The comparison of the level of knowledge (% of correct answers) between the different grade levels (CE1/CE2/CM1/CM2) will be carried out using a linear mixed model with random intercept and will be stratified over the assessment time (T0/T2).
The same statistical analysis will be carried out to compare the level of knowledge (% correct answers) according to the education policy to which the school belongs (priority educational school/non-priority educational school).
Data monitoring
Monitoring is carried out by a clinical research associate (CRA) mandated by the study sponsor, through several sessions spread out over a fortnight. During these visits, the CRA verify all consent forms, regulatory documents and monitoring of good clinical practice. They also write a monitoring report after each visit. There is not an independent body monitoring safety. The consents collected by the field coordinators are stored in a secure locked cupboard (with respect for the anonymity of the consents in closed envelopes), which is located in a secure locked room within the CHU of La Réunion under the responsibility of the investigation team.
Access to the database is secured (access by personal ID and password) with different levels of security depending on the roles assigned. The data hosting is certified ISO 9001:2015. Ennov hosts its customers’ applications in a data centre located in two mirror sites offering optimal security conditions. Ennov Clinical complies with Food and Drug Administration recommendations regarding Computerised Systems for Clinical Trials Management (‘guidance for computerised systems used in clinical trials’) and electronic signature and international standards.
Harms
Undesirable events must be reported to the various health vigilance circuits in accordance with the regulations in force. However, there are no identified risks associated with the questionnaires and the constraints are minimal.
Auditing
A person mandated by the sponsor may conduct an audit at any time during the study, independently of the investigating team.
Patient and public involvement
Patients or the public were not involved in the design, or conduct, or reporting, or dissemination plans of our research.
Ethics and dissemination
Ethics approval was obtained from the Institutional Ethics Committee CPP Ile de France II on 2 May 2022 (ID RCB: 2022-A00350-43). The study was registered with the US National Institutes of Health (ClinicalTrials.gov) under ID: NCT05367180. All methods were carried out in accordance with relevant guidelines and regulations in the Declaration of Helsinki.
The informed consent of both parents or that of the child’s legal representative will be obtained prior to inclusion by the field coordinators. No intervention or data collection was initiated before consent was obtained.
Results will be published in peer-reviewed journals as well as presented and disseminated at conferences.
Protocol amendments and confidentiality
Protocol amendments
A Steering Committee will decide on any relevant amendments to the protocol necessary for the continuation of the research. Any substantial modification, that is, any modification likely to have a significant impact on the protection of persons, conditions of validity, the results of the research or the methods for conducting the research, will be the subject of a written amendment which will be submitted to the sponsor for approval, which must be obtained prior to its implementation, including an assent from the Humans Ethics Committee.
Three amendments to the protocol were made. In May 2022, a first amendment was made to modify an inclusion criterion, add a non-inclusion criterion and adjust the study population. A second amendment was made in September 2022 to change the number of participants included. A final amendment was made in November 2022 to extend the inclusion period and adjust the number of schools and classes to be included.
Confidentiality
In accordance with the legislative provisions in force, persons with direct access to the source data shall take all necessary precautions to ensure the confidentiality of information relating to the investigational medicinal products, to the research, to the persons involved and in particular to their identity and the results obtained. These persons, as well as the investigators themselves, are subject to professional secrecy.
During the research or at its end, the data collected on individuals by the investigators (or any other specialist) will be made anonymous. Each pupil will have an inclusion number assigned according to the following modalities: school number, class number and inclusion number of the child in the class.
Discussion
Sun protection is an important issue in regions with high daily sun exposure. Furthermore, in conjunction with global warming, the intensity and duration of annual ultraviolet exposure are increasing in many parts of the world.24 This study will provide short-term benefits for students by reducing the number of sunburns, which remains a major risk factor for the development of skin cancer. More generally, the results of this study will help validate the best strategy (ies) to employ for sun protection in schools. The optimisation of this strategy will enable the awareness campaign to be extended to all schools in Reunion Island, as well as other DROMs. These results will make it possible to promote the messages delivered by teachers, to raise the awareness of risks within pupils’ family circle and to contribute to the development of the solar protection policy in Reunion Island and France. Finally, in the long term, we may hope to observe a decrease in the incidence of cancers and pathologies induced by UV. In addition to the long-term effects study, future research projects could investigate health benefits within the pupil’s collaterals (eg, siblings, parents, friends, neighbours) using social network methods.25
Ethics statements
Patient consent for publication
Acknowledgments
This project would not be possible without the kind support and assistance of many individuals, and we would like to extend our profound gratitude to each and every one of them. We thank Emilie Techer, Camille Talandier, Cathy Gonthier Charveron, Nadège Naty, Céline Dumez, school directors, school teachers, school pupils, board of education of Reunion Island, Dr Isabelle Rios, Pascal Geslin, Jean-François Salles and Emilie Mourouveya-Mouraman. We would like to thank our copy editor Jennifer Sanders.
References
Supplementary materials
Supplementary Data
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.
Footnotes
AF and CF contributed equally.
Collaborators not applicable.
Contributors All authors made a significant contribution to the work reported, whether that is in the conception and study design. CD, LB, AF and AB drafted the initial manuscript. JV, KB, FD, CF, JD and NS-B critically reviewed the manuscript. All authors read and approved the final manuscript.
Funding This research received a specific grant ‘DEPREV 2021’ from the Institut National Du Cancer (INCa) /16313.
Disclaimer The sponsor had no involvement in study design; collection, management, analysis and interpretation of data; writing of the report and the decision to submit the report for publication. The coordinating centre was provided by the project manager, a steering committee (methodologist, biostatistician and clinician) and associative staff.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
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.