Article Text
PDF
Abstract
Purpose The Lifelines COVID-19 cohort was set up to assess the psychological and societal impacts of the COVID-19 pandemic and investigate potential risk factors for COVID-19 within the Lifelines prospective population cohort.
Participants Participants were recruited from the 140 000 eligible participants of Lifelines and the Lifelines NEXT birth cohort, who are all residents of the three northern provinces of the Netherlands. Participants filled out detailed questionnaires about their physical and mental health and experiences on a weekly basis starting in late March 2020, and the cohort consists of everyone who filled in at least one questionnaire in the first 8 weeks of the project.
Findings to date >71 000 unique participants responded to the questionnaires at least once during the first 8 weeks, with >22 000 participants responding to seven questionnaires. Compiled questionnaire results are continuously updated and shared with the public through the Corona Barometer website. Early results included a clear signal that younger people living alone were experiencing greater levels of loneliness due to lockdown, and subsequent results showed the easing of anxiety as lockdown was eased in June 2020.
Future plans Questionnaires were sent on a (bi)weekly basis starting in March 2020 and on a monthly basis starting July 2020, with plans for new questionnaire rounds to continue through 2020 and early 2021. Questionnaire frequency can be increased again for subsequent waves of infections. Cohort data will be used to address how the COVID-19 pandemic developed in the northern provinces of the Netherlands, which environmental and genetic risk factors predict disease susceptibility and severity and the psychological and societal impacts of the crisis. Cohort data are linked to the extensive health, lifestyle and sociodemographic data held for these participants by Lifelines, a 30-year project that started in 2006, and to data about participants held in national databases.
- COVID-19
- public health
- epidemiology
- genetics
- mental health
This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/licenses/by/4.0/.
Statistics from Altmetric.com
Strengths and limitations of this study
The Lifelines COVID-19 cohort collects data about factors relevant to the impact of the COVID-19 pandemic for >70 000 individuals living in the Northern Netherlands. Participants in the cohort are also participants in the Lifelines prospective population cohort and the Lifelines NEXT mother–baby cohort, which means there is already a rich data background for all participants, and cohort data can be linked to data held in other national databases.
The cohort questionnaire programme began during the height of the first wave of infections in the Netherlands and continued through late 2020 and early 2021. The questionnaires were designed by a multidisciplinary group of researchers to explore factors that may impact COVID-19 susceptibility and severity and the social, mental and economic impacts of the pandemic.
The compiled questionnaire data have been continuously shared with the participants and the public through an interactive website, and new questionnaire modules have been added to address questions raised by participants and researchers.
The region covered by the Lifelines COVID-19 cohort has so far experienced very low numbers of COVID-19 cases, as compared with the rest of the Netherlands. This reduces the power to detect COVID-19 related risk factors but makes the cohort an interesting resource for examining the broader mental health impacts of the governmental measures to slow infection rates and the associated economic slowdown.
Introduction
COVID-19 has now impacted the lives and health of billions of people around the world. Due to the initial absence of a vaccine, lack of effective antiviral medication and limited understanding of the SARS-CoV-2, most governments have tried to slow the growth rate of infections through public health measures including tracking and testing, shutting down of public life, social distancing policies and stay-at-home orders. These measures have had a huge impact on public health and well-being, the economy (including employment and working conditions) and daily life. The effects of the COVID-19 pandemic will therefore be multiple: there will be the impact of the infection itself and the broader societal and health impacts.
To identify genetic and environmental risk factors for COVID-19 and address the medical, social and psychological impacts of the pandemic, a multidisciplinary group of researchers rapidly developed and implemented an extensive COVID-19 questionnaire, leading to the development of the Lifelines COVID-19 cohort. The questionnaire collects data about COVID-19 related symptoms, current health issues and societal impacts from participants recruited from the Lifelines population cohort1 and the Lifelines NEXT (LLNEXT) birth cohort,2 which are both monitoring the health of the northern Dutch population (provinces of Drenthe, Groningen and Friesland). Via a (bi)weekly questionnaire, the project gathers information about COVID-19 symptoms, associated comorbidities and environmental factors, changes in work and employment, COVID-19 related worries, loneliness and the mental health and societal impacts of the pandemic. In addition, all participating parents are asked about their children’s well-being, and LLNEXT parents received detailed questions about COVID-19 related symptoms expressed by their children. Additional questionnaire modules have been included as the project progressed.
The data collected by the questionnaires is being used to address four aspects of the outbreak: (1) how the COVID-19 pandemic developed in the three northern provinces of the Netherlands, (2) which environmental risk factors predict disease susceptibility and severity, (3) which genetic risk factors predict disease susceptibility and severity and (4) the psychological and societal impacts of the crisis.
The initial COVID-19 outbreak in the Netherlands and the northern provinces
The first official COVID-19 cases in the Netherlands were registered on 27 February 2020 (see timeline figure 1A).3 By 24 March 2020, the number of cases diagnosed per day had risen to 1126 (figure 1B). The rapid rise in case numbers led the Dutch government to shut down primary and secondary schools, bars and restaurants, sporting facilities and other public spaces on 15 March 2020, followed by a more extensive shutdown of public life in the weeks that followed (see figure 1A for major events). However, the three northern Dutch provinces did not follow national trends. COVID-19 appeared later here and did not reach the same incidence of cases or infection rates. While the three northern provinces account for 10% of the Dutch population, they only had 2%–3% of cases, hospitalisations and deaths in the period from 27 February to 9 June 2020 (see figure 1C; online supplemental table 1). Multiple factors may explain why the outbreak was different in the north.4 5 Drenthe, Groningen and Friesland together are the least populated region of the Netherlands and contain the fewest urban centres. The early arrival and spread of infection in the southern Dutch province of North Brabant seems to have originated in travel to and from Northern Italy during the school holidays from 22 February–1 March 2020, with the spread of the infection in the southern Dutch region then further facilitated by personal contact during regional carnival celebrations. In contrast, school holidays fell earlier for the northern provinces (15–22 February 2020), which suggests that northerners who travelled to Italy during the vacation had returned before the major expansion of the outbreak in Northern Italy.6 Nor is carnival widely or generally celebrated in the northern provinces. Other factors may also have played a role, for example, use of a stricter COVID-19 testing regime and wider availability of tests in the region.5 The later arrival of COVID-19 to the north meant that the national steps taken to bring down the infection rate were in place before the outbreak had really taken hold in the Lifelines region.
Supplemental material
Timeline of the COVID-19 pandemic in the Netherlands and Lifelines data collection. (A) Important events of the pandemic in the Netherlands from February to June 2020. (B) Daily reported positive infections (grey) and hospitalisations (blue) visualised alongside the change in mobility (black) in the Netherlands. Mobility is quantified using Apple Maps Request data (https://www.apple.com/COVID-19/mobility) with the change over time normalised to 1 February 2020. Change in mobility indicates the percentage change in overall requested driving directions by users of Apple Maps. COVID-19 daily infections and hospitalisations are derived from the CoronaWatchNL github account (https://github.com/J535D165/CoronaWatchNL) and are based on reported numbers from the Rijksinstituut voor Volksgezondheid en Milieu (RIVM). (C) The reproductive number in the Netherlands and the three northern provinces over time. The R(t) is calculated based on incident cases (new positive PCR tests) including healthcare workers and cases appertaining to local outbreaks. National and regional R(t) values in the early phase of the pandemic are not directly comparable, since testing among healthcare workers was more widely adopted early on in the northern provinces. (D) Overview of the Lifelines COVID-19 data collections. The pie chart on the left shows the proportion of participants for each province. The first weekly COVID-19 questionnaire (Q1) was sent out on 30 March 2020. Based on Q1–7, 71 800 unique respondents have filled out at least one questionnaire. From Q7, assessments are biweekly. IC, intensive care.
Cohort description
Participant recruitment
Participants of the Lifelines COVID-19 cohort are recruited from the Lifelines and LLNEXT cohorts. Lifelines is a prospective population cohort following ~167 000 people in the three northern provinces of the Netherlands for 30 years1 (see figure 2 for Lifelines region). Established in 2006, Lifelines collects detailed information about its participants via extensive questionnaires and medical examinations, and the cohort has been shown to be representative of the Northern Dutch population.7 By design, Lifelines recruited multiple participants within families to produce a multigenerational cohort that could map individual and community health across life course.1 Since 2016, LLNEXT has been recruiting an additional generation through inclusion of mother–baby pairs, with partners also invited to participate to generate parent–baby trios.2
Distribution of hospitalisation across Dutch municipalities. The number of hospitalisations per municipality, as reported by the Rijksinstituut voor Volksgezondheid en Milieu (RIVM), were integrated with a geographical map of the Netherlands. For each municipality, the cumulative number of COVID-19 hospitalisations per 100 000 residents is shown. The lifelines region is outlined in grey. These data were downloaded on 18 May 2020.
To recruit participants for the Lifelines COVID-19 cohort, Lifelines and LLNEXT invited their participants digitally to fill out the questionnaires. In each questionnaire round (Q1–Q7 in figures 1D and 3), all Lifelines participants over the age of 18 years with a known email address received a link to the digital COVID-19 questionnaire (see inclusion flow chart in figure 3). Digital invitations were valid for 3 weeks, and the date on which the questionnaire was completed was registered. Since LLNEXT is an ongoing project in which new participants are still being included, the number of LLNEXT participants invited to participate in the COVID-19 cohort increased with each new questionnaire round. In each round, invited participants chose if they wanted to fill in the questionnaire, and the cohort population consists of all those who filled in at least one questionnaire in the first seven questionnaire rounds (figures 1D and 3).
Study inclusion flow chart.
On 30 March 2020, all Lifelines and LLNEXT participants were invited to participate in the first COVID-19 questionnaire round (Q1 in figures 1D and 3), with new invitations to participate sent out weekly. Starting 27 April 2020, an additional questionnaire about children’s health and symptoms was sent to the participants of the LLNEXT (>300 participants). Programme questionnaires were sent out weekly through the week of 18 May 2020, then at biweekly intervals until July 2020, when the questionnaires became monthly. The project has continued through 2020 and early 2021, with the option to increase questionnaire frequency when the local caseload begins to increase rapidly. While the timeline of the questionnaire programme will be decided by the outbreak, the cohort itself will continue to exist as part of Lifelines, which will allow participants to be monitored for the long-term health impacts of the pandemic through the length of the Lifelines programme.
Questionnaire contents
The Lifelines COVID-19 questionnaire includes question modules about sociodemographic parameters, chronic diseases, COVID-19 infection, general health and symptoms, medication use, the mental health/well-being of participants and of children and young adults in their family, COVID-19 related well-being, social life, social relations and lifestyle (see table 1 for modules and questions and https://www.lifelines.nl/researcher/data-and-biobank/wiki). For participants who answer a subsequent version of the questionnaire, these questions are related to their experience in the period since the previous questionnaire. Additional questions and question modules have been added as the questionnaire programme progressed, including the Groningen Frailty Indicator,8 KIDSCREEN-109 and the Positive and Negative Affects Schedule.10
Lifelines COVID-19 questionnaire
Lifelines COVID-19 cohort data can be linked to other participant data held by Lifelines, including detailed biological measurements such as genotype, metagenomics, metabolomics and transcriptomics collected for subcohorts within Lifelines such as Lifelines DEEP11 and Lifelines DAG3. Data can also be linked to the administrative records held by Statistics Netherlands (https://www.cbs.nl/en-gb), which include health-related records on mortality, hospital admissions and healthcare costs, as well as data on employment status, income, wealth and other sociodemographic characteristics. Data can also be linked to drug prescription data held by IADB.nl via the Pharmlines initiative12 and to SARS-CoV-2 testing data (including serological data) held by Certe and other Dutch laboratories.
Participant and public involvement
Projects carried out within Lifelines are discussed with the Lifelines Participant Advisory Board. With respect to the COVID-19 cohort, compiled questionnaire results have been continuously updated after each questionnaire round and shared with participants and the public through interactive infographics on the Corona Barometer website (https://coronabarometer.nl/, snapshot in figure 4), as well as via frequent social media posts, press releases and interviews in the national press. Subsequent questionnaires have also been modified and refined in response to questions from participants. In addition, new questionnaire modules have been added based on the results of previous rounds and to examine the effects of changes in national policy.
Communicating COVID-19 cohort results to the public through the Corona Barometer. Snapshot of the Corona Barometer (https://coronabarometer.nl/), which is updated after every questionnaire round to present the most recent findings of the lifelines COVID-19 questionnaire in a format accessible by the public. The website is now interactive to enable users to look at trends over time and compare variables.
Findings to date
Response rates and characteristics of respondents
In every questionnaire round, 139 679 out of 159 482 current adult Lifelines participants are invited to respond to the COVID-19 questionnaire. In total, 71 992 (51.4%) Lifelines and LLNEXT participants responded to at least one of the first seven questionnaires, and response rates ranged from 33% to 39% (42 917–54 525) (figure 3). Compared with non-invited subjects (those without a known email address), invited subjects are younger, slightly more often female, have a lower body mass index (BMI) and are more often never smokers (table 2). Of the 139 679 Lifelines participants invited, 71 833 (51%) completed at least one of the questionnaires in the first 8 weeks of the programme. Compared with non-responders, these responders were slightly older, slightly more often female, had a higher BMI, were less often current smokers and more often ex-smokers (table 2). While Lifelines as a whole has been shown to be representative of the regional population,7 these slight differences in cohort makeup should be considered when looking at data from the COVID-19 cohort.
Characteristics of adult Lifelines participants invited to participate in the cohort and the participants of the COVID-19 questionnaire cohort during the first 8 weeks of the project (questionnaire rounds 1–7)
For LLNEXT, 321 people were invited to participate in the first 8 weeks of the project, and 159 participated (49.5%) (table 3). Compared with invitees, respondents were more likely to be female (73.5% of respondents vs 50.5% of invitees) (table 3). As LLNEXT recruits women who are currently pregnant, the age range was small and did not differ substantially between invitees, respondents and non-respondents. In LLNEXT, 80% of all parents who responded to the main Lifelines COVID-19 questionnaire also returned data on their children for the LLNEXT-specific module on COVID-19 and its impact on children. In total, we have data for 112 children up until week 8 of the COVID-19 questionnaire initiative: 96 children 0–3 years of age, 14 children 4–7 years of age and 2 children in the 8–18 age group.
Characteristics of LLNEXT participants invited to participate in the cohort and the participants of the COVID-19 questionnaire cohort during the first 8 weeks of the project (questionnaire rounds 1–7)
COVID-19 cases
Of the participants who responded in the first 8 weeks of the project, 1294 (1.8%) responded that they had been tested for COVID-19 and 127 (0.2%) tested positive. In addition, 887 (1.3%) respondents said they had been told by a doctor that they probably had COVID-19, while 5271 (7.3%) participants responded that they thought they had had COVID-19.
Early results and ongoing projects
One of the earliest results of the project was a clear signal that feelings of loneliness and isolation were substantially stronger in individuals who lived alone and that this effect was strongest in the youngest age group of respondents (18–30 years old, see corresponding panel in figure 4). There was also an increase in the number of unemployed respondents who reported losing their jobs due to the crisis, rising from 7.5% of unemployed respondents to the first questionnaire round (Q1) up to 14% by the week 5 questionnaire (Q5). More recent results have shown that by the end of May 2020, as the number of infections and hospitalisations dropped to low levels and schools and business reopened, respondents were reporting less anxiety, better sleep and fewer worries about losing their jobs.
The data collected from the Lifelines COVID-19 cohort is currently being analysed to address the four goals of the project. COVID-19 cases reported by participants are being used to track the outbreak, and the symptoms reported by participants are being used to generate a symptom-based COVID-19 prediction model.13 The data on chronic diseases, medication use and environmental factors (eg, cohabitation or smoking) will be used to look for associations with SARS-CoV-2 susceptibility and COVID-19 severity to help identify risk factors, protective factors and comorbidities. While factors such as age, sex, BMI and certain chronic illnesses have been associated with a more severe COVID-19 and higher mortality,14 there have also been questions about whether recent vaccinations can be protective,15 16 and cohort data should help address this. Important questions about why pregnant women and children seem to be relatively protected will also be analysed. Finally, it will be possible to look at genetic factors in detail as ~18 000 of the 71 922 COVID-19 participants who completed at least one questionnaire have been genotyped through the UMCG Genetics Lifelines Initiative. Next steps include identifying participants who were also participants in the Lifelines cohorts for which we have more detailed data, for example, participants with gut microbiome data, currently available for >10 000 Lifelines participants.
Mental health problems are known to increase in times of physical and psychological distress. The current COVID-19 pandemic is accompanied by strict government measures of social distancing and quarantine. As these events place significant stress on society and increase isolation and loneliness, close monitoring of mental well-being is important for both short-term and long-term public health policies and individual-level care. Alertness in clinical systems and tailored mental healthcare may be needed during and after such a mass traumatic event. The data from the Mini International Neuropsychiatric Interview major depressive disorder (MD) and general anxiety disorder (GAD)17 modules and the societal impact modules of the questionnaire will allow researchers to: (1) longitudinally track the prevalence of symptoms and diagnoses of MD and GAD during the pandemic in the Lifelines and LLNEXT populations, (2) associate symptoms with COVID-19 severity and outcome, (3) identify at-risk groups and individuals and (4) measure the impact of government policies on the overall mental health in the cohort.
The questionnaire will also help address the major impact the pandemic has had on the working lives of people in the Netherlands. Healthcare workers are a particularly vulnerable group due to their higher risk of being infected by SARS-CoV-2 and their working conditions during the height of infections included long work hours, cancelled holidays, adverse physical and psychosocial work conditions, that is, high psychological and emotional demands and low control.18 19 These working conditions, together with moral distress in relation to the family situation, may increase the risk for mental health problems and sickness absence in this female-dominated occupational group with a high baseline risk. Other ‘essential’ occupational groups are experiencing unprecedented changes in their working environments that may affect their physical and mental health as well as their labour market attachment. For many ‘non-essential’ occupational groups that are now encouraged to work from home, the home working environment might not be suitable, and many families now have to combine working from home with caring for children. This will likely impact the productivity and quality of their work and their level of stress.
The lockdown has led to a sudden disruption of the economy, with several economic sectors effectively brought to a standstill. The Lifelines COVID-19 questionnaire is monitoring changes in people’s current work situation by asking if they lost their job because of the crisis, if they are working in an essential job and whether they have to work from home. The answers to these questions will be used to monitor both the impact of the crisis on the short-term and longer term labour market and to identify workers most at risk of losing their job. This is essential information for policymakers to be able to target measures to the most vulnerable groups in society and mitigate the financial impact of the crisis.
Strengths and limitations
One of the main strengths of this cohort is its embedding within the long-running Lifelines prospective population cohort, which provides a rich data background about participants and the knowledge, infrastructure and relationship with participants necessary to recruit and engage participants during an evolving crisis. The high and sustained rate of response and the weekly questionnaires mean that the cohort can provide a detailed longitudinal prospective view of both the outbreak and the long-term impacts of the crisis. Another strength is the collaboration of researchers across a range of disciplines in designing and implementing a questionnaire that can be used to address a wide range of research questions, can have immediate impact on policy and can be used to help design new policies to prevent and/or manage renewed outbreaks. Finally, Lifelines will continue to follow its participants for the coming decade and beyond, providing opportunities to examine the long-term health impacts of SARS-CoV-2 infection and of the pandemic.
The Lifelines COVID-19 questionnaire was also designed to make comparisons with similar projects throughout Europe. Direct cross-national comparisons with projects in Denmark and France are possible, as they are using nearly identical questionnaires and will provide unique opportunities to examine the effects of different governmental measures on mental health and well-being. The cohort is part of COVID-MINDS network of longitudinal studies on the global mental health impact of COVID-19 (https://www.covidminds.org/),20 and the Lifelines COVID-19 project is also participating in the COVID-19 Host Genetics Initiative,21 an international collaboration to share and analyse data to identify the genetic determinants of SARS-CoV-2 susceptibility, COVID-19 severity and outcomes. In addition, the Lifelines COVID-19 questionnaires have been requested by other (inter)national researchers as a basis for designing their own questionnaires, for example, separate research has been done on the experiences of COVID-19 patients, making use of the Lifelines COVID-19 questions.
The timing and nature of the COVID-19 outbreak in the Northern Netherlands, which diverged from that in other parts of the country, is both a strength and a limitation. The relatively low number of cases in the region, even accounting for undiagnosed cases, may seem to pose difficulties for statistical association analyses looking at COVID-19 related factors. However, as of June 2020, >800 participants reported having had COVID-19, as confirmed by a positive test or a doctor’s diagnosis, which permits a wide number of statistical association analyses. Moreover, the impact of the societal steps taken to reduce the rate of infection in more heavily impacted regions of the Netherlands and the impact of the associated economic crises should have similar psychological and social impacts in the Lifelines population. The fact that the initial outbreak in the north was effectively capped by public health steps now puts the questionnaire programme in an interesting position to monitor the immediate health and societal impacts of the lockdown measures and the impact of coming out of lockdown. It also lays the groundwork for steps to be taken if there is a resurgence of COVID-19 infections, and the data generated while local infection rates were low could work as baseline values if subsequent outbreaks in the northern provinces are more intense.
Acknowledgments
We would like to thank all Lifelines and Lifelines NEXT participants for repeatedly filling out our questionnaires and all the experts involved in developing the content of the questionnaires. We would like to thank the Applied Health Research unit of the UMCG Department of Health Sciences, the UMCG Genomics Coordination Center, the UG Center for Information Technology and their sponsors BBMRI-NL & TarGet for storage and computing infrastructure. We would also like to thank Alex Friedrich and Gerolf de Boer for critical input for figure 1 and supplying the R(t) data in the figure and Trishla Sinha for editing the Lifelines NEXT children’s COVID-19 questionnaires.
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
JAMD, JM, HMB and LF are joint senior authors.
Twitter @P_Lanting, @PatrickDeelen, @JOMierau, @LudeFranke
PL, PD, HHW, JMV, APSO and SAJ contributed equally.
Contributors CW, JAMD, JM, HMB and LF conceived and implemented the study. KMI, PL, PD, HHW, JMV, APSO, SAJ, RW, IvB, FB, MXLD, JCH, AC, OB, EALM, UB, AZ, SAR, EZ, MAS, SB, RvO, VA, LHD, AS, SAS, JAMD, JM, HMB and LF contributed to the design and content of the questionnaire. PL, PD, HHW, JMV, APSO, SAJ and RW carried out the data analyses, established the Corona Barometer website and provided all figures and tables. KMI drafted the manuscript with contributions from PL, PD, HHW, JMV, APSO, SAJ, RW, IvB, FB, MXLD, JCH, AC, OB, EALM, UB, AZ, SAR, EZ, MAS, SB, RvO, VA, LHD, AS, SAS, JAMD, JM, HMB and LF. All authors reviewed and edited the manuscript and approved the final version.
Funding The Lifelines COVID-19 cohort is self-funded with in cash and in kind contributions from the initiators. The LLNEXT cohort is funded with a contribution of the UMCG Fund hereditary metabolic disorders (RVB16.0120). LF is supported by a Netherlands Organisation for Scientific Research (NWO) Corona Fast-Track grant (440.20.001), an Oncode Senior Investigator grant, a grant from the European Research Counsel (ERC Starting Grant agreement number 637640 ImmRisk) and an NWO VIDI grant (917.14.374). AZ is supported by the ERC Starting Grant 715772, NWO-VIDI grant 016.178.056, the Netherlands Heart Foundation CVON grant 2018-27 and the NWO Gravitation grant ExposomeNL 024.004.017. JM and LHD are supported by an NWO Fast-track grant (440.20.002). MAS is supported by the EU Horizon2020 EUCAN-connect programme (824989).
Map disclaimer The depiction of boundaries on the map(s) in this article does not imply the expression of any opinion whatsoever on the part of BMJ (or any member of its group) concerning the legal status of any country, territory, jurisdiction or area or of its authorities. The map(s) are provided without any warranty of any kind, either express or implied.
Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement Data are available on reasonable request. The data analysed in this study were obtained from the Lifelines biobank, under project application number ov20_0554. Researchers interested in using these data should contact the Lifelines Research Office (research@lifelines.nl).
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.