Article Text
PDF
Abstract
Introduction Immigrant and forcibly displaced women and girls are disproportionately impacted by the harmful health consequences of intimate partner violence (IPV) in the USA. Economic and Social Empowerment (EA$E), a women’s protection and empowerment intervention, has shown promising reductions in IPV and gender inequities among forcibly displaced populations (FDPs) in low-income and middle-income countries. However, research on the integration of gender equity interventions into economic empowerment programming for FDPs within the USA is lacking. Additionally, there is growing interest in integrating gender equity programmes among US-based refugee resettlement organisations, including the International Rescue Committee (IRC). We describe our study protocol for examining the feasibility, acceptability and appropriateness of EA$E for use with US-based FDPs, and recommendations for adaptation.
Methods and analysis This is a convergent parallel study to guide the adaptation of EA$E for use with US-based FDPs. Mixed methods (quantitative and qualitative) will be used for the adaptation research. Quantitative data will consist of brief surveys, and qualitative data will consist of focus group discussions (FGDs). Our research will be guided by the ‘administration’ phase of the ADAPT-ITT framework, which entails pretesting the intervention with the new target audience and implementation context to examine acceptability, appropriateness and feasibility to receive feedback to inform modifications of the original intervention. This is done via theatre testing, an innovative approach to pretesting that allows the new target audience to experience the intervention and provide feedback. We will conduct FGDs with IRC staff (n=4, total of 24 participants) and refugee clients (n=8, total of 48 participants, women and men, French and English speaking).
Ethics and dissemination The study has received approval from the George Mason University Human Subjects Committee (#1686712–7) and IRC (via reliance agreement). Results will be made available to refugee resettlement organisations, policymakers, funders and other researchers. This study has been registered in Open Science Framework (https://doi.org/10.17605/OSF.IO/SZDVY)
- PUBLIC HEALTH
- QUALITATIVE RESEARCH
- PREVENTIVE MEDICINE
- SOCIAL MEDICINE
- Health Equity
Data availability statement
Data are available upon reasonable request.
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 is the first study seeking to examine the adaptation of an internationally developed, evidence-based intimate partner violence (IPV) intervention for use with forcibly displaced populations in the USA to address IPV and gender equity.
Our multicountry and academic-community partnership and theatre testing allow for community input.
Study protocols were developed to allow for rigorous data collection despite mitigation measures pertaining to the COVID-19 pandemic.
The primary limitation of this study is delays in data collection due to the need for International Rescue Committee staff to be deployed to assist arrivals in the USA from Afghanistan, following the withdrawal of US Armed Forces from Afghanistan.
Introduction
The foreign-born population, otherwise referred to as immigrants, accounts for nearly 14% of the total population in the USA.1 Foreign-born populations can hold many different immigration statuses ranging from temporary protected status to refugee and asylee to those living in the USA without legal authorisation and permanent residents.2 Many immigrants in the USA were forcibly displaced.2 3 The International Organisation of Migration defines forcibly displaced populations (FDP) as the movement of persons who have been forced to flee their homes or place of residence due to persecution, conflict, violence, human rights violations or disasters.4 Globally, 89.3 million people worldwide were forcibly displaced as of 2021, with women and girls accounting for 49% of FDP.5
Research has shown that settings impacted by crises (eg, conflict, natural disasters) give rise to forced displacement. Such settings also have high levels of violence against women and girls (VAWG), with intimate partner violence (IPV) being the most common form of VAWG.6–8 Many US-based, forcibly displaced women arrive from countries where VAWG is prevalent and also face increased risk of IPV prior to, during and after displacement.9 10 Lifetime prevalence of IPV among US-based immigrant and forcibly displaced women range from 24% to 60%.11 Stressors associated with forced displacement and migration (eg, social isolation and financial strain) in the USA can exacerbate forcibly displaced women’s vulnerability to experiencing IPV.9 12 Forcibly displaced women in the USA and their families grapple with financial strain and loss while navigating new institutions.13 Forcibly displaced women in the USA often encounter more economic opportunity in the USA than in their home country, which contributes to rapid changes in traditional gender roles.13 14 Moreover, forcibly displaced women in the USA also report major barriers to well-being that include unequal power to men in their household, unpaid household labour, limited financial autonomy and reliance on their husbands for access to services.13 Unfortunately, despite these experiences, interventions designed to prevent IPV among forcibly displaced women in the USA primarily centre on clinic-based approaches, women-only education-based interventions and/or tertiary services without addressing broader inequitable social norms. Moreover, mainstream IPV prevention and response services often do not take the experiences of displacement (eg, trauma, change in gender roles and financial loss) into account.15
Evidence-based, internationally developed interventions that have shown promise in reducing IPV and promoting gender equity with FDP in the Global South could potentially address IPV and gender equity among FDP in the USA.16 Economic and Social Empowerment (EA$E), an internationally developed intervention, is a promising approach to addressing gender equity and reducing IPV among FDP and those impacted by humanitarian crises.16–19 The International Rescue Committee (IRC, a global humanitarian organisation focusing on humanitarian crises outside and inside of the USA) developed EA$E based on 4 years of research and community input. EA$E was developed to address two underlying drivers of IPV in crisis settings: household financial strain and inequitable gender norms. EA$E is a social norm intervention that invites couples who are engaged in economic empowerment activities to join EA$E Discussion Groups Series, which facilitates dialogue about attitudes and beliefs related to household financial well-being, budgeting, and communication and negotiation skills.20 Themes related to gender equity, gender roles and violence are woven throughout the series. Emphasis is placed on valuing women’s significant contributions to the household and engaging men to seek positive methods for affirming their role and resolving conflict. Through participatory exercises targeted at teaching concrete financial planning skills (eg, tracking income and expenses, setting goals and allocating resources), these discussions (1) equip men and women to plan and budget household resources together, (2) encourage men and women to communicate and negotiate effectively, and (3) challenge men to not abuse power and control in household decision making.20 As a result, women are able to safely voice their priorities in decisions that affect themselves and their families. EA$E’s approach to addressing IPV and gender equity is through a family-centred financial lens.
The overarching theory of change that underpins the combination of economic stability interventions with challenging gender norms for EA$E is shown in figure 1. This draws from both the theory of gender and power and feminist political ideology.20 These theories focus on structural barriers at the individual and relationship levels to address inequitable power structures, household decision making and resource access.21
Economic and Social Empowerment (EA$E) theory of change.
In 2010, EA$E was first implemented in Côte d’Ivoire, a country that has experienced colonialism and widespread conflict. Between 2010 and 2013, EA$E was rigorously evaluated through a two-armed, randomised control trial conducted across 47 villages with 981 women and their male partners (led by prinicpal investigator (PI) (JG) i partnership with IRC.16 This randomised controlled trial found that the incremental impact of adding the EA$E Gender Discussion Series to ongoing economic empowerment activities significantly improved gender equity and mental health outcomes.16–22 Qualitative data also indicated shifts in men’s attitudes towards joint decision making and gender equity.23 This research demonstrated that EA$E is a promising approach to address gender equity and reduce IPV among populations impacted by humanitarian crises in Côte d’Ivoire.18 24 EA$E has since been implemented in 18 crisis-affected countries throughout Africa and the Middle East. However, research on the integration of gender equity interventions into economic empowerment programming for FDP within the US is lacking.
Similar to the communities that IRC works with in Côte d’Ivoire, IRC clients in the US have experienced forced displacement due to war and conflict, and are regularly engaged in services promoting economic self-sufficiency.11 IRC clients in the US also experience or have experienced many of the drivers that research shows can exacerbate IPV among FDP (eg, loss in social status, rapid changes in gender roles, trauma, and social acceptance of IPV).11 13 Other US-based refugee resettlement organisations also offer economic empowerment programming (eg, job and financial coaching), and there is growing interest in integrating gender equity programmes. Thus, there is tremendous opportunity to integrate discussions of gender equity and IPV prevention into financial programming to promote safety, security, and well-being for forcibly displaced women and their families. However, research is needed to inform the adaptation process of the evidence-based EA$E intervention for use with US-based FDP.
To address these gaps in knowledge, we are conducting the currently described adaptation study to examine the feasibility, acceptability, and appropriateness of adapting EA$E for use with US-based FDP in Phoenix, Arizona. The study team is well-suited to conduct this research because the IRC co-PI (JD) and JG (PI) have an established partnership that began over a decade ago with evaluating the original EA$E intervention. Our team also includes international partnerships (IRC Côte d’Ivoire), the original developer of EA$E (KW), qualitative expertise (MW) and an implementation scientist (SR). Our specific research questions are as follows: (1) what programmatic components from EA$E are perceived (or not perceived) as acceptable, appropriate and feasible by US-based FDP communities? (2) why are some programmatic components perceived as acceptable, appropriate and feasible (or not perceived) for US-based FDP? (3) what are the barriers and facilitators to implementing an adapted version of EA$E by IRC’s US-based offices?
Methods and analysis
Overview of study design
The overarching purpose of our study is to investigate the feasibility, acceptability and appropriateness of the EA$E intervention and to generate recommendations for the US adaptation of EA$E for use with US-based FDP. Guided by the ADAPT-ITT model,25 we will employ a mixed-method, convergent parallel study design to guide the adaptation of EA$E. Our current adaptation research will focus on the ‘administration’ step of the ADAPT-ITT model.25 Quantitative data collection methods will consist of brief surveys, and qualitative methods will consist of focus group discussions (FGDs). The study will consist of three phases: (1) video production for theatre testing, (2) theatre testing with IRC staff and (3) theatre testing with forcibly displaced clients.
Adaptation framework
Our study will be guided by the ADAPT-ITT model.25 While originally developed for HIV interventions,25 the ADAPT-ITT model has since been used to adapt interventions for other public health issues both domestically and globally, including gender equity promotion and sexual violence prevention programmes.26 27
The ADAPT-ITT model comprises eight sequential steps to inform the systematic adaptation of evidence-based interventions in new settings and/or with new populations. The eight steps of ADAPT-ITT as it applies to our project are summarised in figure 2. Steps 1 and 2 have been completed; the current study will focus on administration (step 3). Steps 4–8 are planned for future research.
ADAPT-ITT model for EA$E. EA$E, Economic and Social Empowerment; FDP, forcibly displaced population; IPV, intimate partner violence.
The administration step of ADAPT-IT entails pretesting the intervention with the new target audience and context to examine acceptability, appropriateness and feasibility to receive feedback and to inform modifications of the original intervention.25 This is done via theatre testing, an innovative approach to pretesting that allows the target audience to experience the intervention and provide feedback. While the theatre testing is classically conducted with the original intervention in its entirety, we will conduct theatre testing using a preadapted and shortened version of EA$E.26 28 Relatedly, participants will not participate in the intervention during our theatre testing; instead, they will view the intervention by watching videos including role-playing of EA$E being simulated.
We selected ADAPT-ITT because (1) it will allow for meaningful engagement with FDP communities in Phoenix, Arizona, US-based IRC staff and Ivorian IRC staff; (2) we will be able to triangulate quantitative and qualitative data (eg, theatre test surveys and FGDs); (3) theatre testing allows for US-based FDP to have an immersive experience of EA$E in Côte d’Ivoire via video-recording; and (4) the systematic approach facilitates documentation for the adaptation plan.25
Phase I: video production for theatre testing
In our study, theatre testing will entail all research participants viewing video recordings of condensed simulated EA$E sessions. The US-based research team worked together with the IRC Côte d’Ivoire team to develop and video-record simulated EA$E session role-plays. The IRC team in Côte d'Ivoire worked with past Ivorian EA$E participants (ie, graduates) in filming the sessions. The role-plays were recorded in Ivorian French and emphasised the key objectives of each EA$E session. After the recording completion, videos were sent to a professional translation and transcription company for translation into English. The research team used video-editing software to subtitle in English. The development of these videos will allow for US-based populations to view and experience the original intervention while maintaining COVID-19 mitigations.
Phase II: theatre testing with IRC staff
Recruitment and sampling
The first round of theatre testing will be held with IRC’s Economic Empowerment and Women’s Protection and Empowerment (WPE) staff in the USA. Using a purposive sampling approach, we will recruit 10–12 economic and empowerment staff members and 10–12 WPE staff members over a series of approximately four to five FGDs in total. Staff from IRC’s Economic Empowerment and WPE teams will be recruited from 25 offices including Abilene, Texas; Atlanta, Georgia; Baltimore, Maryland; Boise, Idaho; Charlottesville, Virginia; Dallas, Texas; Denver, Colorado; Elizabeth, New Jersey; Los Angeles, California; Miami, Florida; Missoula, Montana; New York, New York; Oakland, California; Phoenix, Arizona; Richmond, Virginia; Sacramento, California; Salt Lake City, Utah; San Diego, California; San Jose, California; Seattle, Washington; Silver Spring, Maryland; Tallahassee, Florida; Tucson, Arizona; Turlock, California; Wichita, Kansas.
Data collection
Theatre testing will occur through virtual FGDs conducted through Zoom29 to allow for COVID-19 safety measures and will be video-recorded/audio-recorded. Informed consent will be obtained electronically, prior to the start of the FGD, followed by a short set of demographic questions (eg, IRC office affiliation, role/position, years worked at IRC, years worked with FDP, main groups worked with and country of origin). A brief overview of the intervention and its sessions will be given, followed by an opportunity for questions and answers. After viewing each session, participants will complete a brief survey through Mural30 to quantitatively assess our outcomes of interest (figure 3). We will then conduct FGDs after each session to allow for an in-depth understanding of adaptation recommendations. Specifically, participants will be asked a series of questions regarding the degree to which components of the session were (or were not) acceptable, appropriate and what they perceived (or did not perceive) as feasible. Participants will also be asked what they liked and did not like, what they would add or take away, what they would like to change, as well as any broader comments on how the sessions would need to be adapted for the USA.
Sample mural survey. FDP, forcibly displaced population; IRC, International Rescue Committee.
Based on feedback from IRC staff, the number of sessions to review and the length of the focus group guide, theatre testing will be split into 2 days. We anticipate that the theatre testing and focus groups will take 6–8 hours over the course of 2 days. We do not anticipate challenges with internet access as IRC programming has been offered virtually since March 2020. Moreover, if a participant’s connection becomes disconnected during the FDG, they will have the option to participate in an individual in-depth interview. In our consultations with IRC in planning the study, many raised the need to provide flexible options for completing the theatre testing due to the ongoing pandemic and the IRC staff’s roles/deployment in the resettlement of Afghans to the USA. Thus, videos and quantitative surveys will also be made available to IRC staff who cannot attend the full focus groups. They will then have the option of completing the qualitative questions via (1) individual interviews, (2) written responses or (3) provision of self-recorded video or voice memos.
Phase III: theatre testing with forcibly displaced clients
Recruitment and sampling
The second round of theatre testing will be held with clients served by the IRC office located in Phoenix, Arizona, as it is Arizona’s largest resettlement office. It has provided services for FDP since 1994 and serves 13 000 clients annually. This site was selected for theatre testing due to the expansive programming and partnerships they have for survivors of IPV and other forms of violence and robust economic empowerment programmes (eg, asset building, business development, financial coaching, job readiness, vocational training and employment services). Similar to other IRC offices, IRC staff in Phoenix have also requested more technical assistance with integrating gender equity into economic empowerment programming. Based on a purposive sampling approach, FDP will be recruited by the IRC staff using texts, phone calls and social media. Inclusion criteria for FDP are (1) at least 18 years old, (2) married or in a cohabiting relationship, (3) French or English speaking, and (4) have resided in the USA for at least 3 years. Theatre testing and FGDs will be conducted in French and English. We will conduct eight client FGDs, organised by language and gender, consisting of five to seven participants each (four French, four English; two women’s groups and two men’s groups per language). Francophone clients will include clients from Francophone countries, while English-speaking clients will come from a range of countries, including Iraq and Afghanistan, and will represent those who generally have been in the USA for longer and/or had access to higher education opportunities or job opportunities that facilitated English language acquisition. This cross section of clients will provide the opportunity to test the adaptability of EA$E across cultures and countries of origin while also looking at clients who have had time to establish some ties to the community in the USA.
Data collection
Informed consent will be obtained prior to theatre testing by a language-matched member of IRC. We will hold separate groups for men and women in each language to promote comfort and more open discussion around topics such as gender equity, norms and financial decision making. Consistent with procedures for the IRC staff, a brief overview of the intervention and its sessions will be given, followed by an opportunity for questions and answers. Participants will then watch each video session followed by an FDG to allow for an in-depth understanding of adaptation recommendations after viewing each session. We will continue to explore the optimal length of theatre testing with IRC clients, as well as format. We do not anticipate internet access challenges with forcibly displaced clients (as IRC has reported improved participation in programming with remote formats). However, we do anticipate potential challenges with childcare needs. For this reason, we will not require that couples jointly participate in the theatre testing. Data collection was held from May 2022 (pilot) and will continue through March 2023.
Implementation outcomes
We will use measures of acceptability, appropriateness and feasibility as developed by Weiner et al.31 These measures were developed in response to the need for implementation measures with strong conceptual clarity and psychometric properties. These measures fill an important gap in providing a reliable and valid approach to assessing stakeholder perspectives of intervention adaptation and approaches. Weiner et al’s measures are also pragmatic in that they are brief, easy to use, and are not dependent on any specific context or intervention type.31 Moreover, these measures have been used in implementation science research, including within low-ncome and middle-income country settings.32 The quantitative measures will be triangulated with findings from our FGDs that will also focus on acceptability, appropriateness and feasibility while also probing for perceptions of comfort, stigma and safety in discussing financial decision making, household communication and IPV among the FDP in Phoenix. Questions will be based on instruments used by our team in Côte d’Ivoire. Sample questions will include ‘What did you like/dislike about the module? What would you take away/add to this module? Tell us about how you think forcibly displaced families in this area would feel about discussing their relationships and household decisions during a programme like the one you just saw’. We will also qualitatively examine implementation strategies (eg, what would be the optimal timepoint following arrival to the USA for offering EA$E, preference for remote versus in-person intervention delivery and barriers/challenges with participating. These outcomes will be explored in both phases II and III (IRC staff and clients).
Patient and public involvement
No patients were involved in this study.
Data analysis
FGDs will be video-recorded/audio-recorded and transcribed verbatim. The transcripts will be entered into Dedoose, a computer software program designed for the coding and analysis of qualitative data.33 A codebook will be created to describe each code and when it should be applied, and each transcript will be topically coded. Two team members will independently code a transcript, and then will compare codes to resolve discrepancies and revise as needed, aiming for an inter-rater reliability of 85.0%.34 35 Constant comparison analysis strategies, aided by visual displays, will be used to analyse participants’ responses.36 37 This method will involve data generated from FGDs to support emerging categories and themes. Early FGD data will be coded, and as more data are collected, these codes will be revised. We will organise the analysis by each module and assess what topics generate the most discussion, agreement and disagreement regarding our questions of interest across all FGDs. Codes that appear to fit with one another will be clustered into themes and compared between FGDs and between women, men and IRC economic and empowerment and WPE staff. We will note key differences and similarities that emerge across different groups of participants (eg, staff vs clients, women vs men).These analytical steps will continue until no new themes emerge and will be documented using audit trail methodology.36 37
Survey data will be analysed using RStudio,38 including a descriptive analysis of demographics. Analysis will include mean scores of the intervention acceptability, appropriateness, and feasibility scores for each module and the overall programme, along with variations in scores across demographics. Findings will be interpreted and synthesised to respond to the key research questions. Data-informed recommendations for the adaptation of EA$E for US-based FDP will be formulated based on the findings using the traffic light approach.39 Recommendations for green light modifications of EA$E for US-based refugees will be adaptations that can be implemented without any concern for safety, ethics or core elements for impact on gender equity outcomes (eg, consolidating sessions and adding additional debriefing sessions). Recommendations for yellow light modifications will be adaptations that we recommend with caution and guidance. Red light modifications will be adaptations that we would not recommend for the programme. Decisions guiding the adaptations will be informed through a series of stakeholder engagement meetings with the research team, US-based IRC staff, US-based IRC clients and international IRC staff, including Côte d’Ivoire-based programmatic staff.
Discussion
This study seeks to examine the feasibility, appropriateness and acceptability of adapting an internationally developed women’s protection and empowerment intervention—EA$E—for FDP in the USA to address the unique challenges forcibly displaced women and households face in achieving gender equity. Study findings will inform the adaptation of promising global gender equity promotion interventions for US-based FDP and will fill an important gap in current intervention research on VAWG.
As with all research on IPV, we do anticipate challenges and have outlined steps to address such challenges. One challenge may be client discomfort in speaking about gender equity and IPV during focus groups. Although we are not asking about personal experiences with IPV, we are aware that such experiences may be inadvertently shared in a focus group setting. We will thus adhere to the WHO protocols for conducting research on violence against women.40 41 The entire team is well versed in conducting research on sensitive topics such as IPV and gender equity. Such protocols are standard in our field, and thus we have extensive expertise in ensuring safety while conducting rigorous research. Examples of such ethical approaches include referring to services should a participant need or request support. Moreover, to minimise chances of disclosure of personal experiences with IPV during FGDs, participants will be guided to speak about acceptability, feasibility and appropriateness within their community, as opposed to disclosing personal IPV experiences. This will also allow us to obtain an understanding of perceived norms. This approach has been widely used for IPV research, including our research team and IRC. We will also use validated measures for feasibility, acceptability and appropriateness that are simple to administer, thus reducing respondent burden. Notably, because many participants will be reading off of subtitles, it is possible that participants’ reporting and description of perceived feasibility, acceptability and appropriateness may be impacted. To reduce potential impacts, we have carefully limited our inclusion criteria to those who have resided in the USA for 3 years to ensure that our participants have had exposure to the English language prior to participation. In terms of the measures used in the surveys, to address lower levels of literacy that may be present among some participants, we will use both text and pictorial Likert scales.42 For the qualitative data, we will train coders to assure consistency and will aim for a high inter-rater reliability. We will also maintain a strong academic and community partnership to ensure that research procedures and protocols, including data interpretation, are ethical and meaningful.
The ongoing and protracted COVID-19 pandemic does pose challenges to the research in terms of limiting opportunity for in-person data collection. However, our research team has planned innovative strategies for remote theatre testing and data collection that will offer important lessons learnt for remote data collection strategies and adaptation research. In response to IRC’s major role in the resettlement of Afghan evacuees, our timeline has shifted. Our original plan was to complete all theatre testing with IRC staff prior to theatre testing with clients. However, we have delayed the start of the IRC staff theatre testing to allow for the IRC staff to complete their deployments. All theatre testing (for both staff and clients) will no longer be held in a sequential manner.
We anticipate recruiting IRC clients from diverse countries, including Afghanistan, Democratic Republic of Congo, Irag and Syria. While clients will originate from very different cultures and traditions, our study will focus on the shared experiences of forced displacement. IRC’s experiences (including many staff who themselves are forcibly displaced) has shown that the commonality of experiences as FDP outweighs cultural differences. This is also consistent with more recent frameworks in addressing migrant health.43
Our expected outcomes from this research study are to produce data-driven recommendations for adapting EA$E for use with US-based FDP. These recommendations will be disseminated to other refugee resettlement and immigrant-serving organisations. We ultimately plan to develop an EA$E–US curriculum for implementation and evaluation.
Ethics and dissemination
The study has received approval from the George Mason University Human Subjects Committee (#1686712–7) and IRC (via reliance agreement with George Mason University). All research activities will adhere to the WHO protocols for conducting research on violence against women.40 41 Results will be made available to refugee resettlement organisations, policymakers, funders and other researchers.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Footnotes
Twitter @jhumkagupta
Contributors JG acquired funding, conceptualised the study and led the writing of the paper. JD acquired funding, conceptualised the study and assisted with the writing of the paper. SK assisted with the writing of the paper. SR advised on implementation science and assisted with the writing of the paper. PCB assisted with the conceptualisation of the study. MW advised on qualitative data collection and assisted with the writing of the paper. KW assisted with the writing of the paper.
Funding This work was supported by the Robert Wood Johnson Foundation’s Evidence for Action, in response to a call for 'Approaches to Advance Gender Equity From Around the Globe' (grant number 223879).
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, conduct, reporting or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.
© Author(s) (or their employer(s)) 2023. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.