Article Text

Motivational interviews to enhance advance care plans in older adults: systematic review
  1. Tongyao Wang,
  2. Mu-Hsing Ho,
  3. Xinyi Xu,
  4. Hye Ri Choi and
  5. Chia-Chin Lin
  1. School of Nursing, Li Ka Shing Faculty of Medicine, University of Hong Kong, Hong Kong, Hong Kong
  1. Correspondence to Professor Chia-Chin Lin, School of Nursing, University of Hong Kong Faculty of Medicine, 999077 Hong Kong, Hong Kong; cclin{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Advance care planning (ACP) improves quality of care.

  • Motivational interview (MI) is a communication strategy to increase positive behaviour changes.


  • Seven MI-based ACP interventions were explored

  • MI-based ACP is feasible and effective.


  • Clinical: training lay-persons to delivery MI-based ACP inthe community is feasible.

  • Research: offers evidence for developing future ACP interventions


Conversations on advance care planning, sometimes known as ACP, can be difficult for many people. ACP refers to a process by which a healthcare provider prepares an individual to plan for the care they would or would not like to receive before any serious illness or unexpected health crisis, to learn about the types of decisions that need to be made, and to document their wishes in the medical records or advance directives.1 An advance directive or advance health directive, also known as a living will, is a legal documentation in which a person specifies the treatments and health services they would like to have when they are no longer mentally competent to make their own decision.2

Most people tend to avoid talking about death, especially older adults, and some are also afraid of not receiving the best treatment if they have advance directives in place.3 Despite the critical role the ACP programmes play, the implementation of the ACP service into clinical practice is still limited. Rates of use of advance care plans remain low, with less than 20% among older adults in western countries such as the USA, UK, Finland, Australia and the Netherland.4 According to a study conducted with physicians, advance practice nurses and other allied health providers,5 although the majority of healthcare professionals recognise the value of providing ACP conversations, few managed to implement it into their daily practices because of competing priorities. To further engage older adults in ACP activities, there were multiple initiatives in countries around the world. For instance, in 2012, the USA also started the Conversation Project as a nationwide, public initiative to help everyone voice their wishes for end-of-life care in the community and clinical setting.6 In 2017, Canada launched the Palliative and Therapeutic Harmonisation model for frail older adults during home visits.7 However, with an ageing population, the gap in ACP service remains largely unfilled, and more effective interventions driven by theories and empirical evidence are needed.

Motivational interviewing is a direct, patient-centred counselling strategy that encourages patients to explore and resolve their ambivalence and promotes their confidence in their ability to change their behaviour. It was originally developed as a treatment for substance abuse and binge drinking in the 1980s.8 It has been applied successfully in a variety of settings, including the management of chronic disease, smoking cessation, weight loss programmes, treatment for substance use disorders and other health behavioural interventions.9

Motivational interview (MI)-based counselling is a set of specific counselling behaviours guided by the principles of resisting the righting reflex, understanding the patient’s motivations, listening with empathy and patient empowerment.8 And it is operationalised by a collection of communication skills such as open-ended questions, affirming, double-sided reflective listening, querying extremes, informing and summarising.8 Motivational interviewing-based counselling emphasises partnership, eliciting patients’ internal motivations and reflecting on perceived barriers to change.

Taking an MI counselling approach in ACP, providers or clinicians are encouraged to resist the urge to address patients’ resistance by correcting their perceptions but to acknowledge and explore the values behind it.10 The MI principles also encourage clinicians to approach patients with open-mindedness, to recognise patients as experts in their own lives, and to respect patients’ autonomy to make their own decisions. The principles further encourage clinicians to accept and give affirmations when patients express their emotions and reflect on patients’ internal motivations and goals for end-of-life care. Using the fourth principle, clinicians can empower patients in decision-making by supporting patients’ self-efficacy in end-of-life care decisions. The Coalition for Supportive Care of Kidney Patients published a training curriculum guide on ACP with an MI approach to assist staff in having ACP discussions with patient.11

The potential advantage of applying motivational interviewing techniques to end-of-life communication has been highlighted in recent literature and clinical practice guidelines. However, no systematic review has been conducted. We conducted a systematic review of MI-based ACP interventions for improving ACP-related outcomes (having advance directives and health delegate documentation) in older adults. The review also seeks to identify and summarise evidence from all relevant studies to inform future research on the feasibility, appropriateness, meaningfulness and effectiveness of MI-based ACP interventions for older adults.


We undertook a systematic review approach, which was chosen so that studies from different research traditions could be synthesised to guide clinical practice.12 To better understand the state of evidence for MI counselling in ACP interventions, the review was synthesised based on the ‘Feasibility, Appropriateness, Meaningfulness and Effectiveness (F.A.M.E.)’ framework, including the following domains: feasibility, appropriateness, meaningfulness and effectiveness. The protocol for this meta-analysis and systematic review was registered on PROSPERO (CRD42022341178). The review process was conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.13

Search strategy

The search question was created in a patient–intervention–comparison–outcomes (PICO) format: in older adults (P) receiving MI-based counselling on ACP (I) in comparison with standard care (C) to examine the impact on the completion of advance directives (O). The initial search was conducted on 27 March 2023 in the following electronic bibliographic databases, MEDLINE/PubMed, Education Resources Information Center, PsycINFO and Cumulative Index to Nursing and Allied Health Literature. There was no restriction on the search period. A base strategy, ‘[motivational interviewing or motivation interview or motivational therapy] AND [advance care planning or end of life planning or advanced directive]’ was developed using MeSH terms. Reference lists of all potentially relevant studies or systematic reviews were screened for any eligible studies that were not found in the initial search.

Inclusion criteria

In the review, the operational definition of the application of motivational interviewing was the specific documentation of using MI techniques in the intervention manual. Studies that met the following criteria were included in this review: (1) one of the following study designs (qualitative, quantitative randomised controlled trials (RCT), quantitative non-randomised, quantitative, descriptive or mixed methods); (2) adults 18 years old and above; (3) reporting an intervention targeting ACP; (4) MI techniques/skills as a major component of the intervention and (5) the comparison group was standard care.

Study selection and data extraction

We followed the PRISMA guideline (figure 1). Based on the inclusion and exclusion criteria, one reviewer screened the list of titles and abstracts identified through a search of databases. The full text of the selected studies was further reviewed independently by two reviewers (TW and MH) independently for quality and validity purposes. Reasons for excluding a study are shown in figure 1. Discrepancies between the reviewers about the inclusion or exclusion of any study were resolved by discussion until both reviewers approved.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow chart.

Assessment of study quality

The assessment of the methodological quality was conducted by two independent reviewers (T.W. and M.H.) using the Mixed Methods Appraisal Tool (MMAT) V.2018, which is a framework for critically appraising studies included in a systematic review of mixed-methods studies.14 The MMAT is a questionnaire examining the overall study quality based on seven methodological quality criteria for each category of a research study. Reviewers’ judgement for each criterion was categorised as yes, no or ‘cannot tell.’ Then, an overall score of study quality was generated.

Data extraction

For descriptive and feasibility studies that did not have means and SDs, descriptive results were reported in online supplemental table 1 by name of the project, reporting studies of the project, study aim, study design, participants, description of the intervention and study results. For RCTs, the mean and SDs of each outcome measurement were extracted from the intervention and control groups. One study reported results in an OR format only. An email request was sent to the corresponding author for the original values.

Supplemental material

Data analysis

The quantitative and qualitative data were first aggregated based on the following four domains in the ‘F.A.M.E.’ framework: feasibility, appropriateness, meaningfulness and effectiveness. Quantitative results were presented descriptively.


Study selection

A total of 744 records were found, and additional 17 records were identified from reference lists and other sources. After screening titles and abstracts, 26 studies were retrieved for full-text review. Additional 20 full-text articles were identified by citation searching and experts in the field. Finally, four RCTs, six quantitative feasibility studies and three qualitative studies were included.

Methodological quality

The quality assessment of the included study is demonstrated in table 1, and all included studies demonstrated satisfied performance during the quality assessment using the MMAT.

Table 1

Quality appraisal via the MMAT

Study characteristics

The review included 14 studies that examined seven MI-based ACP interventions in assisting older adults with completing advance directives, namely, ‘Sharing and Talking About My Preferences,’15 ‘Emergency Department Brief Negotiated Interview,’16 ‘Make Your Wishes Your Way (MY WAY),’17 ‘Support Older Latinos on ACP,’18 ‘Brief Motivational Stage-Tailored Intervention,’19 ‘Thinking Ahead’20 and ‘Apoyo con Cariño/Support With Caring.’21 The review comprised a total of 1265 adult patients. Participants were diagnosed with multiple comorbidities or advanced illnesses such as end-stage cancer or kidney disease, and three studies were conducted on older adults without any health impairment in the community. The settings of the MI-based ACP intervention include the emergency department, dialysis clinics, supportive housing, hospice settings, primary or specialist ambulatory clinics, or residential homes. Trials were conducted between 2015 and 2021. All studies were conducted in the USA: Massachusetts, New Mexico, Washington, DC, California, Alabama, Connecticut and Colorado, respectively. Two studies were conducted to explore cultural-specific ACP interventions for African Americans and Latinos.

Characteristics of ACP interventions

Two key elements in the intervention were ACP education and/or counselling with an MI approach. Four out of the seven interventions provided educational materials.17 20–22 The official advance directive forms, such as the National Physician Orders for Life Sustaining Treatment form17 and the health-literacy adaptive advance directive form,20 were also provided in two studies. In terms of the MI-based counselling section, three studies followed script-based intervention manuals for the MI-based counselling session.17 21 23 The procedure involved in MI-based counselling includes three key phases: (1) encourage the expression of values and beliefs, (2) exploring and empowering internal motivation and (3) supporting decision-making. The counselling ranged from 11 min to 90 min in one or more separate sessions.

Description of the ACP interventionists

Interventionists in the six ACP projects were physicians,23 nurses,17 social workers,17 22 nursing assistants and medical assistants,24 individuals with minimal healthcare experience,20 or health psychologists or psychology interns.15 For projects that targeted Spanish-speaking communities,21 22 interventionists were also required to be bilingual and familiar with the Spanish culture and the communities where the studies were conducted.

Description of the ‘train the trainer’ protocol

The intervention training comprised (1) education on ACP and (2) MI training and role-play practice, with the exception that training for physicians working in the emergency department only included training in MI coaching skills.23 The length of the intervention varied from 3 hours to 200 hours. Training on ACP knowledge and end-of-life care, for those with minimal clinical experience, was provided in structured courses. Huang’s team provided the ACP Facilitator Certification Course from the international, evidence-based, decision-making organisation,20 and Fischer and his colleagues provided the course, Education in Palliative and End-of-Life Care from the Northwestern University, School of Medicine.21 Training on MI skills and supportive counselling were provided via workshops including lectures, discussion and role-play practices. Lupu and his colleagues also established a training curriculum on MI-based ACP specifically for patients with end-stage kidney disease.17 Core MI strategies that were incorporated into the training workshops encompassed reflective listening, resisting the righting reflex, rolling with resistance, eliciting internal motivation, asking open-ended questions, patient-centred feedback, using affirmation, supporting autonomy, change planning and informing.15 17 20 23 Competency in MI skills was assessed using the coding of recorded counselling sessions25 or bedside auditing.23

The feasibility domain

Feasibility refers to whether the intervention delivery process is physically, culturally or financially practical in the context of ACP communication.26

Cultural and financial feasibility

The cultural feasibility was managed by having interventionists who were familiar with the target participants’ culture and languages, and designing the intervention and educational materials at the participants’ health literacy levels.21 In terms of financial feasibility, though no study reported cost-effectiveness data, interventions conducted in community settings used a variety of workforces, such as nurses, social workers, nursing or medical assistants, and lay persons. Moreover, in the emergency department, Leiter and his colleagues also found that there was no significant difference in intervention fidelity between a physician and a physician assistant (p=0.08).23

Recruitment and completion rates

The physical characteristics of intervention delivery were measured by the study recruitment rate, completion and dropout rates, and length of the interventions. The recruitment rates, which were defined as the percentage of eligible people recruited in a study, were 36% in adults with end-stage kidney disease, 47%–55% in older patients with advanced illness in the emergency department,16 23 67% in older adults living in supportive housing19 and 51.5% via telephone call of patients in ambulatory clinics.15 The intervention completion rates, which were defined as the portion of participants who received the full intervention dose, were high, ranging from 94% in healthy older adults,19 81.3% and 100% in community-dwelling Latinos and African Americans20 21 and 53.8% in patients in ambulatory clinics, respectively.15 In addition, interventions delivered in the clinical setting were in one session only and took an average of 11–65 min.17 23 Interventions in the community settings encompassed multiple sessions, and the mean time of delivery was longer as well, from 85 min to 105 min.20 24 Furthermore, one study also reported intervention fidelity of 78%, indicating that a high portion of the ACP counselling had incorporated MI techniques.23

Appropriateness and meaningfulness of MI-based ACP interventions refer to the patient’s personal experience with the intervention, such as their positive and negative views on the intervention, and their interpretation of the interventions in terms of satisfaction and usefulness.26 Relevant information was extrapolated from three studies where participants, navigators or authors commented on the appropriateness and/or meaningfulness of the MI-based ACP interventions or where their recommendations for intervention improvement were discussed.25 27 28

Stages of change in ACP

Guided by Prochaska and DiClemente’s stages of change transtheoretical model,28 29 it was found that participants who received an ACP education plus MI-based counselling interventions were at different stages of change (ie, precontemplation, contemplation, preparation, ACP action and maintenance). Participants in the precontemplation stage felt uncertain about ACP after receiving education and counselling, and they could not recall what ACP meant even after intervention. Some participants expressed confusion or misunderstanding about the purpose of an advance directive, and some did not believe it applied to them because they were not sick enough. Participants at this stage also expressed that it was a hard subject to discuss with family and providers. Participants in the contemplation and preparation stages were open to learning new ideas and how they could prepare for ACP. The authors did not indicate the number of participants at each stage.28 As participants may experience different levels of readiness for ACP after receiving the intervention, it was recommended that providers hold ACP discussions multiple times.28

Intervention satisfaction

All the studies included in the review reported high satisfaction with their programmes from participants, with a satisfaction rate ranging from 70% to 96%.17 19 Latino participants found the ACP intervention particularly useful as it filled their knowledge gap.22 In a study conducted by Ouchi’s team, over 70% (17/23) of the participants who received the intervention viewed the intervention as positive.25 Participants largely attributed their positive experience to the emergency clinicians’ gentle and comforting approach. The intervention identified questions for participants to ask their doctors and provided the opportunity for them to reflect on their future care.25 In Fink’s study, the patient navigators were able to establish relationships and rapport with participants and family caregivers and their knowledge played a large role in assisting patients with advance directive completion.27 Furthermore, patient navigators empowered participants to communicate with providers about their pain and other symptoms. The bilingual and bicultural navigators played a role as cultural brokers and were effective in reducing barriers to palliative care and improving health literacy for underserved populations.

Sensitive conversations at the personal and culture levels

Some participants had mixed emotions during the discussion, but they emphasised the helpful nature of the intervention despite the upsetting topic.25 Two participants in Ouchi’s study had negative sentiments regarding the intervention and wished the discussion was less vague and more direct.25 Lupu and his colleague also reported that 4 out of 109 participants experienced moderate distress during the conversations on end-of-life care.17 Furthermore, ACP is not a lived experience among Latinos and can be perceived as interfering with traditional and religious values. The authors also expressed that the existing primary advance directive documentation was not culturally and linguistically adapted for those with low levels of health literacy.28

The effectiveness domain

Effectiveness refers to the impact of the MI-based ACP intervention on clinical or health outcomes.26 The primary outcome in this review was the completion of advance directives, and, secondary outcomes include ACP engagement (knowledge, attitude and skills) and health service utilisation.

Impact on the advance directive completion

Four interventions, ‘MY WAY,’ ‘Support Older Latinos on ACP’, ‘Apoyo con Cariño (Support with Caring)’, and ‘Sharing and Talking About My Preferences’ reported their impact on advance directive documentation. Participants who received the ‘MY WAY’ intervention were 79% more likely to complete an advance directive ((95% CI, 1.18 to 2.72), 32.8% vs 17.8%, p=0.004), and 2.2 times more likely to have a documented health agent ((95% CI, 1.32 to 3.68), 29.6% vs 13.2%, p=0.001) in comparison to usual care.17 The RCT of ‘Support Older Latinos on ACP’ showed that, in comparison to receiving an ACP education, those who received additional MI-based ACP counselling sessions had a significantly higher completion rate of advance directives (OR=6.901, p<0.05).18 In terms of the ‘Apoyo con Cariño (Support with Caring)’ intervention, the completion rates of living wills and powers of attorney were 39.3% and 61.6% in comparison to 12.6% and 25.2% in the control group that received an educational booklet only.24 In the ‘Sharing and Talking About My Preferences’ study, participants who received MI-based counselling or MI-based counselling enhanced with tailored education handouts were significantly more likely to talk about quality versus quantity of life with a loved one, to assign a health delegate, to complete their living wills and to have their decisions written in their health records that those received usual care.15 Notably, Fried and his colleagues also found no significant differences in ACP-related outcomes between the usual care group and the group that received tailored, stage-based education handouts in their mailbox.15

Impact on ACP engagement (knowledge, attitude and confidence)

The effectiveness of the other three interventions was tested using a cross-sectional feasibility design, which was assessed by the change in ACP engagement measured by individuals’ knowledge, attitudes and confidence.16 19 20 The ‘Brief Negotiated Interview in ED’ intervention showed a significant positive impact (from 3.8 to 4.3 out of 5, p=0.01) on participants’ readiness to sign advance directives and communicate end-of-life care decisions with family members and physicians.16 In terms of the ‘Brief Motivational Stage-Tailored Intervention’ for low-income, supportive housing residents, participants had significant increases in ACP knowledge (p<0.001), attitudes towards ACP (p<0.05), self-efficacy on advance directives (p<0.05) and perceived importance (p=0.05), but no change in negative attitudes (p=0.91) was observed.19 The ‘Thinking Ahead Intervention’ for community-dwelling African Americans resulted in a significant increase (p=0.01, d=1.67) in knowledge related to advance directives.20 In addition, two out of three RCTs also reported significant improvements in ACP engagement measurement (p=0.03)17 and readiness to engage in ACP discussion with family members (p<0.001) and physicians (p<0.03).18

Impact on health service utilisation

The project involving two RCTs21 24 that tested the ‘Apoyo con Cariño (Support with Caring)’ intervention reported health service utilisation. Both studies did not find significant differences in hospice utilisation, length of hospice stay or aggressiveness of treatment received at the end-of-life, between the intervention and control groups.


Main findings

This is the first systematic review to summarise interventions adopting MI in end-of-life care planning and to synthesise the evidence on the effects of seven MI-based ACP interventions. We used the F.A.M.E. framework to integrate quantitative and qualitative evidence to gain a comprehensive understanding of the state of science on MI-based ACP interventions.

The quantitative evidence was complementary to the quantitative evidence, and the aggregated findings support that ACP counselling delivered using MI techniques by health professionals or certified laypersons is effective in improving older adults’ engagement in ACP but has no impact on health service utilisation. The MI-based ACP interventions reviewed in the study demonstrated adequate feasibility, as evidenced by the high recruitment and study completion rates in a variety of patient populations and settings. Notably, we found that the MI counselling on ACP for older adults in the community is feasible to be delivered by lay persons with appropriate training, which indicates a great potential to scale up the intervention for a larger population.

The qualitative data also suggest that the MI-based ACP intervention was appropriate and meaningful for participants, who viewed the interventions as a helpful and new experience in end-of-life discussions and would recommend them to others. Participants generally appreciated the supportive communication styles in MIs to express their personal values and explore the motivation for ACP engagement. This could be explained by the fact that the philosophy of motivational interviewing aligns well with the objective of ACP conversation, which is to encourage patients to discover their own values, as opposed to pressuring them to provide the correct responses or perform the appropriate behaviours.30 In addition, the cultural appropriateness was also analysed in the qualitative data, and indicates that ACP interventions should be flexible and sensitive not only to the recipients’ cultural background on death-related talks but also to their readiness to engage in those conversations.31 Therefore, ACP counselling based on MI revealed a significant potential for practicability and acceptability in routine practice for the elderly.

In terms of effectiveness, all the studies found a significantly positive impact of the ACP interventions on the completion of advance directives and health delegates. This was consistent with previous meta-analysis that showed ACP interventions in general can be effective (OR=3.26, 95% CI (2.00 to 5.32)) in increasing the completion of advance directives.32 Likewise, the effectiveness of the MI-based ACP intervention was in line with the significant effects of brief MIs on changing health behaviours such as weight control, blood pressure and substance use management.9 33 Furthermore, this finding is also in line with the transtheoretical model, which states that individuals’ decision-making for a behaviour change is a continuous cognitive and affective evaluation process through six stages of precontemplation, contemplation, preparation, action, maintenance and termination.29 The benefit of MIs is that they focus on creating cognitive and affective knowledge of ACP by addressing individuals’ ambivalence and reluctance about making a decision before encouraging them to act on ACP.34 Therefore, the effectiveness of the MI-based ACP highlights the potential mechanism underlying the nature of motivational interviewing for the development of ACP behaviours.

What this study adds

The use of motivational interviewing in the ACP programme is an evidence-based and theory-driven approach, and when adopted into routine practice, it can help older adults engage in end-of-life care planning. The current review found that using MI strategies to facilitate older adults’ (1) value expression, (2) motivation recognition and (3) supportive decision-making with compassion and education in ACP counselling is likely to be the core elements of effective ACP interventions. In addition, following the four core principles of motivational interviewing, facilitators for ACP conversations give older adults autonomy and prioritise establishing their readiness to engage in end-of-life planning before the emotionally charged communication. This practice was key for the older adults and family members to find these interventions appropriate and meaningful. Notably, the motivational interviewing theory offers a practical structure and communication strategies for delivering ACP conversations. Furthermore, the results suggested that a multisession or single session of 60-minute, structured MI-based ACP counselling can be feasible and effective in enhancing older adults’ ACP. This preliminary implementation evidence provides a valuable design framework for future research in this area.

Fully powered clinical trials on older adults from different demographic backgrounds and clinical settings are needed to establish the validity of the MI-based ACP programmes as a routine practice. In addition, to develop a full picture of the MI-based ACP implementation, additional studies that involve multiple stakeholders to integrate MI-based ACP practice into routine practice, which helps to engage more patients and family members in end-of-life communications at the point of care, should be conducted as well. More research is also necessary on the effects of MI-based ACP interventions on advance directive completion and compliance, cost-effective analysis, end-of-life decisional conflicts and health-related outcomes such as symptom burden and quality of dying across different patient populations and clinical settings.

Strengths and limitations of the study

The search for the systematic review was guided by PRISMA standards in several databases, and the rigour of the included studies was assessed by two evaluators using the MMAT. However, our review has several limitations. First, shared metrics were not used in all RCTs. Second, heterogeneity in the MI-based interventions indicates that the findings of the review should be interpreted within the relevant context. Third, all the studies did not measure ACP-related outcomes, such as quality of communication, decisional conflicts, symptom burden at the end of life and concordance between preferences for care and delivered care. These measurements may provide additional insights into the application and effectiveness of MI-based ACP interventions.


This systematic review reveals that there is strong evidence to support the application of motivational interviewing skills in ACP programmes for older adults. The MI-based ACP intervention is effective in improving older adults’ readiness to engage in ACP in the short term and the completion of advance directives in the long term. Future research in the field of MI-counselling facilitated ACP is needed across healthcare disciplines, especially in nursing, as nurses play a vital role in delivering compassionate care. Such efforts are essential to the integration of the MI-based ACP intervention into the clinical practice for providing quality end-of-life care for the rapidly ageing population.

Data availability statement

No data are available.

Ethics statements

Patient consent for publication

Ethics approval

Not applicable.


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.


  • Contributors TW: conceptualisation, methodology, software, data curation, literature appraisal, formal analysis, writing—original draft. MH: methodology, data curation, literature appraisal, writing—original draft, review and editing. XX: Methodology, writing—original draft, review and editing. HRC: methodology, review and editing. CL: supervision, resources, methodology, writing—review and editing, guarantor.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.