Article Text

Original research
Use of a 2-year continuing professional development programme to change Japanese physicians’ attitudes to learning primary care: a qualitative study
  1. Masayasu Seki1,
  2. Yasuki Fujinuma2,3,
  3. Masato Matsushima4,
  4. Tatsuhiro Joki5,
  5. Hideo Okonogi5,
  6. Yasuhiko Miura5,
  7. Iwao Ohno5,
  8. Jun Hiramoto5
  1. 1Division of General Medicine, Department of Internal Medicine, The Jikei University School of Medicine, Tokyo, Japan
  2. 2Centre for Family Medicine Development, Japanese Health and Welfare Co-operative Federation, Tokyo, Japan
  3. 3Seikyo-ukima Clinic, Japanese Health and Welfare Co-operative Federation, Tokyo, Japan
  4. 4Division of Clinical Epidemiology, The Jikei University School of Medicine, Tokyo, Japan
  5. 5Division of Nephrology and Hypertension, Department of Internal Medicine, The Jikei University School of Medicine, Tokyo, Japan
  1. Correspondence to Dr Masayasu Seki; masayasu-seki{at}


Objective To evaluate changes in the learning attitudes of primary care physicians.

Design Qualitative study through one focus group interview with the programme’s participants. Analysis of the focus group content using the Steps for Coding and Theorization method.

Setting Japan.

Participants Eight primary care physicians who completed a 2-year continuing professional development (CPD) programme using a problem-based learning (PBL) approach, focused on acquiring the skills needed to practise as primary care physicians in the community.

Results Participants described positive changes in their attitudes and behaviours as a result of the training programme. These changes were grouped into three main themes: ‘changes in learning methods regarding medical practice’, ‘encounters with diverse perspectives and values, and confidence gained from those encounters’, and ‘showing one’s attitude towards learning and its influence on others’. The experienced practitioners participating in this study reported that the programme helped them apply their skills more broadly; for example, searching the literature for psychosocial aspects of practice and engaging more comfortably with diverse perspectives. They reported the positive impact of their learning on their coworkers.

Conclusion A 2-year CPD programme using PBL can influence primary care physicians’ attitudes and learning-related behaviours. Further research is needed to determine which specific aspects of the programme are the most effective and whether the changes in attitudes and behaviours described affect patient care.

  • primary care
  • general medicine (see internal medicine)
  • medical education & training
  • qualitative research
  • education & training (see medical education & training)

Data availability statement

No data are available. Because of the nature of this study, participants did not agree that their data could be shared publicly, so supporting data are not available.

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

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.


  • This study examined changes in learning attitudes (Kirkpatrick model level 3) among primary care physicians and the impact of the changes on other staff (Kirkpatrick level 4) following a 2-year continuing professional development programme.

  • This study had a small sample size and was a single focus group interview conducted in 2018.

  • It is unclear whether changes in learning attitudes among participants have led to improved quality of patient care.

  • Bias may have occurred because of the fact that the programme facilitator was the main interviewer.


Medical education continues from undergraduate education to continuing professional development (CPD), with doctors working in various roles as practitioners, researchers and teachers.1 CPD responds not only to the development of the doctors’ personal professional development but also to the needs of patients, their families and their community.2 Family medicine and primary care are disciplines that provide long-term care centred on people of all ages and situations.3 It is comprehensive, continuing from prenatal care to palliative care.3 No training programme—regardless of its duration or content—can provide the postgraduate medical trainee with all competencies needed for primary care.3 Primary care physicians need to commit to lifelong learning with a deliberate CPD plan to practise with an expert level of clinical skills.4

General practitioners (GPs) in Japan may become family practitioners or hospitalists.5 Approximately one-third of physicians in Japan are in charge of primary care at their own private clinic after 5–10 years of specialist practice training at university hospitals or city general hospitals.6 Many physicians do not have public primary care training but independently undertake learning and training in this area. Unlike physicians in many other countries, they do not need to participate in a specific CPD programme on primary care to maintain licensure.7 The Japan Primary Care Association, established in 2010, is responsible for board certification of senior residents who complete their training programme.5 8 The Japanese Medical Specialty Board (distinct from the Japan Primary Care Association) was newly established in 2017 to manage the certification of GPs in Japan.5 Board-certified GPs were recognised as a new specialist category under a board certification senior resident training programme that began in 2018.8 9 Although an education programme for senior residents is now in place, educational support for veteran primary care physicians, whose training was focused on specific organ systems, is inadequate. Therefore, we consider that the CPD of primary care physicians in Japan should be supported.

In April 2016, we started a 2-year Family Medicine Brush-up Program, which is an interactive CPD programme for primary care physicians with a problem-based learning (PBL) approach. The programme aimed to enable participants to discuss and learn about issues encountered in primary care by studying scenarios based on themes such as those found in online supplemental appendix 1.10 We conducted a qualitative study to clarify participants’ training needs and inform the programme content.10 Three categories of participant statements were established: ‘no standard re-education programme for primary care physicians to respond to changes in the clinical and practice setting’, ‘problems with undergraduate and postgraduate medical education in primary care’ and ‘content of primary care CPD’.10 Through the programme, we aimed to develop the ability to identify problems in the practice of medicine and to continue learning to solve them. Al-Azri and Ratnapalan and Dowling et al reported that a PBL approach can improve physicians’ performance and patient care.11 12 The PBL approach allows learners to actively participate in group activities and helps learners develop into reflective practitioners.13 The field of primary care is fraught with complex problems and uncertainties that make it difficult to arrive at a single correct management pathway.14 We believe that primary care physicians who grow through repeated reflection have a strong affinity with lifelong learning, and for this reason, we have adopted the PBL approach for this programme. The PBL approach we used encompassed working in groups to discuss relevant, real problems. After the 2-year programme that started in 2016 was completed, we considered evaluating the programme to see how the participants had changed. We felt that the completion of the 2-year programme by a number of participants was a good milestone to study the impact of the programme on participants’ attitudes toward learning primary care.

The Kirkpatrick model is used to evaluate educational programmes, including CPD programmes such as our Family Medicine Brush-up Program.1 15 The model focuses on the outcomes of the programme, not just learner satisfaction.16 The Kirkpatrick model was proposed in the 1950s, and a modified model (the New World Kirkpatrick model) was introduced in the 2000s.15 The model consists of four levels.1 13 Level 1 is reaction and satisfaction: do learners respond favourably to the programme? Level 2 is learning measures: do learners acquire the intended knowledge? Level 3 is behavioural change: do learners apply what they learn? Level 4 is results and impact: do the expected outcomes occur?1 15 16

In this study, we aimed to examine the changes that our programme participants experienced in their attitudes towards learning (corresponding to Kirkpatrick level 3) and the impact those changes had on other staff present in the workplace (corresponding to Kirkpatrick level 4). To elicit detailed insights from individual participants, we chose to conduct a qualitative study based on focus group interviews with the programme participants to explore those two dimensions of change and understand how our programme contributed to those changes.


Study design and participants

On completion of the programme (January 2018), we conducted a single focus group interview with programme participants to investigate changes in behaviour that had occurred during the programme corresponding to Kirkpatrick level 3 and to investigate impacts on their immediate colleagues corresponding to Kirkpatrick level 4. Interviews were considered effective for assessing these changes in behaviour and their impacts.1

Eight participants completed the Family Medicine Brush-up Program targeting physicians who had not undertaken specialist training in family medicine and had qualified at least 10 years previously. The interview was conducted at the end of the programme with the eight physicians (A–H, table 1). The results were presented following the Consolidated criteria for reporting qualitative research (COREQ) guidelines for reporting qualitative studies17 (online supplemental appendix 2).

Table 1

Attributes of participants

Data collection

The participants received an explanation of how the interview would be recorded and conducted, and consented to be interviewed. The focus group interview was conducted with the guiding questions: (1) ‘What kind of changes do you have in your awareness and behaviour after taking this programme?’; and (2) ‘Do you notice any change in the behaviour or attitude of staff at your workplace?’

The participants were interviewed in a quiet room undisturbed by daily activities using a digital recorder. Three authors (MS, YF and TJ), all primary care physicians, managed the interviews. In this study, we considered it important to use and analyse the interactions generated by group discussions and adopted the focus group interview method. Focus group interviews are also suitable for investigating attitudes and experiences.18 19 This method is reported to encourage people to talk about difficult content and voice critical opinions.18 19 Interviewers need to establish a positive rapport quickly during in-depth interviews.18 In response to the interviewer’s questions, participants verbalise their own experiences. That verbalisation builds on the interactions and social constructions created between the interviewer and the participant.20 Based on this constructivism recognition, we considered that the authors, who ran the programme and facilitated the participants, should act as interviewers, rather than having a third party involved. We felt that this would better promote group dynamics and elicit discussions among the participants.20 Therefore, the authors acted as interviewers for the focus group interviews. YF had the most experience with interviewing and was therefore the main interviewer, with MS and TJ assisting. These three authors had also managed the programme and facilitated the participants’ learning over the past 2 years.

The interview time was set at 60 min. When one participant responded to a question, several others typically added their opinions. YF asked all the participants questions using the guide questions in chronological order and encouraged participants with relatively few responses to provide additional opinions. In actuality, the interview took 72 min. At that point, the interviewer decided that theoretical saturation had been achieved without any further opinions from the participants.

Data analysis

We analysed the interview records with the Steps for Coding and Theorization (SCAT) method, which is a grounded theory-based thematic analysis approach. SCAT is an analytical method that adds codes in a four-step process, from raw interview data to themes (table 2).21–23 We used this method when conducting a previous study on the needs of participants for the programme.10 SCAT is suitable for the analysis of relatively small samples, such as those used in the previous study, and it was considered appropriate to use SCAT for this study with a similarly small sample.21 23 The SCAT method improves reflexivity by looking back at each step and can be expected to improve the possibility of falsifiability by clarifying the analysis process.21–23 Therefore, the SCAT method was selected as the analysis method of this study. Using the tape transcript, two authors (MS and TJ) independently coded the text for SCAT steps 1–3.21 23 The two authors conferred on conflicting opinions about the content of the code until they reached a joint consensus. Three authors (MS, TJ and HO) independently conducted the coding for SCAT step 4.21 23 The three authors again conferred and agreed on common themes and constructs about the content of the code.

Table 2

Four steps following the Steps for Coding and Theorization method

Patient and public involvement

There was no patient or public involvement in the design or implementation of this study.


Although our programme took place over 2 years with nine participants enrolled, one participant dropped out after only 1 year because of changes in the participant’s medical practice hours. Eight persons completed this programme, and all agreed to participate in the interview. The participants’ interview records were organised into three categories: ‘changes in learning regarding medical practice’, ‘encounters with diverse perspectives and values, and confidence gained from those encounters’ and ‘showing one’s attitude towards learning and its influence on others’ (table 3). This section presents excerpts from focus group interviews on these categories.

Table 3

Themes and constructs about changes in behaviours

Theme I: changes in learning regarding medical practice

This theme was subdivided into ‘search for material and literature’ and ‘psychosocial problems’. The participants talked about how they moved from investigating biomedical problems in their daily practice to investigating problems involving biomedical and psychosocial factors.

Search for material and literature

As primary care physicians, the participants are solving clinical problems related to individual patient consultations. They had few opportunities to reflect on their practice, such as the evidence behind their treatment choices.

I had never given much thought to my routine practice before, but the program made me dig deeper again into questions such as what guidelines said and what kind of literature there was. (Participant B)

Secondary materials were often used to search for evidence to support daily practice and to resolve clinical problems. A change in participants’ learning occurred in their search for primary materials and raw data, such as statistical data about their learning tasks.

Now I search not only for secondary materials but also primary materials. (Participants C and D)

Searching for primary materials was a shift in attitude toward generating opinions based on the participants’ own ideas, to present their findings to other participants for discussion.

All of us in the program gave presentations and had discussions based on statistics we looked up for ourselves. (Participant G)

Psychosocial problems

Participants were experienced in searching mainly secondary materials about biomedical problems. However, they had limited experience in searching material for information about psychosocial problems. Participants’ learning attitude towards problem solving for various clinical problems changed.

I was able to learn systematically, not only biomedical issues but also psychosocial ones, by finding learning topics in scenarios, searching for literature, and considering it logically. (Participant A)

Theme II: encounters with diverse perspectives and values and confidence gained from those encounters

This theme was subdivided into ‘confidence, non-judgemental attitude about other’s opinions’, ‘tolerance of diversity’, ‘no standard re-education programme’ and ‘loneliness about own practice’. Participants who were inexperienced in primary care and operated in isolation at their workplaces described how they had changed after attending the programme.

Confidence, non-judgemental attitude about other’s opinions

When presenting their ideas to others, participants were concerned that they would be judged on whether they were correct or incorrect in their presentations. However, the non-judgemental atmosphere supported participants’ learning.

I felt like I would be judged for my presentation, but there was no critical atmosphere around presentations at all. It was an environment where I could research my learning topic freely and get feedback from everyone. (Participant D)

Tolerance of diversity

The non-judgemental attitude was based on an attitude of respecting individual values and tolerating diversity. These attitudes also encouraged participants to use primary materials and express their own ideas.

I recognized that it’s not really about whether someone is right or wrong, but that maybe there can be all kinds of physicians. (Participant E)

No standard re-education program

One of the reasons participants lacked confidence in their own thinking and were afraid of being judged was that they had not received standard retraining in primary care. They gained knowledge and skills in primary care by attending the programme, but also rediscovered the joy of learning through encounters with diverse values.

I dove right into practicing family medicine without training in it. I had no confidence in myself, and I worried about what I should do and how I should study. The first thing that changed in me through participating in this program was meeting all kinds of physicians and encountering many ways of living. The program reminded me of the truth of how enjoyable it is to learn, even though my daily work as a physician is overwhelming, to think hard about my next own learning topic and compare it with what I actually see in my own patients. (Participant H)

Loneliness about own practice

Another reason for the lack of confidence and fear of judgement was the loneliness that participants felt in their daily practice. They were generally administrators in their own healthcare organisations and had no colleagues to talk to about various issues such as patient care, staff management and their own concerns. Encountering diverse values helped to alleviate this loneliness.

In the clinic, in my position as the manager, even when I get lonely or worry about my relationships with my staff, I have no one to turn to for advice where my clinic is located. The only choice I ever had was to sort things out in my own head. However, by going to a place far away from my clinic and opening up to the people I met there, I learned that I’m not the only one who feels lonely. (Participant H)

Participants felt less lonely, and dealing with diversity allowed them to open up. As a result, the participants realised the depth of their learning.

I have the impression that the level of learning varies quite a bit depending on how much someone opens themself up. (Participant C)

Theme III: showing one’s attitude towards learning and its influence on others

This theme had only one subtheme, ‘active transformation of colleagues’ learning motivation’. Participants saw their own learning change, gained confidence and also shared their learning with their colleagues. Their own development led others to change too.

Active transformation of colleagues’ learning motivation

Even without setting up a formalised learning session, showing a learning attitude is linked to the learning motivation of other colleagues.

My staff told me that seeing me hard at work researching issues between examinations showed them that it’s possible to learn even when you’re busy. They said that when they saw how I studied, it made them want to work harder too. (Participant H)

Showing colleagues the learning content increases their motivation to learn.

I now make it a point to tell all of my staff everything I learned about in this program. I make sure to jot down what I learned and put it up in the meeting room. (Participant A)

Based on the needs of the medical facility to which participants belong and the needs of their colleagues, the sharing of their learning content also led to changes in patient care.

For instance, I have the staff at my clinic actually write out genograms based on what I learn from my patients. I think it’s given my staff the ability to look at things from the perspective of the families and lifestyles of our patients. (Participant A)


The first behavioural change that emerged in the participants’ statements was a change in learning (theme I). One participant stated that their literature searches and logical reasoning had changed regarding not only biological issues but also psychosocial issues. Psychosocial problem-solving is a core competence in family medicine and primary care.24 The participants in our programme have a great deal of practical experience as specialists of different organs and are well versed in literature searches and logical reasoning for biological issues. In addition to this capacity, our results suggested that completing our programme may help participants acquire literature search and logical reasoning capacities for psychosocial issues.

The second behavioural change that emerged was related to encounters with diverse perspectives and values and the confidence gained from those encounters (theme II). As previous studies have found, the absence of re-education programmes often leads to learning in a solitary environment.6 25 In Japan, many private physicians engaged in primary care have solo practices.26 By providing participants with an arena for learning, our programme may have encouraged positive changes in the participants’ attitudes. Providing an arena for learning and forming a learning community may be important, regardless of learning style. Further study is necessary to determine whether confidence, a specific change in the participants’ attitudes, results from the PBL approach.

Similarly, participants also spoke favourably about the effect on diversity of our programme being held away from the locations where they practice. However, for physicians in rural areas, travelling to such programmes is often considered an obstacle to participation.12 Holding programmes online facilitates participation from remote areas. In comparisons of online and on-site education, results are mixed.27 One participant in our study stated that it is difficult to consult with other medical professionals in her own community about issues encountered with patients. For learning about content highly relevant to the participants’ practices, providing a learning community away from the areas where they practise may foster better learning. Previous studies have also shown that traditional face-to-face lectures are preferred by many CPD participants.28 However, during the current COVID-19 pandemic, hosting the programme online would reduce the risk of infection. Additionally, healthcare use in Japan has changed. Aoki and Matsushima highlighted the need to strengthen primary care functions such as support for populations with social isolation and multimorbidity.29 Further research should consider changing the programme to an online format and modifying the primary care learning topics to be covered.

Again on the exposure to diverse perspectives, one participant in our study also noted that discussions regarding the results of learning topics and participants’ practices and values did not lead to a judgemental atmosphere. A positive atmosphere in classes and groups is considered to bring about cooperative learning, while positive discussions and a learner culture are thought to diversify learning, encourage flexible thinking and increase creativity.30 In East Asia, the learning style in medical education is based on Confucian culture.31 The communication style is expressed as ‘cultural reticence’32—a tendency not to actively express what you know or feel.32 The level of learning may change depending on the degree to which someone opens themselves up, and a facilitator of learners’ presentations and discussions may therefore need skills to provide the learners with a safe discussion atmosphere in which the learners’ presentations are not judged as right or wrong and which promotes self-disclosure. Currently, no formal training exists for such facilitators. Going forward, training to help facilitators promote discussion should be conducted while the programme is administered.

The third and final behavioural change was the influence on others (theme III). A previous study suggests that programme participants can promote a positive attitude towards learning in their workplace staff and others around them by demonstrating their own positive attitude towards learning and sharing what they have learnt.33 In East Asia, where Confucian influences are strong, students respect teachers, learn from them and imitate their attitudes.31 Such a cultural background may also improve the learning attitude of the workplace staff. Further examination of the effects of learning programmes will require surveys of the participants’ staff and confirmation of changes in patient care.

The Kirkpatrick model was used to evaluate this programme.16 This model is useful because of its clarity in focusing on programme outcomes and its clear description of outcomes beyond simple learner satisfaction.16 However, this model on its own does not provide educators with a complete evaluation of their educational programmes.16 34 The model has been criticised on the grounds that it does not include intervening variables, such as motivation and learner’s entry level, and the relationship between programme elements and context.16 35 36 It is necessary to investigate the intervening variables that have affected prior learning and then to conduct interviews with the intervening variables in mind regarding changes in behaviour in the study group.

In terms of the three changes in attitude, we will consider whether attending this programme was an effective learning exercise for the participants. The feedback, activity, individualisation and relevance principles are known to be associated with effective learning.37 The points of activity and individualisation were achieved by the use of small groups and a learning strategy in which the learner selects the learning theme using the PBL approach. These points are evident from both the observed change in attitude toward the learning group shown in theme II and the change in learning shown in theme I as a result of the learning environment. In addition, the point of relevance is also satisfied by using a scenario that assumes the site of primary care. This was evident from the fact that the programme became a place to learn about problems faced in clinical practice, as described in theme II. Under the conditions of a solo medical practice and learning environment, and with self-judgement of the correctness of learning tasks, appropriate feedback cannot be obtained from facilitators and other participants. The interview results on theme II suggest that participating with confidence among participants with a diverse set of values in a non-judgemental environment provided sufficient feedback. Additionally, providing appropriate feedback is one of the competencies required as an educator.38 Acting as a facilitator is one of the 12 roles of the educator, and feedback is included in this role. The third attitude change in theme III applies to participants being viewed as role models. Studying in this programme may also enhance participants’ ability to support other learners as a faculty member. By observing how participants behave as facilitators or role models in clinical and learning settings, it may be possible to assess level 4 stages of the Kirkpatrick model for this programme. This aspect could be a subject for future research.

As we aimed for an evaluation that went beyond the satisfaction of taking the course, we chose to address the programme evaluation using dimensions corresponding to Kirkpatrick’s levels 3 and 4. We evaluated one aspect of level 4 of the Kirkpatrick model measured through the impact the practitioner had on their colleagues. However, we did not evaluate another aspect of the impact on patient outcomes. As Samuel and Cervero state in their review, the outcomes corresponding to level 4 of the Kirkpatrick model from CPD programmes are not supported by sufficient evidence.28 Measuring outcomes in terms of patient health and medical economy may be a future research topic for the CPD programme. This would require a survey of individual patients’ illnesses and health conditions, as well as a survey of management conditions. The outcomes should also investigate what changes have occurred in the staff of the medical institutions to which the participants belong, using the participants as role models.


The interview in the present study may not necessarily reflect all changes in the attitudes to learning among the programme participants. It would also have been helpful to include the views of the participant who did not complete the programme.

This study is an analysis of a single focus group interview with all participants who completed the programme. Although the participants are experienced primary care physicians, they do not all have the same level of medical competence and knowledge on the themes of health problems that are addressed in primary care. In addition, the level of their medical skills and knowledge was not verified beforehand. It is possible that changes in the learning attitude of each participant may have been overestimated or underestimated. Future research will require multiple focus groups with larger numbers of participants divided by their subspecialty.

The interview was conducted by facilitators who had been involved with the programme for its 2-year duration. Close involvement in the learning process may have enabled the facilitators to encourage deeper discussion than an interviewer without such involvement. Conversely, the involvement of the interviewers in the learning process may have influenced the discussion about the effective outcomes of the programme, as participants might not have wanted to offend the facilitators.


This study confirmed that participation in our 2-year CPD programme changed participants’ learning attitudes and education-related behaviour. Our results suggest that support of CPD for primary care physicians requires the preparation of a learning community based on diverse values and perspectives and the capacity for facilitation to foster the learning community.

Data availability statement

No data are available. Because of the nature of this study, participants did not agree that their data could be shared publicly, so supporting data are not available.

Ethics statements

Patient consent for publication

Ethics approval

This study involves human participants and was approved by the institutional review board of the Jikei University School of Medicine (study number: 27-277(8162)).


We thank Michelle Pascoe, PhD, from Edanz ( for editing a draft of this manuscript and helping to draft the abstract.


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 MS conceived the study, contributed to the development of its design, received the JSPS KAKENHI grant, collected the data and analysed the qualitative data. YF conceived the study, contributed to the development of the design and interviewed the participants. MM conceived the study, contributed to the design and facilitated the focus group interview. TJ facilitated the focus group interview and analysed the qualitative data. HO analysed the qualitative data and contributed to the design. YM, IO, and JH conceived the study and contributed to the design. All authors contributed to the drafting of the manuscript and read and approved the final manuscript. MS is responsible for the overall content as guarantor.

  • Funding The Family Medicine Brush-up Program was funded by the Jikei University School of Medicine as part of a project entitled 'Building General Practice Capability from Pre-graduate to Life-long Learning – for the Promotion of Clinical Research in the Community' (no grant number). This work was supported by the Japan Society for the Promotion of Science KAKENHI Grant-in-Aid for Young Scientists (B) (grant number 16K19179).

  • Competing interests MM received lecture fees and lecture travel fees from the Centre for Family Medicine Development of the Japanese Health and Welfare Co-operative Federation. MM is an adviser for the Centre for Family Medicine Development Practice-Based Research Network. The other authors report no conflicts of interest.

  • 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.

  • 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.