Narrative based medicine

An action research project to develop group education and support for bilingual health advocates and elderly South Asian patients with diabetes



The objective of this study was to develop bilingual health advocate (BHA)-led group education and support groups for South Asian patients with diabetes, based on a storytelling format. The design and setting were action research in a deprived inner London district.

The research team worked developmentally with BHAs, service users and other stakeholders to assess need, train staff, and set up and evaluate the storytelling groups. Stories were chosen as the main vehicle for education because of their universal appeal, memorability, holistic perspective, ethical dimensions, and potential for multiple interpretations.

Key challenges in the project were: high burden of need; multiple stakeholders; a shifting baseline of policy directives; frequent staff changes; resource constraints; and resistance of top and middle management to new service models. These difficulties made progress slow, and several early initiatives failed to take root. However, we succeeded in developing a nationally accredited training programme (‘Storytelling for group work in health and care’) for BHAs, and in supporting the graduates from this course to set up two groups for South Asian patients with diabetes—one in a hospital and one in a community centre. The patient storytelling groups, once established, were popular and well evaluated. Blood glucose levels in regular attenders fell, but changes did not reach statistical significance.

An action research approach allowed us successfully to develop an innovative training programme for BHAs and support them in developing and mainstreaming new service models. BHA-led diabetes storytelling groups, supported by health professionals, are feasible and popular with non-English speaking Asian patients. Their impact on diabetes outcomes should now be formally evaluated in clinical trials. Copyright © 2005 John Wiley & Sons, Ltd.


Ethnic Asian people have a three to six-fold excess of diabetes compared to whites. Age-standardised morbidity from diabetes related complications is higher than in whites.1 Reasons for this excess are complex, and include: socioeconomic deprivation (including the impact of displacement); genetic risk; discrimination and racism; cultural and religious influences on behaviour (including exercise and food choices); communication; and access to services.

At a stakeholder conference organised by the British Diabetic Association in October 2000, people from minority ethnic groups identified the three most important priorities for diabetes services as: poor cultural awareness (leading to lack of understanding and sensitivity by NHS staff); the unfulfilled potential for partnerships between NHS and community organisations; and poor communication between staff and ethnic patients. Solutions suggested included active recruitment of staff from minority ethnic groups, more opportunities for working in partnership, and extending the role of the interpreter or bilingual health advocate (BHA).1

Our previous research had showed that positive behaviour changes (such as giving up smoking) in British Bangladeshis are almost always attributed by them to a story told in an informal setting by another Bangladeshi.2, 3 The unique advantages of the story form in education are shown in Table 1.4–7

Table 1. Unique benefits of stories in professional and patient education4–7
• Stories are a natural and universal form of communication
• Stories create engagement through metaphor, rich imagery, suspense and other literary devices
• Stories are sense-making devices—i.e. they allow people to make sense of events and actions and link them to past experience
• Stories embrace complexity. They can capture all the elements of a problem
• Stories offer insights into what might (or could or should) have been, and hence consider different options and their likely endings
• Stories have an ethical dimension, and hence motivate the learner
• Stories occur in both formal and informal space. Hence, story-based learning can occur from a very wide range of sources
• Stories are performative. They focus attention on actions (and inactions) and provide lessons for how actions could change in future situations

The present study aimed to develop and evaluate BHA-led storytelling groups for South Asians. The study was funded by Diabetes UK, approved by the local research ethics committee, and delivered through Social Action for Health, a local voluntary sector charity.

Method and findings

We used the action research cycle (Figure 1),8 in which researchers work with local health care staff, service users and other stakeholders to identify a focus of action, gather data to illuminate the problem, take action aimed at a change, and gather further data to evaluate progress.9 Because the findings of one phase in the action research cycle are used to feed into the method for the next phase, we have not followed the conventional separation of ‘methods’ and ‘results’, but rather presented our work as an unfolding story with successive phases.

Figure 1.

The process of action research (adapted from Somekh)8

Getting started

The project required a lengthy set-up phase in which we mapped existing diabetes and ethnic health services with a view to developing partnerships with organisations (both NHS services and the voluntary sector). The study was undertaken in a deprived, under-resourced, multi-ethnic inner city area. While there were exemplary diabetes services in some areas, major service gaps were apparent. For example, at the time the project started (1999), there was no community based diabetes specialist nurse and virtually no community dietetic or chiropody services. Hospital clinics were overcrowded and advocacy services severely stretched. There was duplication of provision in interpreting and advocacy services, with five different organisations providing these services to the public sector.

Against this background, we spent several months building relationships, presenting ideas for the storytelling initiative, and seeking specific mainstream projects that might link with our own. We found that motivation to work with us was high, but capacity was often low, with the top management of organisations distracted by ‘must-do’ policy directives and middle managers struggling with profound staff shortages and low morale.

Developing the BHAs' training

With the co-operation of advocacy managers in the different organisations, we collected data on the BHAs' background and experience, and their perceived training needs. We used three data collection methods: semi-structured (qualitative) interviews with the BHAs' managers; a structured questionnaire for BHAs; and our own participant observation in ‘taster’ group sessions (see below).

In retrospect, we seriously underestimated the challenges of this phase of the project. BHAs come from a very diverse background (their education level, for example, ranging from five years of formal education to MSc degree). Their knowledge and confidence about health in general and diabetes in particular varied widely. They dealt with huge numbers of patients with diabetes, yet few had had any training in diabetes and none had been formally trained in patient education. They rated their own knowledge and confidence in dealing with diabetes as poor. There is no such thing as a ‘locum’ BHA, no routine ‘cross cover’, and no contractual right to study leave for this staff group. The nature of the training (group based storytelling) required a minimum of six (and ideally eight or more) participants at each session—leaving a significant dent in local service provision for half a day a week. Furthermore, advocacy managers in some organisations, who had been raised on a much more didactic and expert-led model of training, saw the course (‘sitting round telling stories’) as unconventional and rather amateur.

Having met with so many initial obstacles, we decided to ‘show rather than tell’ the potential benefit of a learning set for BHAs based on storytelling about patients. We offered two ‘taster’ sessions (with generous funding reimbursement to organisations who sent staff to these), in which the focus was for BHAs to identify, and reflect on, their learning needs in relation to diabetes. Using small group brainstorming and ‘Post-it’ exercises, we established a list of potential topics and the level of interest in each topic, and used this to develop learning objectives and a syllabus for future sessions. These taster sessions were very well attended, and enthusiasm for the course (especially the storytelling approach, which was seen as non-threatening, different and fun, but also as an effective way of learning) soon spread by word of mouth. Organisations began to sign up to it.

We established a weekly learning set, with the aim of: (a) promoting self-esteem, motivation and a professional approach to practice; (b) increasing participants' knowledge of, and confidence in, the principles of diabetes care; and (c) ‘modelling’ the principles and practice of adult learning,10 effective small group work,11 and the storytelling (‘case history’) approach to exploring patients' needs.4 The programme consisted of 10 structured three-hour sessions, each on a theme identified by the BHAs in the taster sessions. Participants shared stories in small groups (of three or four) and then fed back selected case studies to a larger group who reflected on patients' unmet needs and professional educational and support needs.12 The session leader (TG or AC) adopted a facilitative, non-didactic approach in which stories, not textbook ‘facts’, were used as the basis for learning.

Because we were keen for the BHAs to acquire a formal qualification, we asked them all to write up one story each week as ‘homework’ on a structured template (Table 2). These written accounts and commentaries formed an important source of data for our own learning as course organisers.

Table 2. An example of a story written up by a BHA on the standard template
The story of AB
A healthy 10-year-old boy, AB, discovered a few months ago that he had diabetes. One day he lost a lot of weight, felt thirsty, frequently went to the loo and his lips were very dry. His Mum who cares for him—his father is not around as they are divorced—took him to the hospital and his blood sugar was tested. It was very high and the doctors said he should go on a careful diet (as well as insulin). So his Mum always gives him boiled food and now he is fed up and frustrated and he feels tired most of the time. His Mum is very worried and so are the doctors at his lack of enthusiasm and lack of concentration and interest in his studies. He knows how to take his insulin himself. His Mum taught him. Sometimes, AB sneaks down chocolates and his blood sugars get very high and he needs to stay in hospital and his insulin dose is increased. But he is scared about insulin and blood testing in hospitals. He doesn't want to go but his Mum forces him.
Could anything have been done differently?
The boy's mother should not have been so strict with his diet by limiting it to only boiled vegetables. She should vary the diet and adapt recipes with, for example, lower oil content, less salt and sugar which is always found in large amounts in Asian Bangladeshi foods
What questions or issues does this story raise?
It is a very emotional upheaval for a 10-year-old boy to cope with diabetes and a big responsibility for a young person going to start secondary school soon. The mother feels guilty and occasionally when her other son eats chocolate she allows AB to have it as well
What are the learning points?
The diet, most of all, should be varied. Although the boy is not overweight, he should exercise regularly to improve his well-being and raise his energy levels so that his studies, which are very important, improve. If he follows a regular exercise programme and controls his diet, it will give him an advantage when he is elderly. It will also become enjoyable to a young, growing boy

The BHAs' learning set was very well attended. Themes raised included: a deep commitment to their most needy and vulnerable patients; high personal workload; lack of specific training for key aspects of their role; lack of protected time for reflection or discussion about difficult and stressful cases; lack of clarity of roles and responsibilities; and (in some) a profound feeling of professional isolation.

At the end of the course, using short questionnaires, we documented high levels of satisfaction with the course in general and the ‘storytelling in groups’ approach in particular. We also showed a significant increase in self-assessed knowledge and confidence in supporting patients with diabetes. In liaison with a local community education and training organisation (‘Making Training Work’) we registered the 12-week course as a nationally accredited training option with the London Open College Network (‘Storytelling for group learning in health and care’). Further details of the aims, objectives and syllabus of the course are given in the trainers' workbook, available as a book.13

The first learning set was restricted to 13 female Bangladeshi BHAs, but in subsequent courses we included Gujurati, Persian, Somalian, Turkish, Arabic, and Chinese speakers of both genders.

Setting up storytelling groups for patients

About two years after the start of the project, we had a cohort of BHAs who were confident in the principles of diabetes care and keen to apply group based storytelling techniques with patients. However, most of them held low status positions in their organisations, and had neither the authority nor the resources to set up and run such groups. On three occasions when BHAs we had trained had successfully set up patient groups, the post holders moved on within a few weeks (for personal reasons such as maternity leave, or because staff shortages in other parts of the organisation created needs that were perceived to be more pressing).

In retrospect, our belief that training the BHAs would lead in a linear and unproblematic fashion to patient groups being set up was naïiacute;ve. Because of repeated disappointments in the early phases of this part of the project, we decided to focus intensive support on one or two groups where the host organisations were supportive of user groups in principle but faced logistical problems. We worked with middle and senior managers to gain further organisational buy-in to the project, so that the BHAs could be given appropriate autonomy, support and resources.

In these two settings (a local district general hospital for Gujarati speakers and a community based lunch club for Bangladeshis funded by social services), both top management and middle management support was strong, and the BHAs were able to move forward with their initiative. We describe one of these in detail below.

Storytelling groups in action: one example

St Lucia's (pseudonym) is a community centre in a deprived estate in Stepney. There is a weekly lunch club, attended mainly by elderly Bangladeshi women who speak little English. Many are strict Purdah observers. Around half had known diabetes. The BHAs booked a side room in the Centre and passed a message by word of mouth that a diabetes group was starting for those who wanted to stay on after lunch. AC and NB attended all the sessions, with NB acting as interpreter; TG attended around one meeting in four.

The group was popular from the outset, and membership grew from eight to 42 in 18 months. The storytelling took a very different format from the BHAs' learning set (in which we had been able to organise the group and facilitate storytelling and reflection in a highly structured and timetabled way). The service users, in contrast, were not willing to sit around in a circle taking turns telling stories, nor were they willing to be organised or facilitated by the research team. Discussion often took the form of animated arguments between participants, which we were often called in to resolve. Multiple conversations could often be heard at once, and women wandered about the room, coming and going as they pleased and sometimes bringing friends or grandchildren with them.

However, despite the casual (and sometimes frankly chaotic) appearance of the group, most women attended consistently and the focus of discussion was invariably diabetes—especially the ongoing stories of regular attenders. On one occasion, a BHA from a different organisation came into the room to ask for the women's input to the design of an incontinence education project. She was politely told by the women that this was a diabetes group and they would not be able to help until their discussions were complete.

The focus of the group was on action rather than words. For example, many women attended to check their blood glucose levels on our meter, even though we had provided them with their own meters and shown them how to use them. Interestingly, they compared their own glucose level with other people's levels rather than with their own previous level. To assist this process, we made a colour chart annotated with smiley (and sad) faces so that numerical values could be readily interpreted by illiterate group members. This enactment of blood glucose testing led to a lot of reflection and attempts at sense-making of the general format ‘why is my result worse than hers?’ (and, very rarely, ‘why is my result worse than it was last week?’). In some cases, though not as often as we had expected, women sought lifestyle advice from the better-controlled individuals.

Overall, the group's mean glucose levels did not change, since new attenders were often poorly controlled. Our sample was systematically biased in this respect, since many people were initially sent to the group by local primary care staff concerned about their poor control, so we had a steady influx of poorly-controlled new members. Random glucose levels in longstanding attenders decreased over time, but the change did not reach statistical significance.

When the group was well established, we asked the women's consent to tape record a focus group discussion, which we translated and transcribed. (Incidentally, our initial request to tape a session was turned down, but a few weeks later when we told the women that our funding would soon be finished, they spontaneously asked us to bring in the tape recorder because they wanted the funders to hear their enthusiasm for the group.)

In the evaluation, led by NB, we asked the women what they would tell a friend about the group. The responses were unanimously positive and included: ‘If someone has diabetes and doesn't know what to do, they should come to the group first, and they will learn what to do.’

At the time of writing, the St Lucia's group and one other group based in a community hospital setting have continued short-term with mainstream NHS funding and diabetes specialist nurse support. The longer-term sustainability of these groups, however, remains unknown.


This study has demonstrated that setting up BHA-led education and support groups for South Asians with diabetes is feasible but not straightforward. BHAs lack a formal career structure. Their knowledge, experience, confidence and study skills vary, and their support needs are often considerable. Their low status in health care organisations means that training for the individual must be linked with change at strategic and managerial levels within the organisation.

Story-based group education for South Asian patients looks and feels very different from group work in a professional setting and from traditional ‘diabetes group education’. There is resistance to formal facilitation and to a pre-set agenda. Stories are often fragmented, and may be enacted rather than told. The shared social event of reacting to the story and discussing different interpretations and possible endings can lead to both reflective learning and empowerment to take action. A key outcome of the group in the eyes of the participants was ‘knowing what to do about diabetes’.

This study was not designed to formally test the impact of the storytelling group on blood glucose control or other biomedical or psychological outcomes. Our preliminary results suggest, but do not prove, that diabetes control in regular group attenders tends to improve.

A theoretical limitation of patient groups led by non-clinicians is the danger that incorrect and even dangerous clinical advice might be shared in the group. We did not encounter any such examples in the groups we studied, and indeed we were surprised how adept the group was at identifying areas of confusion and seeking the advice of one of the research team. Our findings suggest that hands-on support by a suitably qualified health professional (ideally, a diabetes specialist nurse) is advisable.

Previous studies of group education and support in type 2 diabetes14–18 have included people from many different ethnic groups. However, they were not designed to investigate the specific cultural or linguistic needs of particular minority groups nor to evaluate the impact of the group process (as opposed to the educational package delivered to the group). The role of BHAs in group based health education more generally is virtually unexplored. Yet a MORI survey in 1994 showed that nearly three-quarters of advocacy service users had not had any kind of group based health education with the BHA. Of the quarter who had, 80% had found it ‘very useful’ and another 14% ‘fairly useful’.19

Our qualitative findings suggest that the mechanism by which group participation achieves positive learning outcomes is not principally via acquisition of knowledge (though this is clearly important) but by providing a forum in which participants can negotiate the meaning of knowledge and by prompting action. In both the BHAs' learning set and the user groups, knowledge was repeatedly discussed, reframed and challenged by the group, and only then made meaningful for the participants.

In conclusion, this study has provided unique insights into the feasibility of setting up narrative based diabetes education and support groups for both BHAs and service users, and into the process of learning, empowerment and change in such groups. This innovative service model could now be evaluated formally in a randomised controlled trial.

Conflict of interest statement