Jump to ContentJump to Main Navigation
Understanding and Using Health ExperiencesImproving patient care$

Sue Ziebland, Angela Coulter, Joseph D. Calabrese, and Louise Locock

Print publication date: 2013

Print ISBN-13: 9780199665372

Published to Oxford Scholarship Online: May 2013

DOI: 10.1093/acprof:oso/9780199665372.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 24 November 2020

Story gathering: collecting and analysing spontaneously-shared stories as research data

Story gathering: collecting and analysing spontaneously-shared stories as research data

Chapter:
(p.60) Chapter 7 Story gathering: collecting and analysing spontaneously-shared stories as research data
Source:
Understanding and Using Health Experiences
Author(s):

Trisha Greenhalgh

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199665372.003.0007

Abstract and Keywords

This chapter, illustrated by a worked example of diabetes self-management, describes an approach in which illness narratives are collected to extend understanding of the lived experience of the condition and inform the design of education and support programmes. In the example, we describe how informal, unstructured story-sharing groups were held for 82 participants in six different ethnic languages, facilitated by bilingual health advocates. Some but not all stories were translated by the facilitator or by members of the groups. Researchers sat in on groups, made contemporaneous notes and typed these up afterwards. We collected and analysed around 300 story-fragments, producing a new theoretical model of how the story form enables people to make sense of abstract diabetes knowledge and apply it to their own lives and lifestyles. Particular ‘storylines’, which were common across the ethnic groups studied, gave the biomedical tasks of self-management both social meaning and moral worth, allowing them to be embraced (to a greater or lesser extent) in personal and family routines. Using this empirical example, we consider the strengths and limitations of naturalistic story-gathering as a data collection method in qualitative research

Keywords:   Story-gathering, Naturalistic research, Self-management, Patient education

Background

Stories are rich research data, but the stories people tell in formal research interviews differ from those told spontaneously in real life. Collecting stories in ‘naturalistic’ settings has both strengths and limitations (see ‘Discussion’). In this chapter, we describe a new approach to naturalistic story gathering using diabetes self-management as a worked example.

We make sense of our lives by retrospectively and prospectively ‘storying’ our experiences (Bruner, 1990). Stories are defined by chronology (unfolding over time), characters, setting, trouble, and plot (Aristotle, 1996). In illness narratives, ‘trouble’ comprises pain, disability, disfigurement, loss of independence, and so on (Frank, 1995). Coping with chronic illness and minimizing its impact requires sense-making and biographical work (Strauss, 1975). In cross-cultural illness narratives, personal experience is nested within a wider narrative of personal, family, and community—which for the immigrant patient may include displacement, loss of social status, economic hardship, and perhaps discrimination following immigration.

A ‘textbook’ narrative is structurally coherent with a beginning, middle, and ending, but real-world narratives may emerge as a bricolage of incomplete, inconsistent story fragments (Riessman, 2008). Stories are image rich, action-packed, and laden with emotions—hence are potentially powerful vehicles for educational messages (McDrury and Alterio, 2002).

Whilst narratives have often been used to explore illness from the patient’s perspective, there are far fewer examples of such narratives being used to inform a grounded and nuanced approach to patient education. The question arises: how can we provide education that is useful and meaningful to people with a condition unless we know how that condition affects both their identity and the practicalities of daily life for them? Yet as policy in chronic disease management shifts from (p.61) a focus on episodic, clinician-led care to one that emphasizes self-management, the importance of understanding the patient narrative increases.

‘Self-management’ is a contested term—referring (in different literatures) to particular behaviours and lifestyles; coping emotionally and practically with illness; participating in illness-related networks and social movements; and challenging inadequate provision of services (Hinder and Greenhalgh, 2012). In the study which forms the basis of the example in this chapter, we sought to explore the stories told (and enacted) by people with diabetes and use these stories to design self-management education and support.

Our research questions included: (1) what stories do people from minority ethnic groups in socio-economically deprived areas tell about their diabetes and efforts to live with it? and (2) how can these stories inform the design of culturally congruent self-management programmes?

The ‘naturalistic story-gathering’ approach

Building on earlier work to develop and validate a group-based story-sharing intervention (Greenhalgh et al., 2005), and as part of a randomized controlled trial to test that intervention (Greenhalgh et al., 2011a, 2011b), we collected stories shared in peer support groups held in six different ethnic languages. The study was based in a deprived and ethnically diverse part of London with high diabetes prevalence. Eighty-two people had been referred for diabetes education and randomized to story-sharing groups. They ranged in age from 25 to 82 years; 20 were on insulin and 22 were male. All were first- or second-generation immigrants from African Caribbean (seven participants), Bangladeshi (23), Tamil (11), Punjabi/Urdu (34), and Somali (seven) communities. Each group was homogeneous for language and ethnicity (with some overlap of Pubjabi and Urdu), but mixed gender (though groups sometimes spontaneously split into men and women for discussions of gender-sensitive topics).

Because our early research had identified informal stories (such as gossip or hearsay) told by an equal-status peer (as opposed to ‘standardized’ or ‘expert’ accounts) as the trigger for behaviour change (Greenhalgh et al., 1998), the intervention was intentionally unstructured and informal, and didactic teaching by staff was strongly discouraged. Following Hawe et al. (2004), we recognized that complex interventions may include flexibility in application in different settings yet must have well-defined core components. These were:

  • Sessions must involve spontaneous, informal, and unstructured story sharing (participants should be encouraged to tell whatever stories they want about their condition, in whatever order, with no stipulation of what is ‘important’ or ‘legitimate’).

  • (p.62) Sessions must be facilitated non-directively by a non-clinical professional or volunteer trained in the sharing stories model.

  • Input of clinical professionals (doctors, nurses, dieticians, etc.) must be in the form of a response to the stories shared by group participants.

In practice, the groups were run as follows. Each group had ten to 12 participants; sessions lasted around two hours (the venue was booked for three hours to allow for a ‘social time’ and to accommodate late arrivals); they were held two-weekly for six months.

Participants were encouraged to identify a theme for a forthcoming session, which was typically a topic that had emerged in discussion the previous session (e.g. ‘feeding the family’, ‘medication’, ‘dealing with doctors’). They were asked to bring personal stories on the agreed topic next time, and the principle of participant-led, story-based topic discussions was followed as far as possible.

On arrival, participants had ten to 15 minutes (sometimes more) ‘social time’ before being asked to get into ‘buzz groups’ of two to three and share their stories. After about 20 minutes, the facilitator invited them to regroup in a plenary format and asked ‘Does any group have a good story to share?’. This often allowed a story with important learning points to emerge, after being ‘tried out’ in the safety of the buzz group. Informal discussion on this story was encouraged, with a focus on the practicalities and moral challenges that the story raised. Whilst the generic design for the group session was around the ‘told story’, in practice many stories were enacted with visual ‘props’—for example, pills are often passed round when discussing medication; or samples of rice when discussing glycaemic index. If stories prompted action (e.g. stories about exercise prompted the group to try out chair-based or stretching exercises), this was encouraged.

Thus, each story-sharing group was led by a bilingual health advocate (BHA) with some input from a clinician—most usually a diabetes specialist nurse but depending on the chosen session topic, a chiropodist, dietician, general practitioner (GP), pharmacist, or physical trainer was invited to attend either as well as, or instead of, the nurse. The BHA facilitated the group in a non-directive manner, encouraging the sharing of personal stories about diabetes and its impact, which s/he interpreted and invited the health professional to comment on. The BHA sought to work actively with the health professional to ensure that the focus of discussion was on ‘explaining why X happened in the story’ rather than ‘giving a talk on X’.

To capture the stories, we used quasi-naturalistic story gathering (collecting naturally-occurring talk during real social interaction), though groups were not strictly naturalistic since they were set up as part of a trial. (p.63) Members of the research team attended a total of 65 group sessions. We sat in the group circle and made field notes on group dynamics, issues discussed, emotions displayed and actions observed, which we typed up immediately afterwards. To be captured, stories had to be translated by the facilitator without interrupting the group discussion or spontaneously retold in English by a group member. This required the researcher to exercise flexibility and judgement, not least because—as in any multilingual group where informal stories are being shared—the narratives overlapped; different fragments of the narrative were sometimes told in different languages; and each narrative attracted a response and/or a counter-narrative (that is, a story which illustrated the opposite point to the one being made in a previous story). There was no self-evident or technical way for the researcher to decide which stories to record in a notebook or which to ask the BHA or another group member to translate in real time. Rather, we used the narrative ‘tool’ of simple curiosity, selecting those stories that surprised, intrigued, delighted, or shocked us.

Analysing naturalistic narratives

Stories can be analysed as pieces of text (for example, the researcher could go through a printed story with a highlighter pen and mark up particular words or phrases, but take no account of the narrative features such as whether the story is a ‘comedy’ or a ‘tragedy’, who the ‘heroes’ and ‘villains’, are and what constitutes the ‘plot’), and/or they can be analysed as literature with attention to all these features and more. In particular, a narrative analysis goes beyond the text itself and asks questions about the narrator and the act of narrating.

In our example, we undertook a preliminary thematic analysis mainly to gain familiarity with the data. For this, we used Ritchie and Spencer’s framework method to chart data on a spreadsheet and iteratively refine a taxonomy of issues raised (Ritchie and Lewis, 2003). In a subsequent, more detailed, narrative analysis, we identified stories or story fragments and for each one identified, we considered literary features such temporality, characterization, contextualization, and emplotment (e.g. how events and actions were juxtaposed in a particular way to imply motive or causality). We also considered the context in which a particular story was told, the narrator’s (sometimes urgent) desire to tell this story to the group, and the reciprocal desire of the audience to hear it, participate in its interpretation, and contribute to its resolution. To achieve this, we drew on Bakhtin’s dialogical approach and Riessman’s notion of storytelling as performance (Bakhtin, 1981; Riessman, 2008). (p.64) Thus, our interpretive analysis took account of both the product of storytelling (narrative as noun) and its process (narrative as verb).

The messy and heterogeneous nature of narrative data

The researcher who seeks to collect a tidy set of ‘textbook’ stories from naturalistic story gathering, each with a clear beginning, middle, and end, will be disappointed. As suggested earlier, narratives shared in groups come in many different shapes and sizes, and are not infrequently incomplete. The messy nature of the data may be further compounded by the lack of verbatim recordings (for example, if many stories are told in quick succession the researcher may capture only parts of some of them).

Many ‘stories’ captured in our handwritten field notes of the group sessions were thus short and fragmentary, and some were offered in response to another fragment on the same topic told by another participant (hence could only meaningfully be analysed in conjunction with those fragments). Other stories were longer and had a more definitive narrative structure; these sometimes became the focus of the day’s meeting with the facilitator and other participants asking further questions, which shaped the unfolding narrative. Some longer narratives (e.g. a person’s efforts to get a referral from a reluctant GP) became recurring topics in successive meetings, with the narrator offering (or being asked for) an update by the group. Some topics (e.g. efforts to lose weight) began as one person’s story but became collectivized as other group members offered similar or contrasting ones.

Occasionally, a narrator would claim the floor by standing up or presenting an artefact (e.g. a hospital letter or packet of tablets) and offering a story about this. Many stories were presented emotionally—with laughter, tears, or apparent anger. These contrast markedly with the fictional vignettes or images typically given in educational materials, in which self-managing patients are depicted somewhat blandly as doing context-independent tasks such as ‘eating healthily’ or ‘exercising’. In contrast, real-life accounts depict the work of self-management as having social meaning and moral worth. Indeed, this was why they came across as stories, however fragmentary, and why group members chose to afford significance to them.

Finding ‘storylines’ in narrative data

Every story is unique, but at a higher level of abstraction several stories may share a common thread or ‘storyline’. A challenge in analysing a large dataset (p.65) of naturalistic stories is to identify these storylines and reflect on their implications (see Chapter 5).

Our preliminary ‘framework’ analysis revealed seven self-management challenges facing the person with diabetes: knowledge, diet, exercise, medication, foot care, self-monitoring, and attending check-ups. These map closely to the domains covered by previous studies of diabetes knowledge, and to self-management education guidelines (Funnell et al., 2009). But an important finding of the study was that these challenges were never presented purely as ‘tasks’ but were framed within particular storylines, which were common to all cultural groups studied. Next, we describe two of these storylines in detail to illustrate the point that storylines add much to the research findings over and above the specific topics talked about. The other six storylines are illustrated briefly in Table 7.1 and presented in more detail elsewhere (Greenhalgh et al., 2011b).

Storyline 1: becoming sick

(I was) panicky, fear gripped me…I didn’t think I would die but I was very frightened—I had heard about diabetes in the family. I cried for 3 days because I was frightened and everyone said to me awful things would happen to me, strokes and I don’t know what…I am still frightened. (Participant in Tamil group, T1.)

In all of the groups, the first session was spontaneously spent with each participant in turn describing the experience of diagnosis. Most had felt shock, feelings of hopelessness, loss, and isolation—reactions which link to Frank’s notion of ‘deep illness’, affecting all life choices and decisions (Frank, 1995). Many told stories of others who had developed terrible complications (e.g. amputation) and anticipated that they would too. Some people revealed fear of stigma, ill-defined fears, and/or specific knowledge gaps. One of the commonest questions was whether to conceal or share the news of the diagnosis, and with whom.

Storyline 2: becoming a practitioner of self-management

They [various tablets for diabetes] are like a herd of cattle for me and I remember each one like a different cow. (Participant recently immigrated from Somalia where she had lived a nomadic life herding cattle, group S4.)

Around half of all the stories exchanged in the groups involved either seeking or providing tacit knowledge relating to diet, physical activity, self-monitoring, and managing medication, thereby operationalizing self-management within the social and practical constraints of their daily lives. Strategies employed were often grounded in particular cultural symbols or gender roles, but the (p.66)

Table 7.1 How the findings from a diabetes story-gathering study informed diabetes self-management education

Storyline

Implication for education

1. Becoming sick

Participants who have not come to terms with the diagnosis may not engage with self-management

People should have an opportunity to tell ‘the story of my illness’

2. Rebuilding spoiled identity

Moving on from the diagnosis involves biographical work linked to practical actions

Educators should understand how physical and practical measures can help identity-rebuilding

3. Becoming a practitioner of self-management

Self-management is learnt partly through watching, listening, and building social networks

Educators should promote tacit knowledge exchange through stories

4. Living a disciplined and balanced life

Many great religions uphold the principles of diabetes self-management: looking after the body, resisting excess and indolence, and observing rituals

Programmes should reflect the fact that some people make sense of their lives through religion

5. Mobilizing a care network

The ‘self-managing’ patient is usually sited in a care network comprising family, community, and health professionals

Input from family and peers should be supported flexibly

6. Navigating and negotiating in the healthcare system

Access barriers and power differentials can hamper diabetes management

Programmes should address accessing healthcare and communicating with staff

7. Managing the micro-morality of lifestyle choices

Self-management often poses small-scale moral choices, especially how to spend limited family income

Educators should draw on stories to explore moral dilemmas

8. Taking collective action

Stories may be enacted by the group as well as told orally. Enacted stories have great potential to energize and motivate.

Educators should recognize and support the ‘social drama’ aspects of group work

Reproduced with permission from Greenhalgh, T. et al. New models of self-management education for minority ethnic groups: pilot randomized trial of a story-sharing intervention. Journal of Health Services Research and Policy, Volume 16, Number 1, pp. 28–36, January 2012, Copyright © The Royal Society of Medicine Press Ltd, UK, DOI: 10.1258/jhsrp.2010.009159.

(p.67) task of ‘becoming a practitioner’ was a common challenge for all. Talk about the practicalities of self-management was led by a few experienced members in each group, but others joined in with time. For example, a man who had recently been widowed kept quiet for several sessions whenever the subject of shopping or cooking came up, but eventually he began to tell stories about trying out recipes learnt from the women in the group.

The exchange of tacit knowledge through stories resonated with Lave and Wenger’s notion of community of practice—an informal group in which learning is not about amassing socially disembodied facts but about ‘legitimate peripheral participation’ (watching, listening, developing an identity through practice, and building the social networks that form the scaffolding for tacit knowledge acquisition) (Lave and Wenger, 1991).

Application of naturalistic story gathering

It is worth reflecting on what this novel approach (which produced a messy and difficult-to-analyse dataset) achieved over and above what might have been achieved through more conventional qualitative research (such as one-to-one semi-structured interviews or focus groups). By collecting and analysing narratives from people with diabetes from minority ethnic groups in a deprived part of London, we confirmed previous biomedical research on knowledge and skills needed for self-management (Funnell et al., 2009) and the sociological notion of illness narrative as biographical work (Strauss, 1975; Bury, 1985). We also made a number of original contributions both to the clinical literature on diabetes and to the methodology of narrative research.

Cross-cultural research traditionally seeks to identify and explain differences between ethnic or social groups. One of the most striking findings in this study was the common storylines occurring in all the groups despite a wide range of different cultural backgrounds, religions, and the inclusion of both first- and second-generation immigrants. We found that commonalities across groups, especially the triple jeopardy of immigrant status, material poverty, and low functional health literacy, were more striking than the differences between them.

By synthesizing these overarching storylines into a more-or-less holistic picture of what it is to live with diabetes, we were able to draw a closer link between biomedical and sociological research on self-management of chronic illness. Previous literature on this topic had been somewhat polarized, considering either the biomedical skill set (can the patient demonstrate diabetes knowledge and perform particular self-management tasks?) or the illness narrative (can the person make sense of their illness, restore coherence and balance in their (p.68) life and address moral questions?) (Greenhalgh, 2009). Through naturalistic story gathering followed by an explicitly narrative analysis of the dataset, we concluded that these two perspectives are not incommensurable but two sides of the same coin: behaviours expected of the self-managing patient come to be applied in practice by being included within particular illness storylines which give them meaning and purpose.

The study allowed us to refine a complex intervention for education and peer support in minority ethnic groups with diabetes (Table 7.1).

An individual’s ability to manage their diabetes is largely unrelated to their level of factual knowledge, a finding that confirms other work in low-literacy minority ethnic groups (von Goeler et al., 2003; Carbone et al., 2007; Tang et al., 2008). Participants appeared to recognize the distinction between abstract knowledge (‘knowing-that’), which they said they gained from health professionals, and tacit or practical understanding (‘knowing-how’), which they felt they gained from sharing stories. One participant put it like this: ‘We learn the facts from you [health professionals] but we learn the meaning and how to do things from one another’.

Naturalistic story gathering has potential to inform a radical new curriculum for patient education and support. The approach may be particularly pertinent when the patient’s ‘real life’ includes hardship, material constraints, and barriers of language and literacy.

Our findings suggested the need for two radical developments in the design and delivery of self-management education for people with chronic illness. First, programmes should be oriented more explicitly towards capturing and bearing witness to the overarching storylines through which the experience of chronic illness, and the work of self-management, is patterned. Importantly, these storylines are discernible when stories are told informally and spontaneously. Second, diabetes education should be more firmly grounded in the pedagogical distinction between know-that knowledge (easy to codify and standardize) and know-how knowledge (tacit, applied, and context-specific, hence impossible to codify or standardize). ‘Structured’ self-management education will cover only the former (Funnell et al., 2009); sharing stories informally is one way of covering the latter.

Strengths and limitations

Naturalistic story gathering has methodological strengths and limitations. Creating an environment in which people feel free to share stories presents challenges to the researcher who wishes both to preserve the naturalism, and record the content of, the encounter. Of course, the very presence of a researcher or (p.69) health professional (especially perhaps if not from the same ethnic group as the participants) may also influence the nature or format of the stories told.

Our own decision not to record the group sessions was made because, in early pilot work (in which we held ‘drop-in’ groups for people with diabetes), group members were emphatically opposed to tape recording of the sessions. This may have been because some attenders were illegal immigrants or had other reason to distrust the use of recording equipment. Other researchers may decide to audio or video recording a storytelling group if they feel that they need to transcribe detailed verbatim accounts for data analysis (and translation).

Whatever the chosen approach to capturing stories as research data, there is a trade-off between the authenticity of the story (for example, its informality, spontaneity, and original language) and the accuracy with which it can be recorded—the control exercised by the research participants and the extent to which health researchers are visible in the process. In a multilingual storytelling group the researcher may have to rely on translations (and, sometimes, a re-telling).

However, selective translation (and the fact that the researchers in our study were monolingual) could be seen a positive feature, since it allowed the group to control and even negotiate what was to be shared as ‘data’. Our data suggest that people talked very freely when they were assured that the researcher would only hear the stories if they or the BHA chose to translate them.

Conclusion

While this conclusion needs to be confirmed in other studies, we tentatively conclude that naturalistic story gathering may be especially suited to capturing the process by which tacit knowledge is exchanged among peers. Furthermore, the approach led us to a new hypothesis: that ‘non-compliance’ with self-management tasks is more often linked to a lack of ‘know-how’ knowledge (a problem for which the exchange of tacit knowledge through stories may offer a solution) or to the ambiguous social meaning and moral worth which particular tasks may hold (a problem for which group negotiation of meaning and morality may offer a solution) than to lack of factual, know-what knowledge.

The empirical study that forms the basis of this chapter illustrates some but almost certainly not all the potential advantages and disadvantages of naturalistic story gathering in clinical research. Many methodological questions (such as the trade-off between recording and not recording stories verbatim, and the use or not of bilingual researchers) remain unanswered. Indeed, it (p.70) could be argued that these questions are unanswerable in the abstract, since the ‘right’ method can only be decided in relation to a particular research question, context, sample, and budget. We would be most interested to hear from other health researchers who have used the naturalistic approach in narrative research.

Acknowledgements

We thank group participants for their stories and permission to share and learn from them; clinicians, managers, advocates, and volunteers who participated directly in the study; external advisors who served on our steering group; and the NHSSDO Programme who funded the study.

References

Bibliography references:

Aristotle (1996). Poetics, trans. M. Heath. London, Penguin.

Bakhtin M. (1981). Discourse in the novel. In: Holquist M (ed.) The dialogic imagination. Four essays by M.M. Bakhtin. Austin, TX: University of Texas Press.

Bruner J. (1990). Acts of meaning. Cambridge, MA: Harvard University Press.

Bury M. (1985). Chronic illness as biographical disruption. Sociology of Health and Illness, 4, 167–82.

Carbone E.T., et al. (2007). Diabetes self-management: perspectives of Latino patients and their health care providers. Patient Education and Counseling, 66, 202–10.

Frank A. (1995). The wounded storyteller: Body, illness, and ethics. Chicago, IL: University of Chicago Press.

Funnell M.M., et al. (2009). National standards for diabetes self-management education. Diabetes Care, 32 (1), 87–94.

Greenhalgh T. (2009). Patient and public involvement in chronic illness: beyond the expert patient. British Medical Journal, 338, 49.

Greenhalgh T., Helman C., and Chowdhury A.M. (1998). Health beliefs and folk models of diabetes in British Bangladeshis: a qualitative study. British Medical Journal, 316, 978–83.

Greenhalgh T., Collard A., and Begum N. (2005). Sharing stories: complex intervention for diabetes education in minority ethnic groups who do not speak English. British Medical Journal, 330, 628–33.

Greenhalgh T., et al. (2011a). New models of self-management education for minority ethnic groups: pilot randomized trial of a story-sharing intervention. Journal of Health Services Research and Policy, 16, 28–36.

Greenhalgh T., et al. (2011b). Storylines of self-management: narratives of people with diabetes from a multiethnic inner city population. Journal of Health Services Research & Policy, 16, 37–43.

Hawe, P., Sheill, A., and Riley, B. (2004). Complex interventions: how ‘out of control’ can a randomised controlled trial be? British Medical Journal, 328, 1561–3.

Hinder S. and Greenhalgh T. (2012). ‘This does my head in’. Ethnographic study of self-management by people with diabetes. BMC Health Services Research, 12, 83.

(p.71) Lave J. and Wenger E. (1991). Situated learning: Legitimate peripheral participation. Cambridge: Cambridge University Press.

McDrury J. and Alterio M. (2002). Learning through storytelling in higher education: Using reflection and experience to improve learning. Auckland: Routledge Falmer.

Riessman C.K. (2008). Narrative methods for the human sciences. Thousand Oaks, CA: Sage.

Ritchie J. and Lewis J. (2003). Qualitative research practice. London: Sage.

Strauss A.L. (1975). Chronic illness and the quality of life. St. Louis, MO: Mosby.

Tang Y.H., et al. (2008). Health literacy, complication awareness, and diabetic control in patients with type 2 diabetes mellitus. Journal of Advanced Nursing, 62, 74–83.

Von Goeler D.S., et al. (2003). Self-management of type 2 diabetes: a survey of low-income urban Puerto Ricans. The Diabetes Educator, 29, 663–72.