Challenges of participatory research: reflections on a study with breast cancer self‐help groups

Abstract

Objective

To review and discuss issues related to participatory research, as they apply within the arena of cancer control.

Design

A participatory research study with breast cancer self‐help groups is referred to for description and discussion purposes. That study employed primarily individual and group interviews to assess benefits and limitations of self‐help groups.

Settings

Four breast cancer self‐help groups in Ontario communities provided the core involvement in the participatory research project.

Results

The values and practices of mainstream academic research often conflict with those of research emphasizing participation and control of communities under study, leading to a variety of challenges for the latter approaches. Practical constraints faced by many community groups have important implications for participatory research approaches.

Conclusions

A balance needs to be found for participatory research within cancer control – one that ensures that the core aims of participatory research are maintained, while simultaneously acknowledging the various challenges that make a fully participatory project unrealistic. Steps can be taken to achieve a workable balance.

Introduction

In recent years, social scientists have shown increasing interest in participatory research. This trend has begun to find expression in cancer research, where there is now considerable discussion about the optimal ways to involve patients. Unfortunately, this discussion is taking place largely without reference to what has been debated and learned in the social sciences. In this paper, we intend to review the existing body of knowledge about participatory research and use our experience with breast cancer self‐help group research to comment on issues that make the approach challenging.

Although there are currently numerous interpretations of participatory research, a recent comprehensive report, commissioned by the Royal Society of Canada, defines it as ‘systematic inquiry, with the collaboration of those affected by the issue being studied, for purposes of education and taking action or effecting social change’. ¹ Participatory research is characterized less by any specific methodologies, and more by the philosophical commitment to determine methods on the basis of needs, as defined by the community under study. ²

What is currently referred to as participatory research was influenced primarily by two traditions. The first, which actually coined the term ‘participatory research’, was rooted in community development initiatives in the third world, ³ , 4, ^–5 and drew on Paolo Freire’s ⁶ work on adult education and empowerment. The second (earlier) tradition, was coined by Kurt Lewin ⁷ as ‘action research’. It emphasized the link between knowledge development and action, and often included increased participation by communities under study. ^{8 , 9} For many years these two traditions developed in isolation from each other, but recently interchange has increased, and writers have focused more on similarities. Indeed, Green and co‐workers ¹ argued that enough convergence has now occurred for both traditions to be subsumed under the unifying umbrella of participatory research.

Secondary influences on modern participatory research are many and include the following:

(1) Agricultural initiatives, like Rural Rapid Appraisal and Participatory Rapid Appraisal; ¹⁰

(2) Feminist research; ¹¹ , 12, ^–13

(3) Applied anthropology; ^{14 , 15}

(4) Community psychology. ¹⁶ , 17, ^–18

Proponents of all of these approaches have expressed concerns about reducing the distance between researcher and researched, and in increasing the substantive involvement of ‘subjects’ in studies about themselves.

Early contributions to the participatory research literature were steeped in the recognition of inequities and oppression in third world countries. ^{19 , 20} In this context, the value of egalitarian relationships between the researched and researchers had strong currency and the importance of maximal community control over research process and product was continuously stressed. In essence, proponents were more intent on making the case for a new, transforming model of research and less on the issues actually faced in applying the model in local settings. This early conceptual and political work could easily lead to an impression that more control for the group under study was always better, and that transfer of power from the researcher to the researched was relatively unproblematic.

As researchers have gained experience with participatory research, and the field has broadened by virtue of converging influences, there has been less rhetoric about the absolute value of maximal control and participation by communities under study. Instead, writing has begun to reflect a greater focus on what actually occurs in practice, and more detailed discussions about the possibilities and limitations inherent to participatory approaches. ^{21 , 22}

Results from the Royal Society of Canada study showed that most contemporary participatory research lies in the middle of a continuum. At one end there is the ideal of equal roles and responsibilities for researchers and the researched while at the other end there is only token participation from selected community representatives. ¹ In a similar vein, Borkman & Schubert ²³ categorize participatory research projects as having ‘slight’, ‘moderate’ or ‘extensive’ participant power and ownership. What is becoming more obvious is that much research described as participatory has achieved the goals of community participation and control in selected and partial ways. This reality can be understood from different perspectives.

One view is that there is a trend towards less responsible adherence to fundamental principals of participatory research. Cornwall & Jewkes ²¹ write, ‘Much of what passes as “participatory research” goes no further than contracting people into projects which are entirely scientist‐led, designed and managed.’ Many current ‘participatory’ projects are, indeed, missing important ingredients or principles that most researchers agree define the territory of participatory research. ¹ Although some of this ‘failure’ may be due to researchers’ poor planning or cynical appropriation of the increasingly popular terminology of participatory research, also relevant is that many researchers appear to have modified their expectations of how much participation and control is possible for communities under study. An example of such changing attitudes is provided in descriptions of the relationships between researchers and study participants. Whereas earlier descriptions focused on equality, more recent writing has emphasized the unique strengths that each partner brings to the endeavour.

The trend to more critical and reflective writing about participatory research is consistent with a ‘problematizing’ of the field. A recent review ²¹ of points of conflict or concern includes the following: scepticism and lack of interest on the part of communities; time constraints of participants; the heterogeneity of ‘community’ and associated conflicts within communities; unintended negative consequences for participants; ill‐preparedness of researcher training for relinquishing control; and conflicts between expectations of funding agencies and/or academic settings and those of the participatory research approach. The authors of this review ²¹ note that it is important to recognize potential issues and pitfalls if they are to be avoided or handled, and if the participatory research agenda is to remain viable in the future.

Participatory research and self‐help groups

By definition, self‐help groups are concerned with participant empowerment and democratic decision‐making. ²⁴ Given these values, it comes as no surprise that self‐help groups have tended to be suspicious of traditional research approaches that emphasize researcher control and standardized methods. ² Indeed, it has been argued that adherence to these mainstream approaches may alter the nature of self‐help group experience and render research results meaningless. ²³ , 24, ^–25

The limited self‐help research that has been successfully conducted has typically featured careful negotiations between researchers and group participants about issues of power. ²⁴ Indeed, some of the groundbreaking participatory work in North America was conducted by Rappaport and his co‐wokers ²⁶ with self‐help groups for the mentally ill. In the United Kingdom, similar sensitivities have been shown in relation to cancer self‐help groups. ²⁷ More recently, Borkman & Schubert ²³ have argued for the importance of a participatory approach which highlights self‐help groups as ‘culture creating and transmitting groups that develop their own world views and accompanying experiential knowledge of their problem and its resolution’. The participatory approach, with its emphasis on conceptualization of issues for study by research participants, is particularly relevant in that self‐help group members’ perspectives on their problems often differ from the perspectives of professionals or the lay public. ²⁴

Research with breast cancer self‐help groups

We decided, for reasons argued above, that a participatory approach would be most appropriate for research with breast cancer self‐help groups. Very broadly, the goals for our research included to help identify health benefits for self‐help group members, and to illuminate some of the strengths and limitations of breast cancer groups, as well as some of the variations among groups. But, just important to us as these outcomes, was to explore the ways in which we could engage self‐help groups in the process of research that they would largely define, own, and come to see as valuable.

The study on which this paper is based ran from 1993 to 1997. The primary work was with four breast cancer self‐help groups in Ontario communities. We interviewed 24 women from those four groups initially and then again 1 year later. ²⁸ Several group interviews were conducted in each community and 13 women from other breast cancer self‐help groups were also interviewed. Interviewees included older and younger women, long‐term survivors who were well and women with advanced disease, women who were new to the self‐help group and those who had helped form the groups. We are very pleased with how the study went. Thus, our desire to be critically reflective about participatory research stems not from disappointment, but from observations we have made through a primarily positive process.

Members of our research team were not explicitly trained in participatory research. Our interest in a participatory approach developed over time as we watched and attempted to facilitate the development of cancer patient empowerment vis a vis the health‐care system. ²⁹ , 30, ^–31 Our concerns about the need for patient voices in clinical decision‐making and in health‐care planning led us to be more reflective about our own research practices. It was only as a result of these experiences that we turned to the literature on participatory research and began to implement principles we found in that literature. Because we are relatively new to this work, we acknowledge that some of the tensions and issues experienced within our project, and ‘problematized’ in this paper, may have been due to limitations in expertise. However, most were, we believe, related to structural issues that are unavoidable and that have been experienced by other researchers in the field.

In what follows we critically review our experience in undertaking participatory research with breast cancer self‐help groups. To determine the content of this review, our research group held a series of meetings to review process notes collected over the duration of the project. Rather than try to be comprehensive, we will highlight those issues that we have been most affected by and about which we have the most to say.

Issues related to funding

Until very recently, Canadian organizations responsible for research funding of health‐related issues were entirely dominated by research scientists. Lay or consumer involvement in setting policy directions was seen as inappropriate, as was involvement in reviewing study proposals. However, changes in policy have occurred over the past few years, stemming primarily from the 1993 National Forum on Breast Cancer. At the Forum, breast cancer survivors argued that they be involved in the design and implementation of the breast cancer research agenda. As a result, representation was secured on the management committee of the new Canadian Breast Cancer Research Initiative. More recently, survivors have participated as non‐voting members on research review committees of the National Cancer Institute of Canada. Similar advances in patient participation have occurred in both the United States and the United Kingdom.

Despite the changes noted above, the control of research funding decisions remains predominantly in the hands of scientists and administrators. This has implications for participatory researchers, in that there is less valuing of participation built into the process of proposal generation than might be the case if there were more consumer control. Zarb, ³² in writing about disability research, makes an interesting observation about this issue in terms of his own research activities. He notes that in two studies where funding was provided by a consumer organization there was much more consultation and shared decision‐making than in two other studies funded by traditional scientific agencies. His conclusion was that the ‘material relations of research production’ were very relevant to the level of participation of the disabled community.

In beginning our search for funding, we suspected that traditional funding agencies would not consider cancer survivor participation as a high priority issue. Hence, we spent a good portion of our proposal providing a rationale for the participatory approach with self‐help groups. We explained that many details of the research methodology could not be specified ahead of time because of the need for on‐going discussion with the groups. We also compromised and included in the design of the study some procedures and measures that traditional researchers would recognize as valid and duly ‘rigorous’. This subsequently created some tension with self‐help group members, who weren’t entirely happy about catering to prevailing scientific values. In the final analysis, two traditional funding agencies for cancer research rejected the proposed project. The primary reasons were lack of sufficient control in the design and lack of specificity about project details. It seemed that the need to leave issues open to negotiation with self‐help groups could not be accommodated within the then prevailing criteria for recommending proposals. Fortunately, we later received funding from the Health Promotion and Programs Branch – Ontario Region, Health Canada, which was sympathetic to the community partnership aspect of the proposal and, indeed, encouraged us to make this the primary feature of the study. The point to be made, though, is that as long as research funding remains primarily in the control of the scientific community, few participatory research studies are likely to be funded. This impression is supported by findings that many participatory researchers disguise their projects in ‘respectable quantitative clothes’ when seeking funding. ¹

A second issue relevant to funding relates to who controls research funds once they are granted. It has been suggested that more researcher accountability is ensured when funds are held by the group under study. ³³ This was how things were initially designed in our study. Funds were to flow to the Burlington Breast Cancer Support Group, a self‐help organization that had agreed to be official sponsor of the study. But it quickly became apparent that the job of money management would be just one more burden for an already overwhelmed volunteer group. Steps were taken for the financial work to be assumed by the research team. The Burlington group was grateful for the relief and it was easier for the research team to manage day‐to‐day operations. Practical considerations had overtaken the well‐intentioned goal of participant control.

Issues related to participant engagement in conceptualizing the research

The idea for this study was born when a prominent breast cancer activist approached one of us (RG) at a meeting and suggested the need for research to prove to physicians and scientists how helpful self‐groups are. This began a discussion, carried on with all the groups in the study, of how difficult (i.e. impossible) it is to provide proof of the benefits of self‐help groups sufficient to satisfy the scientific community. This is because the gold standard of traditional research is the randomized, controlled design using standardized assessment instruments and standardized interventions. When we explained what would be involved with this type of research, breast cancer survivors agreed that it would violate the spirit of self‐help and the needs of individual women. They welcomed the suggestion that the focus should instead be on what groups themselves want to know, and on user‐friendly methods which might help them to learn. However, despite all of our conversations about these issues over the past 2 years, we continue to encounter intermittent requests for study findings to prove to reluctant health professionals the value of self‐help groups. Because of this experience (and related others), we think it is unreasonable to expect most members of self‐help groups, having little or no research background, to develop sophisticated critiques of research issues and/or to lead the way in conceptualizing research design.

Another issue related to study conceptualization is how the involvement of the community under study fits with the process of writing proposals and submitting them for funding consideration. As discussed above, we found ourselves unable to provide sufficient detail in our proposal to satisfy funding organizations. In traditional research planning, projects are developed in detail by researchers with minimal, if any, consultation, with the potential study population. But we understood the need for substantive and on‐going consultation appropriate to a participatory approach. It was a classic ‘catch‐22’. We needed to spend significant time consulting with groups in order to arrive at a mutually acceptable design and methodology. Yet we had no funds to support the costs of such consultation, nor could we easily justify to the groups the value of time spent if it were to transpire that our proposal was rejected. Thus, in addition to details that could only be worked out over the course of the research process, we were unable to resolve some issues that might have been addressed up front. As it turned out, most of the final details of study procedures and methods were revised, in consultation with our newly created research advisory committee and the four community self‐help groups, after we received funding. We were in the fortunate position of having the support of our funder in making these revisions.

There is one further issue related to the early phases of the research endeavour. It is important not to underestimate the impact of unbalanced power relations between researchers and members of the group under study. In our case, despite our best efforts, this power differential manifested itself in self‐help group members agreeing too readily with our original plan, which included administration of a self‐report questionnaire to newly attending group members. They provided input into the development of the questionnaire package and agreed to pilot test it. But the questionnaires did not get distributed. It took some investigation to uncover that the women were uncomfortable with, and thus resisting, the procedure to which they had agreed. With further discussion, it became apparent that the questionnaire unduly interfered with usual group processes. The women were, thus, relieved when we suggested we look for other ways to gather the required information.

Issues related to instrument development, ongoing communication, and data gathering and analysis

Learning from our early difficulties with piloting the questionnaire packages, we took a less directive approach to designing questions for individual and group interviews. We generated some initial questions and then asked self‐help group members on our advisory committee to critique them and generate additional questions that would be of interest to them. They made helpful suggestions and were pleased to be engaged with the process.

We tried to have on‐going dialogue with self‐help groups by keeping in telephone contact with group leaders, by making periodic visits, and by bringing key individuals together in Toronto for planning meetings. However, we quickly came to realize that more contact is not always better from the perspective of self‐help group members. Because they are so busy with their regular meetings, advocacy projects, and assorted fundraising and ‘business’ concerns, they are typically cautious about being over‐burdened by the demands of research. We, thus, had to be careful to space our contacts out and carefully negotiate the conditions of our consultations. This contributed to their sense that we were respectful of their needs and wishes. An additional issue related to on‐going communication about the research is that commitment to the research process differed between groups and changed for individual groups over time. For example, while members of one group eagerly awaited interview transcripts to review, another group complained that transcripts were too long and boring to read through.

Group valuing of the research project seemed to relate to the interests and enthusiasms of a core group of women in each community. Indeed, one of the concerns of our research team was how information about the research process sometimes did not get beyond the core group. We were disturbed to find in interviews with several women on the periphery of groups that they knew little about the on‐going research process. The failure of this information to be disseminated to the larger group by core members suggests that our emphasis on on‐going communication was not entirely embraced by the groups.

We were aware that some participatory researchers had hired members of the community under study to act as research assistants and do the bulk of data gathering. ³⁴ Thus, at the beginning of the project, we entered into discussions with a breast cancer survivor and self‐help group member about possibilities of her working as a research assistant. Unfortunately, this became entangled with existing conflicts between breast cancer groups and among prominent individuals within the breast cancer advocacy movement. Self‐help group members expressed concerns about possible failures of objectivity that might be unavoidable because of our candidate’s position within activist politics. Consequently, we changed our plans and hired a doctoral student in psychology with clinical experience working with cancer patients. She was well received by the self‐help groups, and proved an excellent choice because of a combination of strong interviewing skills and personal sensitivity to women’s issues.

We continue to believe that there are important advantages to supplying members of a group under study with the opportunity to be engaged in interviewing and other data collection. We also think the value of ‘objective’ researchers has historically been overstated (and, as poststructuralism suggests, may be fundamentally misleading). Nevertheless, our philosophical position was no match for the political realities we faced and we were forced to adjust to fit the circumstances.

Although we were aware that the participation of those under study in the analysis of data may add to the richness of interpretations and has the potential of increasing participant control, in our study this was impractical. The community self‐help groups in our study were scattered around the province, making frequent visits to the groups difficult. But even more important – as mentioned above – was our concern not to overtax the capacities of groups. Although all groups have, for the most part, been excited and engaged by the research process, they have given as much as they are willing and able. We did send interview transcripts to individuals and asked them to review and make any changes they might think appropriate. We also sought the feedback of our advisory committee, and of individual groups, about the validity of preliminary analyses the research team conducted. However, members were not asked to participate in the actual data analysis. One reviewer suggested that such an omission might constitute an unconscious paternalism on our part. Certainly this is possible, but we were at least conscious of the issue, and of walking a fine line between how much it was fair to ask (and the anxiety for them in saying no) and being unduly protective of them.

Issues related to ownership of research products and dissemination of study results

We have produced research products that have been clearly under the direction of study participants. These have included a paper targeting physicians about the benefits of self‐help groups, ³⁵ as well as group stories for each of the breast cancer self‐help groups. The idea to produce group stories developed from our initial interviews with the four community groups. When we discussed this among members of the research team, we envisioned narratives with nuances of both the benefits and the limitations and/or struggles experienced by the groups. But when this was discussed with our advisory committee, comprised of self‐help group members, it was quickly apparent that they had a different vision. They were interested in stories that would inform their local community about the benefits of breast cancer self‐help groups. Essentially, they wanted a product to help market their groups. This was fine with us, but only when we were able to reassure ourselves that the more difficult issues we knew existed would find expression in other research products. Four thousand copies of stories for each group were printed, with the text shaped and approved by group members, and cover designs customized for groups. It seemed a great success, and made the potential usefulness of participatory research much more tangible for group members.

There is one caveat to the success of the group stories. One of the groups was particularly adamant about stressing the inadequacies of physicians’ response to women with breast cancer. Unlike the other groups, they wanted to minimize historical details around the initial formation of their group in favour of an emphasis on the devastation they felt at not getting the support and information they needed from the medical community. While the women’s unhappiness with local physicians was clearly connected to real and important problems, we were initially uncomfortable with exclusively focusing on these problems in the booklet. We felt that background material would be helpful and of interest to most readers, and perhaps shouldn’t be neglected. We also wondered whether the harsh criticism of doctors might alienate some of the women who would pick up the booklet. Would physicians agree to distribute it to newly diagnosed women? We carefully raised these issues with group members, but their position was firm. So we supported what they wanted, while asking to soften a few of what we perceived to be inflammatory phrases. We also explained to our funding organization, which expected to review all project products, that the group had rejected their and our comments. We were fine with all of this, and were feeling rather pleased with our ability to set aside our biases and follow the group’s wishes. But then we later heard that there was some unhappiness with the final version of the story among a few group members. They felt that we had watered down their criticism of physicians. One woman commented that this was typical of professionals, always protecting each other. We wonder what to make of this. To what extent does it point to a failure of process on the part of the research team? To what extent might it reflect tensions within the group about what is important and how their group will be defined? Beyond the specifics of this situation, what is highlighted is how difficult it is to gain universal consensus both within community groups and between those groups and research teams.

Particularly in the early participatory research literature, emphasis was placed upon the rights of the group under study to control research products. However, this failed to acknowledge the situation of professional researchers. Because of the political and structural parameters of academic institutions, ³⁵ researchers need to be seen as being productive. We are typically recognized when our names appear on grant proposals and journal articles – that is, when we take ownership for research products. In our work with breast cancer self‐help groups, we have been very conscious of having to balance the needs of the groups with our own needs for vocational survival and career advancement. It seems to us better to acknowledge these latter needs as legitimate rather than to pretend that we do not have them.

In order to meet our needs for research products (that we can be seen as being responsible for) we maintained primary control over writing directed towards the academic and professional communities. For these products, we tried to ensure substantive research participant input, especially from our advisory committee of survivors. However, that input was most often on a general level. We did not seek extensive reviews of drafts of academic papers by individual members of self‐help groups, although members of the advisory committee have been involved. This was a conscious choice, fuelled by our desire to avoid potential public relations disasters. Was our process too high‐handed? Did it violate principles of participatory research? We don’t know the answers to these questions. In retrospect, perhaps we were overly cautious in our approach. But it did allow us to complete the work and, in the final analysis, we have had only positive comments from self‐help groups about the published papers.

Lessons learned and suggestions made

Drawing on our successes and mistakes, we offer the following tentative suggestions for researchers interested in participatory research.

• Seek funding for participatory research projects only from funding agencies that acknowledge and value participatory principles. This context is changing and broadening over time, as shown by the recent funding of a larger participatory project by the National Cancer Institute of Canada.

• Assume that each aspect of the study will have to be negotiated separately with the community under study, as opposed to assuming that the community either always wants involvement or control (or never wants involvement or control).

• Attend to the reality of power imbalance between researchers and the community under study, making special efforts to obtain community input in ways that extend beyond responding to researcher plans or agendas.

• As much as possible, have frank and clear discussions with key members of the community, both at the beginning of the overall project and with the introduction of each new phase. Make researcher needs explicit in relation to academic/professional products. Develop guidelines for the level of engagement required if community members are to be included as co‐authors on academic papers.

• Be prepared to have the project agenda shift according to the needs of the community. Build in budgetary and time accommodations to develop products, or undertake side projects, that are identified by community members as being of prime importance (even when they are secondary to the main research objective).

• Anticipate the need for a higher than usual time commitment for core members of the research team. Success in a participatory project is predicated on building and sustaining relationships with members of the community. Relationships take time.

Conclusion

Participatory research has made a valuable contribution to the research community by challenging the historical tendency to ignore and discount the rights of communities under study. But the actual working out of this approach is complicated and sometimes difficult. The identification and naming of issues that are frequently encountered by participatory researchers is important to ensuring a more practical and valid understanding of what the approach actually entails. A balance needs to be found in the field between ensuring that the core aims of participatory research are maintained, while simultaneously acknowledging that ‘it is nearly impossible to create a 100% participatory project’ (Les Voakes, quoted in Green et al. ¹ ).

Acknowledgements

The authors wish to thank Dr Kathryn Church for her exceedingly helpful comments on various drafts of this paper. Thanks are also due to Ann Wray Hampson, a key member of our Advisory Committee, for taking the time to review and comment on the paper. For their on‐going commitment and contribution to this project we are grateful to Ann Griffith, Mercedes Sandhar, Joan Davey, Sandra Lockwood, Linda McCorkindale and Wanda Risso. And thanks also to Pat Kelly for the initial prompting to undertake the study.

The study discussed in this paper was funded by the Health Promotions and Programs Branch – Ontario Region, Health Canada.