Community-based participatory research (CBPR) is essentially an egalitarian and action-oriented endeavour that strives to equitably involve community members, organisational representatives and researchers in all aspects of the research process. The partners contribute unique strengths and share responsibilities to enhance the understanding of a given phenomenon and the social and cultural dynamics of the community, and integrate the knowledge gained with action to improve the health and well-being of community members.1

Integrating action with research is an integral part of most known definitions of CBPR,1^–4 representing one of the goals and key principles of CBPR, but there has been little effort to synthesise the literature to evaluate to what extent such integration is occurring. Most of the previous reviews have focused mainly on the process of CBPR to document institutional, methodological and other issues and challenges involved in building and sustaining CBPR partnerships and the strategies to cope with them.5^–7 To help CBPR deliver on its promises, it is imperative to take the next step to evaluate the extent to which CBPR involves action to effect community-level changes and what factors facilitate such action. The present review represents such an effort.

One of the previous reviews, the most comprehensive one to date,8 touches upon this issue to provide a brief account of CBPR projects in which research findings were translated into policy change, albeit with fundamental differences from the present review. First, even though they acknowledge the unique benefits of CBPR, the authors of this previous review take the perspective of the prevailing biomedical tradition of health research in evaluating the research quality of CBPR, mainly with respect to research design rigour (such as internal and external validity, intervention fidelity, outcome measures, statistical analysis, and blinding) and formulating recommendations on how to incorporate community involvement in a way that enhances such research quality in CBPR. Although an emphasis on design rigour is fundamentally legitimate, such criteria need to be reconsidered in evaluating CBPR in which scientific pursuit of knowledge is not an overarching goal but integrating action to address the problem investigated is considered an equally justifiable element of scientific endeavour. Second, the recommendations formulated in this comprehensive review are more conceptual than empirical, and thus limited in providing meaningful, qualitative information that can help CBPR move forward. The present review offers such qualitative information.

The scope of the present review is limited to environmental and occupational health areas in the USA. More resources have been made available for CBPR projects in these two areas in recent years, with more public funding provided for CBPR in environmental health9 and with labour unions able to support CBPR to address health problems affecting their constituents.10 It is thus meaningful to take stock of the progress made in these areas. Also, given that occupational and environmental hazards and their remedies tend to be of a collective nature in their origins (that is to say, environmental and workplace hazards are likely to affect all those who live and work in the same physical environment), it is particularly appropriate to examine whether CBPR projects in these two areas have yielded community-level action to address such problems. On the basis of the recent developments in CBPR in these areas, the present review attempts to glean the facilitating factors for CBPR that successfully integrate research and action.

METHODS

Inclusion criteria

To identify references for CBPR to address health disparities in environmental and occupational health, the MEDLINE database was searched using the key phrases “community-based participatory research”, “participatory research”, “participatory action research” (PAR) and “health disparities” in combination with “environmental health”, “environmental justice”, “worker health”, “worker safety” and “occupational health”. A manual search was conducted on all the references thus identified. Searches were not restricted to a specific time period.

Studies that met all of the following criteria were included:

• original articles published in English

• CBPR/PAR studies conducted in the USA

• studies that investigated the effect of environmental or occupational risk factors on human health or those that implemented interventions to address them.

When more than one article was published from the same study, all the articles that documented the inception, process, findings and the use of findings were included.

Key aspects of CBPR to be reviewed

First, I examine whether the research problem originated from the defined community. This is of critical importance in CBPR because a research problem originated from the community may be more likely to reflect genuine and central community concerns and the research project is more likely to be supported by members of the community. The three categories used to determine the degree to which the research problem was posed by the community are whether the project was initiated by (1) community members or groups, (2) academic researchers or (3) mutual agreement between community partners and researchers. These categories—and most other categories used in the present review—are used mainly as “sensitising concepts”,11 rather than strictly operationalised variables, to allow further generation of qualitative information.

Second, I examine whether the collaboration was relatively equitable. Researchers possess inherent power when forming partnerships with community members,12 and they conventionally take the lead role in designing and implementing research. An equitable partnership in the CBPR process was operationalised as community partners’ active participation in decision-making in central research areas, that is to say, over setting research agendas/priorities, developing and implementing the research project and how to use the findings, beyond the customary roles of providing consultation and volunteering or working as supportive research staff.

Third, I evaluate research designs and methods used in the CBPR projects in search of clues for those more compatible with action-oriented CBPR. The categories used to classify research designs are (1) randomised control trial, (2) retrospective or prospective cohort study and (3) cross-sectional study. I also examine whether a given project used quantitative methods, qualitative methods, or both.

Fourth, I examine whether a given CBPR project led to community-level action, which was operationalised as collective efforts to effect community-level changes that go beyond efforts to modify individual-level behaviours of community members. With little prior research to guide the conceptualisation of community-level action, further operationalisation was not attempted.

On the basis of the evidence provided by this review, suggestions are formulated concerning appropriate methodological approaches and funding policies that might facilitate CBPR better integrating research and action to improve the health and well-being of community members.

RESULTS

Thirty-three articles reporting on 20 studies—17 in environmental health and three occupational—met the inclusion criteria. They were conducted in collaboration with low-income communities, mostly African American and/or Latino communities, with the exception of two Native American communities and a Vietnamese community. A summary of the studies included is presented in table 1.

The inception of the studies

Ten of the 20 studies reviewed were initiated by the affected communities: seven of them by community-based organisations (CBOs) or groups (studies 1, 2, 6, 9, 12, 13, 20); one by a labour union (study 11); one by community residents (study 14); and one as an outgrowth of preliminary research initiated by a community member and a nurse (study 5). The rest of the studies were initiated by university researchers (studies 4, 8, 17, 19), by US Environmental Protection Agency (EPA) scientists (study 7), by both a CBO and university researchers with mutual understanding that CBPR was needed (study 3), and grew out of an existing CBPR partnership (study 10). Authors of three studies (studies 15, 16, 18) did not explicitly state who initiated the projects.

It appears that community partners who initiated the research partnerships recognised the strategic importance of collaboration with academic researchers. Communities tend to lack the expertise needed to conduct health research, and academic researchers can lend support to them by compiling credible evidence to validate community concerns. For example, a CBO routinely collected information on the community’s exposure to environmental hazards, but was unable to collect, synthesise and present the data in a way that would effectively influence public opinion and policy, so they sought to collaborate with researchers (study 3). In other cases, community partners chose to work with researchers because of the credibility lent by the latter’s affiliation with academic and other respectable institutions (studies 7, 9). In one study, this validation of community concerns helped spur the community’s readiness to act, further promoting changes (study 11).

Academic researchers reported that community collaboration had been valuable in making the studies possible and valid and in generating credible data. Community partners helped academic researchers to recruit and retain study participants (studies 2, 4, 8, 10, 11, 12), to increase compliance and improve accuracy of reported information (study 14) and to render the research more culturally sensitive and acceptable to the participants and relevant to the local context (studies 4, 8, 10).

Community collaboration

The most common role that community partners played was to provide consultation, as staff of a CBO, as individual community members or as members of a community advisory board or a steering committee. In just one study (study 17), consultation was the only role that community members played in the research process otherwise dominated by academic researchers. In the rest of the studies, consultation was provided in conjunction with other research-related activities.

It appears that community partners in some CBPR projects were in more equitable partnerships than others: they defined research questions, set research agendas or determined research priorities (studies 2, 9, 11, 13, 14), jointly made decisions on the design of research or intervention (studies 2, 5, 6, 7) or how to use the findings to create change (study 1), and approved grant applications or publications/presentations (study 6). In one study (study 11), community partners appear to have been more dominant than the academic researchers, with one of them—a labour union—providing the bulk of the research funding and setting research priorities.

It should be noted that community partners tended to play a more integral role in studies initiated by community members or organisations: eight (studies 1, 2, 5, 6, 9, 11, 13, 14) of the nine (studies 1, 2, 5, 6, 7, 9, 11, 13, 14) studies in which community partners shared power with academic researchers in defining the research problem and developing the project on a more or less equal footing were initiated by the community.

Research designs and methods

Four (studies 4, 8, 10, 15) of the 20 studies reviewed involved randomised controlled trials (RCTs); the rest were observational studies. One of the observational studies was a combination of cross-sectional/case–control design; another a combination of cross-sectional/cohort design; two were prospective cohort studies; and the rest were cross-sectional studies. In most of the cross-sectional studies, both quantitative (ie administering structured questionnaires and conducting exposure assessment) and qualitative methods (observation, in-depth interviews, focus group sessions, etc.) were used.

All of the 10 studies initiated by the community were observational studies: six (studies 1, 2, 6, 9, 11, 20) of them were cross-sectional, using both quantitative and qualitative methods; three involved prospective cohort studies (studies 1, 12, 14); and one (study 5) combined cross-sectional exposure assessment and a case–control design. None of the studies originated from the community involved a RCT.

Research quality of the observational studies included in this review may not rate high with respect to research design rigour. Most of these studies relied on purposive or snow-ball sampling, finding willing study participants through community partners in communities that were willing or sought to work with professional researchers. Cross-sectional studies may have fared worse because of the limitations inherent in the design—that is, the inability to establish temporality between the exposure(s) and the outcome(s). Prospective cohort and case–control studies were designed to address this issue but were not free from the limitations that plague exposure assessment in environmental and occupational health—that is, the difficulty of assessing simultaneous and cumulative exposures to multiple hazards/risk factors and their unknown combined effects.45

In the four studies that involved RCTs, arrangements were made to enhance design validity by randomly assigning study participants to intervention and comparison groups. However, community partners in all four RCTs found the design too complex or simply unacceptable, with the main contention being over the withholding of beneficial interventions from the control group. In response to such community reactions, investigators of three RCTs instituted staggered designs in which what initially was to be the control group became a “treat later” (studies 4, 10) or a “lower-intensity” (study 15) group. Still, field staff members recruited from the community had “concerns about asking low-intensity groups to wait one year before receiving the benefits of the high-intensity intervention” (study 15). These findings indicate that the design rigour involved in the RCT may be fundamentally at odds with the healthy commonsense of community partners that beneficial treatments should not be withheld from those who could benefit from them just to satisfy rigorous research requirements. To the extent that such a design appears unreasonable to community members, it may undermine the trust that the academic researchers worked hard to forge with community members: in one study, the use of a RCT gave the impression to some community partners that “research mattered more” to the academic partners than their well-being (study 10).

Outcomes of CBPR

Did CBPR lead to action?

In 14 (studies 1, 2, 5, 6, 7, 9, 11, 13, 14, 15, 16, 17, 19, 20) of the 20 studies reviewed, CBPR led to action to improve the health and well-being of the community members; for example, to force the cancellation of plans to build industrial hog plants (study 2); to clean a river bank and to identify alternative sources of food to fish caught in contaminated river (study 7); to successfully negotiate a new union contract that reduced workload for hotel room cleaners (study 11); and to force the local housing authority to replace the central heating system at a public housing scheme to improve air quality (study 17). In some studies, the intervention itself—for example, one that involved soil lead abatement for homes (study 16) and another to reduce indoor asthma triggers (study 15)—represented a significant action to improve community health.

Nine (studies 1, 2, 5, 6, 9, 11, 13, 14, 20) of the 14 studies that led to community-level action were initiated by the members or CBOs of the affected community. This may be because community partners who initiated CBPR continued to search for solutions to the problems investigated during the course of the research. In a way, the CBPR that they initiated may in itself constitute a community-level action to address the problems that negatively affected their communities.

It is noteworthy that CBPR projects initiated or prompted by government agencies also led to action to improve the conditions within communities. The US EPA seems to have led the way: in at least three studies reviewed, the partnership of community groups, EPA scientists and university researchers investigated environmental health risks, and developed/implemented interventions to address them or engaged in action otherwise. In study 4, the federal Agency for Toxic Substances and Disease Registry played a similarly pivotal role. In five other studies reviewed, state or city departments of health, environmental protection and housing were integral partners, providing funding and technical assistance in designing and implementing interventions. Only two of the CBPR projects that led to action were initiated by university researchers (studies 17, 19), one of which was initially proposed by the company medical director who became aware of similar health problems plaguing the workforce (study 19). Thus, CBPR projects that led to action are much more likely to have been initiated by the affected community or by government scientists/agencies than by academic researchers.

With the exception of one study that involved a RCT (study 15), all of the studies that led to action were observational studies, most of which had cross-sectional designs employing both qualitative and quantitative methods. While quantitative methods were useful in collecting and synthesising data to estimate epidemiological measures, qualitative methods were invaluable in developing research tools, for example survey instruments (studies 1, 9, 10, 11) and culturally appropriate and contextually relevant interventions (studies 3, 8), and incorporating community knowledge and observations to uncover new sources of health risks previously unknown to researchers (studies 6, 7).Thus, qualitative methods played a more central role in many of the CBPR projects reviewed than in conventional biomedical health research to produce more informative findings.

Did CBPR help to build the community’s capacity?

Enhanced capacity of the community has been reported in several studies in terms of the acquisition of research skills (studies 3, 6, 8, 9, 12, 14) or an improved ability to secure funding (studies 3, 6, 8, 15). However, for CBPR to truly energise and empower communities, a critical consciousness on the part of community members and their increased desire and ability to mobilise their community to effect change are essential. In this respect, building leadership skills is critical.

Two studies stand out in the ways in which community members built and demonstrated their leadership skills. In one study, youth interns became more knowledgeable about air pollution and air monitoring and cultivated and exercised key leadership skills while engaging in public speaking, strategising and educating community members to successfully rally the community behind them (study 9). In another study, some hotel room cleaners—the core members of the research team—emerged as peer leaders to serve on a hotel’s joint labour management problem-solving committees even after the completion of CBPR (study 11). Both of these studies were initiated by the communities and further aided by the related social movements—the former by the environmental justice movement in which the youth leaders were initially trained and the latter by the labour movement that supported the CBPR project.

DISCUSSION

To summarise the findings, community partners in studies initiated by the community are more likely to participate in the CBPR process in a more integral and equitable manner, build leadership skills and use the findings to take community-level action to address the problems investigated. The integration of research and action in CBPR is more likely to occur in observational studies incorporating qualitative methods.

The limitations of the present review should be noted. First, in relying on published articles to synthesise information, I was unable to use unpublished articles or independently verify the information provided by the authors in published articles. Although inherent in most systematic reviews, this limitation may have biased the findings, to the extent that pertinent information was omitted or misrepresented in the published articles. Second, in evaluating the integration of research and action in CBPR, the present review was not able to take into account the circumstances under which the practitioners of CBPR deemed no immediate action to be the best option, mainly owing to little documentation of such circumstances in the literature reviewed.

The findings of the current review offer clues on key facilitating factors for successful CBPR. First, who initiated the study is of critical importance in shaping the process and outcome of CBPR. The needs of the researcher are often guided by theoretical research questions, whereas the needs of the community are practical and action oriented.46 Therefore, it is not surprising that, when the impetus of CBPR came from the affected community, the research problem was more likely to reflect genuine community concerns and the research to be carried out with potential, relatively near-term solutions in mind and to lead to meaningful action.

Second, CBPR can be more successful when supported by social movements. The findings of this review offer examples of CBPR in which community partners who had built their organisational strength in related social movements—that is, the environmental justice and labour movements—came to play an integral role in CBPR to generate tangible benefits for the community. To conduct successful CBPR, therefore, the community itself should be mature enough to offer a vision and the ability to jointly investigate community health issues and to plan a course of action to address them.

Third, CBPR projects that integrated research with action are more likely to be observational studies, most of which incorporated qualitative methods. With the exception of one study (study 15) that led to advocacy, the RCTs reviewed did not lead to community-level action geared towards delivering tangible benefits for the community; none of the RCTs were initiated by the community. The findings involving the contention over RCTs—that, by design, they deny the benefits of interventions to control group members, albeit temporarily in some RCTs—suggest an inherent tension between the rigour required of RCTs and the core principles of CBPR to integrate research and action for the mutual benefit of all partners.

Critics have questioned whether the preoccupation with RCTs is appropriate in addressing research questions posed in community interventions with long and complex causal pathways,47 especially for those that require flexibility to incorporate the perceptions and characteristics of the community.48 RCTs involving community participation have been used mainly in phase 2 translational research that promotes the adoption of an intervention whose efficacy has been established in clinical research.49 50 A case can be made that in action-oriented CBPR a higher priority should be given to making a beneficial intervention available to the community than to testing its effectiveness in the most rigorous manner. Observational designs closely integrated with surveillance51 might be more appropriate in evaluating the effectiveness of the intervention in CBPR.

Qualitative methods are conducive to eliciting in-depth information from the participants without imposing more structured research protocols on them. Also, while quantitative, individual-level data collection is likely to decontextualise the ways in which risk-taking behaviours and environment are generated and sustained, well-designed and carefully conducted qualitative studies can enhance one’s understanding of the problem within the context it is generated.52 These qualities render qualitative methods particularly useful for CBPR in that they are likely to help generate actionable knowledge—better informed by community members living under local conditions—and integrate it into contextually relevant strategies to effect community-level changes.

Fourth, the collaboration among government scientists, university researchers and community partners emerged as a new model of CBPR that effectively integrates research and action. Armed with scientific expertise and access to resources and governmental decision-making, and yet unencumbered by the requirements made by funding agencies of research fund recipients, government scientists may be in a unique position to freely investigate and promote action to better serve disadvantaged communities. Such is the flexibility that should be allowed to CBPR researchers to enable them to conduct truly participatory research that integrates research and action.

CONCLUSIONS

The focus on action in the present review is not intended to downplay CBPR projects that do not lead to action. Even when not generating immediate benefits, CBPR has been an important vehicle in public health research by incorporating community involvement to better identify the community needs and to conduct a more valid and relevant study, while engaging the community in a more or less equitable manner. Such progress should be duly recognised.

Rather, I argue that action-oriented CBPR seeking solutions to the community’s problems though community-level change—one that is perhaps more consistent with the proclaimed goals and ethical principles of CBPR—should be recognised and promoted. Some federal research funding agencies that have been at the forefront of promoting CBPR, most notably the National Institute of Environmental Health Sciences (NIEHS), recognise that, to be successful, research and prevention/intervention strategies must address the concerns of the community residents and that the scientific knowledge that community residents acquire may be used to inform policy and regulatory decisions.53 However, reflecting the currently prevailing paradigm of biomedical research, their funding criteria favour biomedical, hypothesis-driven research questions.54

Advocates of CBPR have already argued that funding and academic institutions need to extend their criteria for research excellence and be flexible enough to incorporate the input of community partners.6 I go further: I recommend that “action” should be a legitimate component of the criteria. It might be helpful for funding agencies to create programmes to support action-oriented CBPR or institute special emphasis panels to evaluate the merits of proposed CBPR projects in their ability to integrate research and action.

When communities engage in CBPR, their primary concern is to generate usable or actionable knowledge.19 55 To help communities generate actionable knowledge and use it to promote the health and well-being of community members, the extent to which research findings are used to effect change should be an integral part of the criteria to evaluate the quality of outcome-oriented CBPR. A new set of funding criteria and priorities or, indeed, a paradigmatic shift in determining what constitutes good public health research is called for to promote community-initiated and action-oriented CBPR.