BEHAVIOR change interventions designed to improve women's health that are derived from traditional empirical approaches may be unsuitable and unacceptable to women. In contrast, behavioral solutions that integrate women's local knowledge, incorporate local resources, and capitalize on community assets may be more effectively implemented, acceptable, and sustainable. Researchers are recognizing that collaborative methodologies that include local women in all stages of research are important for developing, implementing, and evaluating programs that can effectively address health disparities in underserved women.1 Examples of these methodologies are positive deviance (PD), participatory action research (PAR), and community-based participatory research (CBPR), whose goals and processes are discussed in this article.
BACKGROUND
Theoretical models that underlie common theory-based interventions, such as the health belief model,2 the theory of planned behavior,3 or the health-promotion model,4 may assume common motivational and cognitive paradigms, such as social cognitive theory,5 that may be essentially biased toward the dominant culture and insensitive to differences between communities, thereby leading to health disparities. In addition, ethnic minority women and low-income women more frequently experience disparities in health, in part because they often lack the socioeconomic resources to address health that are available to more advantaged women.6 Conventional theories aimed at modifying the proximal behavioral antecedents of health disparities have been widely used in intervention research and programs among low-income and ethnic minority women; however, they may not be effective in producing the desired behavioral change. For example, interventions and educational materials derived from social cognitive theory5 did not result in a healthier diet or appropriate gestational weight gain nor did they decrease the incidence of gestational diabetes and neonatal macrosomia in pregnant Canadian women from the Cree Indian nation.7
Low-income and minority women may not have the resources to follow the intervention recommendations, resulting in perceived "noncompliant" behavior and subsequent poorer health.6 Underserved women may perceive other health concerns as needing more immediate attention and thus disregard the recommended changes. Moreover, early externally derived research-based intervention approaches proposed to apply the results from scientific experiments conducted on men to women. For example, the use of aspirin was found to be effective for the primary prevention of myocardial infarctions in men.8,9 These results led to a public campaign to encourage aspirin intake by the general population. However, scientists recently noted that aspirin was found to be ineffective in preventing myocardial infarctions in women.10 This emphasizes the need for methodologies that are sensitive to gender, socioeconomic, and ethnic differences.
Furthermore, randomized clinical trials, which exemplify positivist approaches to research on women's health, are designed to test the efficacy of a single treatment while strictly controlling environmental factors. This often makes them impractical for testing multifaceted interventions designed to change behavior to improve the health of underserved women.11,12 Newer approaches that use knowledge held by women, that is, local knowledge, to identify important health problems and use local resources to design community-specific solutions may more readily lead to behavior change that improves women's health.13
Using collaborative methodologies for women's health research is consistent with feminist perspectives.14 Applying quantitative and qualitative methodologies that emerge from feminist perspectives offers a unique way to address health disparities in underserved women. Although feminist theories differ, they share some basic principles. First, feminist perspectives recognize that women's diverse life situations are important to study. Second, the interaction of gender with other sociocultural factors, such as socioeconomic status and race, needs to be acknowledged. Third, the mutuality between researcher and subject must be recognized. Fourth, dualistic, hierarchical, and single-causal approaches are rejected. Fifth, and most important, women's experiences are important to acknowledge.15,16
Collaborative methodologies recognize the local knowledge that women hold regarding effective health behaviors and use resources that already exist in their community. Also, research approaches whereby women in underserved communities are actively involved throughout the research process may promote sustained behavior change. For example, developing an intervention to improve the health of low-income women may be more successful if the problems addressed were important to the women and if women were involved in designing the intervention using local resources.17 Low-income women may be more likely to change their health behaviors if researchers facilitated the development and testing of an intervention designed to address a health problem of interest to them.14
Understanding various collaborative methodologies can assist the researcher to select the strategy that is most consistent with the study purpose and setting. This article describes the foundations of these approaches, the role of women in the research, and the researcher's role. Then it compares how these methodologies differ in identifying and evaluating healthcare outcomes for underserved women.
COLLABORATIVE METHODOLOGIES FOR RESEARCH ON BEHAVIOR CHANGE IN WOMEN
PD approach
The PD approach operates from the assumption that solutions to community problems may already exist within that community. The researcher following the PD approach seeks to identify the uncommon but beneficial behaviors that allow some at-risk individuals to have better outcomes than do their equally at-risk neighbors. PD is similar to PAR and CBPR in that it does not follow the traditional needs-based approach to addressing health problems to solve problems. Instead, the PD approach inductively seeks to identify and optimize existing resources and solutions within a specific community to solve its health problems through behavior change.
The concept emerged in the 1960s in the field of nutrition18 and has spread to education, management, and health. The Save the Children program has used this approach to improve maternal and child nutrition in Bangladesh, Vietnam, Mali, and Nepal.19,20 In Vietnam, a PD initiative reduced moderate to severe malnutrition in children by an estimated 55%-85%, creating sustained behaviors that led to healthier growth patterns in subsequent siblings.21 The PD approach has also been used to identify positive behaviors that prevent teenage obesity in the United States.22 In addition to addressing community health problems, the PD approach has also been applied to business culture and making organizational change using solutions and processes that are organic to the corporation.23
Sternin22 suggests that the PD approach is appropriate where the problem of interest is widespread (ie, the norm), when those exhibiting the desired behavior are in the minority, and when the goal of the initiative is behavioral change in current prevalent practices. Furthermore, the PD approach promises an alternative to complex cognitive and behavioral theories for identifying interventions that may improve women's health behaviors. The latter are cumbersome and may lack the specificity to provide effective strategies for low-resource communities.24
There are 6 steps in the PD approach: (1) define the problem, perceived causes, and community norms; (2) identify individuals in the community who already exhibit the desired behavior; (3) discover the unique practices/behaviors that enable those positive deviants to find better solutions to the problem than do other people in the community; (4) design and implement interventions that enable others in the community to access and practice new behaviors; (5) determine the effectiveness of the intervention; and (6) disseminate the intervention more widely.19
Role of the research community and participants
The goal of a PD approach is to encourage adoption of the behavior(s) practiced by the positive deviants using existing resources in an impoverished community to improve a specific health outcome. The researcher collaborates with community members to identify individuals who demonstrate behaviors that enable them to have better health outcomes than their neighbors. After analyzing and isolating these behaviors, the successful strategies can be shared with the rest of the community. The positive deviants demonstrate to others in the community how to practice the healthy behaviors using local resources.
Also, the PD approach can inform program design. Women from the affected community should be actively involved in developing initiatives or programs that enable the positive deviants to share their strategies with other women in the neighborhood and in fostering the practice of the PD behaviors.25 For example, using PD methodology, Fowles et al24 identified strategies that enabled some low-income pregnant women to eat healthy meals while others did not. Women with healthy diets knew to eat balanced meals, had family support, were willing to prepare and eat foods that were different from those for family members, and ate at home more frequently than did women with unhealthy diets. In order to disseminate the positive deviants' knowledge, local public health agencies might sponsor educational programs in which positive deviant women share their successful methods with other women: how to purchase and prepare healthy meals, where to buy low-cost healthy foods that are accessible in the same community, how to adapt meals for themselves while preparing food for other household members, and how to increase fruit and vegetable intake. These sessions could be supplemented by a nutritionist-led discussion on healthy food options when women eat out at either fast food restaurants or other restaurants.24
Role of the researcher
The researcher using a PD approach collaborates with individuals within the community and uses local resources to ensure that PD behaviors can be disseminated and maintained.26 After a particular widespread health problem (eg, childhood malnutrition) has been identified, the researcher assesses the extent of the problem within the community and collaborates with community members to identify families who do not exhibit the particular health problem. Then the researcher conducts extensive observations of these families to identify the PD behaviors that facilitate the family's ability to avoid the health problem. After identifying the PD behavior, the researcher works in partnership with community members to facilitate sharing it and encouraging the community to practice it. In this way, the PD approach may lead to lasting behavior change, which could result in improved health for underserved women.25
PAR
PAR evolved from sociology and industrial psychology and attempts to find practical solutions to complex problems. It has been widely applied in the corporate world, educational settings, and the farming industry.27 Recently, the PAR approach was used to address chronic or community health conditions. For example, the PAR approach has been used to develop an empowerment model for persons with chronic fatigue syndrome,28 and to address domestic violence in 9 different cultural communities.29
Role of the research community and participants
The goal of PAR is to develop a plan of action to address the problem facing an organization or industry, for example, to increase production while decreasing costs, in partnership with the individuals affected by the problem. "In PAR, some of the people in the organization or community under study participate actively with the professional researcher throughout the research process from the initial design to the final presentation of the results and discussion of their action implications."30(p20)
In PAR, a strong collaborative relationship is established between the researchers and the population being researched. PAR relies on key informants who possess insights regarding the relationships within the organization or community. The researcher interviews these individuals to identify underlying problems and develops an ad hoc research team consisting of members representing all levels in the community. The researcher then guides the team in discussions intended to discover solutions that are practical and acceptable to all levels of the community. After implementing the intervention, the researcher guides the team in evaluating the effectiveness of the solution and adapting the intervention as needed.31
Not all members of the research team are involved in every aspect of the process. Key informants may be called upon only during the problem identification and/or evaluation phase. Members of the intervention team control the development and implementation of the intervention. However, in PAR, the initially identified solution may change when it is implemented. Creative surprises or new ideas that arise unpredictably during the intervention process may alter the original action plan.27
When applied to health-related problems rather than solving an organizational concern, PAR is very similar to the CBPR approach, which is described later in this article. Both approaches involve members of a targeted population or community to identify and evaluate solutions to existing health problems. These methodologies can be easily used to address health conditions affecting underserved women. In these instances, the behavior changes needed to improve health are more likely to be adopted if women targeted for behavior change are actively involved in the stages of the research process.32
Role of the researcher
PAR focuses on interpersonal relations and calls for the researcher to serve as a consultant/facilitator in all steps of the research process. The PAR researcher leads the initial assessments; trains community members on sampling, surveying, and communication techniques; and collaborates on the selection of suitable participants who appropriately represent the community. The researcher then designs the field research projects and tests the effectiveness of the potential solutions that coincide with participants' expectations and constraints.31(p212) PAR forces researchers to go through a rigorous process of checking the facts with those individuals involved with the research team before any reports are written or findings disseminated.
CBPR
PAR and CBPR are terms often used interchangeably, particularly because both use community members. However, subtle differences in underpinnings and goals between the approaches exist. PAR was developed from a social-cognitive psychological perspective, whereas CBPR draws from critical theory, feminism, and interpretive and other postmodern approaches that call for reflection and inclusion.32 In CBPR, particular attention is paid to issues of gender, race, class, and culture throughout the research process.33,34 Issues of power and personal agendas between researcher and various community constituencies are central and addressed in detail prior to initiating a community assessment.17,35-37 Programs that emerge from CBPR are culturally and politically appropriate for an area because they are grounded in the community environment.38,39
Role of the research community and participants
CBPR aims to eliminate health disparities and promote community and social change through research and action conducted via extensive collaboration between community members and researchers. The CBPR methodology has been used to manage diabetes in medically underserved minorities in Chicago,40 address health disparities in east Detroit,41 prevent HIV among teenage Latinos42 and low-income Latinas women,43 and prevent chronic disease in a southern California Latino community44 through cooperation among community residents, academics, and health practitioners.
With the CBPR approach, the community is the focus of the research, and members of the community identify health concerns and seek workable solutions to them. Community strengths and resources, both in terms of personal skills and community organizations, are organized into a network of individuals, social structures, and processes. This partnership among community members and researchers involves an empowering and power-sharing process that can make a significant improvement in the health of the community.35
Role of the researcher
The CBPR approach requires participation from the academic researcher that may be less traditional for academics. For instance, the researcher adopting a CBPR approach serves more as a collaborator with community members than as a director of a research project. The CBPR researcher may also be a consultant to community groups to collect and analyze initial assessment data, find core research-based articles to inform the community members' decisions, or serve as an educator as a component of an intervention program designed to address the health concern. The CBPR researcher collaborates with community partners to decide which parts of the research process are under the researchers' control and which components are controlled by the partners. These nontraditional research responsibilities of serving as leader, educator, and organizer enable the CBPR scientist to contribute to community health.42,45 Because this approach requires a wide range of abilities (team building, consensus development, and group processing as well as research design, instrument development, data collection, and data analysis), graduate education programs and professional organizations should expand to provide training in the skills needed for CBPR projects.
COMPARING COLLABORATIVE METHODOLOGIES
Identifying the problem
When a PD approach is used, a health problem is typically defined by others outside the community, such as ministries of health or, in the United States, the Centers for Disease Control and Prevention. Initially, the researcher collaborates with community members to conduct a rapid assessment designed to discover PD behaviors exhibited by a few individuals or families that enable them to avoid or manage a health problem.19
In contrast, the PAR or CBPR researcher identifies influential representatives of a community, then guides them through identifying problems that impede optimal community health and well-being. When starting a PAR or CBPR project, researchers need to carefully approach members of the target community so as to earn trust, gain entree, and identify healthcare agencies willing to cooperate with the team (Table 1).1,34,44 (See articles by E. Clingerman and M.L. Adams in this supplement for further discussion related to entry into the community.)
Solutions to problems
After community members identify individuals who exhibit PD behaviors, the PD researcher collaborates with appropriate community officials to develop interventions and programs so that successful strategies can be shared and practiced by others in the community in hopes of creating sustainable change.26
PAR and CBPR approaches require the researcher to serve as a facilitator and partner with community representatives who are part of the research team to identify solutions and to develop programs that can effectively address the community's health problems. The research team, including the community representatives, is responsible for evaluating and disseminating the program more broadly.27,28,32,35 In addition to disseminating program results in academic and clinical venues, distributing program products and results often requires establishing a network of service organizations and support agencies that demonstrate a commitment to continuing the program after the funding period ends.44
Evaluating outcomes
PD interventions and programs are designed to improve a widespread health problem by promoting multiple behaviors that contribute to improved health status. Yet it can be difficult to isolate which specific behavior had the greatest influence and contributed to the positive health outcome.49 The PAR or CBPR research team, which consists of community dwellers, sets parameters to determine when the intervention is considered successful. Community members of the research team may be involved in collecting and analyzing data that reflect the intervention's outcome. Using this data, the members of the team determine the level of success of the intervention and may suggest revisions that would need further study.27,28,32,35
Cost considerations
Although each of these collaborative methods capitalizes on assets already available in the community, they entail extensive personnel costs, not only of the external research team but also for reimbursement to participating community members and agencies. Furthermore, with PAR and CBPR approaches, providing financial incentives, such as providing small stipends or travel subsidies for community members is often necessary to ensure their continued involvement. CBPR projects may have to incur additional unexpected costs related to developing educational materials that reflect community-specific needs and values. Additional funding may be needed to train adequate numbers of community members to serve as lay healthcare advocates/educators.1,35,44
Furthermore, securing funding for all aspects of the collaborative projects can be difficult. National agencies that fund health disparities research subsidize personnel and cover training expenses.56 However, the need for additional funding for educational materials that were not anticipated during budget development may arise. The lack of reimbursement for providing meals for community members involved in the project, an expense often not permitted by federal funding sources, needs to be reconsidered. In their successful program aimed at reducing cardiovascular disease in a low-income Latino community, Kim et al found it necessary to provide meals, which were not funded directly by the national agency, to lay health advocates during long training periods and to provide food to low-income adults who participated in the health education classes. Securing adequate funding to support all aspects of these projects continues to be challenging.45
IMPLICATIONS FOR ADVANCING THE HEALTH OF UNDERSERVED WOMEN
Collaborative methodologies offer a unique approach to address health disparities in women, particularly minority women, by building on women's strengths. PD strategies frequently have been successful in reducing childhood malnutrition in several developing countries by drawing on the knowledge and behaviors of local women. This approach has been used to identify and disseminate positive behaviors that have helped (1) African women avoid genital mutilation using a grassroots approach49 and (2) peer support programs to encourage condom use in Vietnamese commercial sex workers, leading to a reduction in HIV infections.57 Its advantage lies in the rapid dissemination of information about effective and sustainable interventions that improve women's health, social mobilization, and rapid gathering of information that identifies health-promoting behaviors within a specific community. This approach has consistently demonstrated positive behavior change that leads to improved health for a wide variety of health problems in developing countries26 but has not been used widely to examine health disparities in the United States. Applying a PD approach to a population of underserved women in the United States may address their specific health concerns as well.
CBPR processes that give special attention to concerns related to gender, class, and race may be particularly effective for encouraging behavior change that address health disparities of minority women,1,34,44,58,59 with the focus on establishing an equitable power relationship among all team members.1,34,44 Women from the community who are included on the research team may offer a distinctive perspective on important community health concerns by naming factors that may contribute to positive health outcomes and enlisting support from other women in the area to participate actively in the program.
Researchers may find collaborative methods useful when attempting to address underserved women's health concerns. These approaches support health-promoting behaviors and incorporate preexisting community resources and are more effective and sustainable than traditional theory-based approaches. In addition, collaborative approaches are more likely to foster a broad-based, multifaceted change in behavior within a community, particularly in underserved, low-income, minority women, and women from indigenous cultures, than would an isolated study or program targeting a specific behavior or intervention developed from a positivist perspective.60 Collaborative approaches, if consistent with the goal of the research, offer a promising and effective strategy for addressing health disparities and advancing women's health.
REFERENCES