Abstract

BACKGROUND

Community-based participatory research (CBPR) is an approach to research that includes community stakeholders as partners rather than subjects in the research process (Israel et al., 2010; Minkler, Blackwell, Thompson, & Tamir, 2003; Minkler & Wallerstein, 2010; Viswanathan et al., 2004). CBPR has emerged as an effective approach to addressing health disparities among disenfranchised communities facing health inequalities (Freudenberg & Tsui, 2014; Israel et al., 2010, 2010b; Minkler, 2010; Minkler & Wallerstein, 2010). CBPR can be conceived on a continuum with limited input from stakeholders on one end and full engagement of stakeholders in all aspects of research on the other end (Minkler & Wallerstein, 2010). The National Institutes of Health (NIH), the Agency for Health Research and Quality (AHRQ), and the Environmental Protection Agency (EPA) have promoted CBPR to varying degrees over the past twenty years. In 2010, the Patient-Centered Outcomes Research Institute (PCORI) was established with funding created through the Affordable Care Act. PCORI funds patient-centered outcomes research (PCOR) that evaluates research “questions and outcomes meaningful and important to patients and caregivers” (Frank, Basch, & Selby, 2014, p. 1513). PCORI states that “incorporating the patient perspective into health care research is inherently valuable and that including the end user of research in the research process enhances usefulness and speeds the uptake of research into practice” (Frank et al., 2014, p. 1514). Much like CBPR, PCOR promotes involvement of patients and other stakeholders in the entire research process.

Pacific Islanders are one of the fastest growing populations in the United States, and are underrepresented in research. Arkansas has the largest Marshallese community in the continental United States. The Marshallese have significant health disparities, which include high rates of type 2 diabetes, obesity, and infectious diseases (Center for Disease Control and Prevention, 2009; LeDoux, 2009; Minegishi et al., 2007; Woodall, Scollard, & Rajan, 2011; Yamada, Dodd, Soe, Chen, & Bauman, 2004). To understand the context of the Marshallese community’s health disparities, it is important to understand the historical relationship between the United States and the Republic of the Marshall Islands. The Marshall Islands consists of 1,156 individual islands and atolls in Micronesia. Between 1946 and 1958, the US military tested nuclear weapons in the Marshall Islands (Barker, 2012; Cooke, 2010; Guyer, 2001). These tests were equivalent in payload to more than 7,200 Hiroshima-sized bombs. The largest test had a yield of 15 megatons, over 1,000 times the strength of the bomb dropped on Hiroshima, and exposed Marshall Islanders to significant levels of nuclear radiation (Barker, 2012; Cooke, 2010; Dobyns & Hyrmer, 1992; Guyer, 2001; Niedenthal, 1997). The Marshallese who lived on the bombed islands and atolls were relocated. However, Marshallese living on nearby atolls were not relocated, and they suffered immediate and long-term exposure to nuclear fallout (Conard, Dobyns, & Sutow, 1970; Dobyns & Hyrmer, 1992; Guyer, 2001; Yamada, 2004). According to the Atomic Energy Commission, the Marshall Islands are one of the most contaminated places in the world (Cooke, 2010). There have been numerous studies between 1965 and 2004 that have documented the ongoing health effects from the nuclear testing (Conard & Hicking, 1965; Dobyns & Hyrmer, 1992; Hamilton, van Belle, & LoGerfo, 1987; Howard, Vaswani, & Heotis, 1997; Kroon et al., 2004; Lessard, Miltenberger, Cohn, Musolino, & Conard, 1984; Pollock, 2002; Takahashi et al., 2003). In the midst of the nuclear testing, US scientists developed Project 4.1 to be implemented in the Marshall Islands in order to study the effects of nuclear radiation on humans (Barker, 2012). Exposed Marshall Islanders were interned in a camp on Kwajalein Atoll to be studied, and research was conducted without informed consent and without translation of study information into their native language (Barker, 2012).

In 1986, the United States and the Republic of the Marshall Islands signed the Compact of Free Association (COFA), which granted the US military exclusive use and control over a strategic portion of the Pacific and also allows the Marshallese to freely migrate to the United States without a visa (108th United States Congress). The Marshallese first arrived in Springdale, Arkansas, in 1986 after the COFA was signed. The relatively strong economy and low cost of living in northwest Arkansas continues to attract Marshallese migrants, with the population doubling between 2000 and 2010 (Jimeno, 2013). Approximately 10,000 Marshallese reside in northwest Arkansas. There are over 2,000 Marshallese students enrolled in the local school districts, and 93% of these students report speaking Marshallese at home (Springdale School District, 2014). Other sizable populations of Marshallese reside in Hawaii, California, Washington, Utah, and Oklahoma, and there is frequent travel between these states and Pacific Territories. The largest employer of the Marshallese in Arkansas is the poultry industry; 76% of employed Marshallese individuals work for Tyson Foods, George’s, and Butterball (Jimeno, 2013).

While COFA migrants can freely enter the United States without a visa, they are not US citizens and are ineligible to vote. The COFA migrant status also limits their access to many health and public benefits. Prior to the Personal Responsibility and Work Opportunity Reconciliation Act of 1996 (PRWORA), commonly referred to as welfare reform, COFA migrants were eligible for federally-funded health care programs including Medicaid and the Children’s Health Insurance Program. However, PRWORA excluded COFA migrants from the category of “qualified immigrants” for purposes of eligibility for federal public benefits including Medicaid (Asian and Pacific Islander American Health Forum, 2014; McElfish, Hallgren, & Yamada, 2015; U.S. Government Printing Office, 1996; United States Court of Appeals for the Ninth Circuit, 2014). The limited access to health care further exacerbates the health inequalities of this underserved and vulnerable population.

Research with the Marshallese community is constrained by a lack of prevalence data specific to the Marshallese living in the United States and is complicated by their lack of access to health care (McElfish, Hallgren, et al., 2015). Furthermore, the Marshallese have been skeptical of participating in research because of the historical trauma originating from the nuclear testing in their islands and subsequent medical experiments (Barker, 2012). The University of Arkansas for Medical Sciences Northwest is committed to a CBPR approach that engages Marshallese stakeholders to overcome these challenges to research, and collaboratively addresses the health disparities identified and prioritized by the Marshallese community (McElfish, Kohler, et al., 2015).

This article describes the process of developing patient-centered outcomes research with the full engagement of Marshallese patients and community stakeholders. Specifically, the article provides information on the process used and the specific ways that community stakeholders input shaped the research design and PCORI proposal. PCORI requires “meaningful involvement of patients, caregivers, clinicians, and other healthcare stakeholders throughout the research process—from topic selection through design and conduct of research to dissemination of results” (PCORI, 2015). PCORI holds “that such engagement can influence research to be more patient centered, useful, and trustworthy and ultimately lead to greater use and uptake of research results by the patient and broader healthcare community” (PCORI, 2015). PCORI provides an engagement rubric that outlines: 1) what they require for stakeholder engagement; 2) their requirements for judging the level of stakeholder engagement; and 3) their requirements for stakeholder engagement in planning a study, conducting a study, and disseminating the study results (Beal & Sheridan, 2014; PCORI, 2014). Components of the PCORI engagement rubric are utilized to evaluate our engagement process and organize the article.

METHODS

In 2011, the lead investigator became aware of the significant health disparities of the Marshallese community, and began meeting with Marshallese community members, Marshallese community organizations, and Marshallese churches with the goal of setting a community-driven research agenda using a CBPR approach. There are only two Marshallese community organizations in Arkansas: the Arkansas Coalition of Marshallese and Gaps in Services to the Marshallese Taskforce. The lead researcher partnered with these two community organizations, the Marshallese Consulate, and Marshallese churches to conduct a needs assessment. This needs assessment process took 18 months and used a mixed-methods design of qualitative interviews and a broad quantitative survey of the community. This mixed-methods needs assessment process is described in a separate article (McElfish, Kohler, et al., 2015). The top priority identified by the Marshallese community was type 2 diabetes (Center for Disease Control and Prevention, 2009; Hallgren, McElfish, & Rubon-Chutaro, 2015; McElfish, Kohler, et al., 2015). In order to address type 2 diabetes, the lead investigator brought together a diverse interprofessional research team and engaged 31 community stakeholders, including patients, family members, and health care providers (hereafter referred to as “stakeholders”) to begin planning an appropriate intervention with community support. The stakeholders included 16 patients with diabetes, 11 family members of patients with diabetes, and 4 community health care providers (see Table 1). Representatives from the two Marshallese community based organizations, the Marshallese Consulate, and Marshallese churches were included as either patients or family members of patients and represented these groups as well as their organizations. Discussions were conducted both in English and in Marshallese (the native language) depending on the preference of stakeholders. Since the researchers did not speak Marshallese, a bilingual translator provided verbal interpretation and took notes in English. The freedom to use their native language increased the stakeholders comfort and shifted the power of gaining knowledge and sharing information to the stakeholder.

Following stakeholder input from preliminary planning, an interprofessional CBPR team was formed and included: two Marshallese community co-investigators, two academic investigators with expertise in CBPR, two individuals holding the Pharm.D degree and a nurse diabetes educator to provide diabetes education, a PhD-prepared biostatistician with expertise in comparative effectiveness research, and two endocrinologists with expertise in the treatment of type 2 diabetes.

Over an 18-month period, between June 2012 and December 2013, the CBPR team engaged in multiple discussions with stakeholders regarding all aspects of the research design and PCORI proposal. As outlined in Table 2, and described below, decisions regarding the project took multiple meetings with stakeholders and often more than one element of the research design was discussed at each meeting. The information for this article is derived from the process notes from our research meetings. The PCORI application, as well as this article, was written with full involvement of the two Marshallese community co-investigators.

PROCESS AND RESULTS

DISCUSSION

Much of the prior literature has outlined the principles and processes for CBPR, but have not provided specific details of how those processes resulted in specific changes to conceptualization and conduct of research. Furthermore, many CBPR scholars eschew randomized control trials as inappropriate for CBPR, and there is little literature published on how to engage patients in the process of designing randomized control trials. The PCORI engagement rubric pushes the field forward by providing a practical tool to measure the process and outcome of engagement, and perhaps most importantly, apply engagement practices to comparative effectiveness research that is often conducted with randomized control trials. This article provides an example of how patient-engaged methods can be used to design and conduct a randomized control trial with a population who has been underrepresented in research and suffered significant historical trauma. In doing so, this article adds to the current engagement literature and encourages researchers and funding agencies to examine their practices to encourage inclusive engagement of stakeholders in all kinds and at all levels of research.

When stakeholders were asked what worked well and what did not work well concerning the engagement and consensus decision making process, they stated: “You listened to us. We told you what we want to do and you helped us do it. We are doing it together. It is hard to make the meetings and talk about research, but we are not just being studied by some outsiders, we are doing the research, and we are doing the research on something that we want, not just what the researchers want. I don’t really understand research sometimes, but I understand my community and that is important too” (McElfish, 2013).

Broad stakeholder input set the research agenda for our CBPR team, and then continued to provide input on all aspects of the research design. Stakeholder input has made a significant difference in the research design of our PCORI application. The inclusion of stakeholders in research opens the protocol to multiple and divergent points-of-view, which requires greater flexibility of the researcher, a willingness to explain complex research methodology and the methodological requirements of funders to stakeholders, and a willingness to change study designs to ensure community support. Even after the stakeholders and research team make a decision on the protocol or operations, additional discussion was often required and sometimes those discussions led to change. This fluid process continues to require open and extensive dialogue to balance community wisdom with scientific methods.

While much progress has been made in our CBPR partnership with the Marshallese community, we acknowledge that PCOR and CBPR, with full involvement of multiple stakeholders, requires significant time, effort, flexibility, and financial resources, and there are many potential impediments to the process. Most of the barriers we experienced are within the university system rather than the community. Our current research enterprise is designed to operate in a principal investigator model and rewards research that is conducted independently rather than collaboratively. In order to address health disparities and translate research into better health, more interprofessional, patient-centered, community-engaged research is needed. This entails a much more facilitative process and academic researchers who are willing to engage in research that they do not design on their own. Our current academic research incentive and promotion structure is not designed to support this type of research. In order for CBPR and PCOR to be implemented broadly, university research must create structures that promote and reward more collaborative, interprofessional, and facilitative models of research that focus on meeting patient and community needs.

CONCLUSION

Research with underserved, minority populations who have experienced significant historical trauma is strengthened through true stakeholder engagement. Utilizing a fully-engaged CBPR process, community stakeholders and the interprofessional community-academic research team designed a randomized control trial through multiple discussions about the research question, intervention, participant assignment, comparison group, data collection intervals, instrument selection, and recruitment. We had easy and quick agreement on some items while other topics included significant debate, education, and compromise between the stakeholders and community and academic researchers. As a result, we developed a hybrid research design that joins the scientific process with the wisdom of the community. Balancing the input, opinions, and desires of stakeholders with the requirement of a rigorous research design proved to be both challenging and invigorating.

Acknowledgments

Contributor Information

Pearl Anna McElfish, Director of the Office of Community Health and Research, University of Arkansas for Medical Sciences Northwest, 1125 North College Avenue, Fayetteville, AR 72703, Office phone (479) 713-8680, Cell phone (479) 264-8690.

Peter A. Goulden, Assistant Professor of Medicine, Director of UAMS Diabetes Program, College of Medicine, University of Arkansas for Medical Sciences, 4301 West Markham Street, Little Rock, AR 72205.

Zoran Bursac, Division of Biostatistics and Center for Population Sciences, Department of Preventive Medicine, University of Tennessee Health Science Center, 66 N Pauline St. Suite 463, Memphis, TN 38163.

Jonell Hudson, College of Pharmacy, University of Arkansas for Medical Sciences Northwest, 1125 North College Avenue, Fayetteville, AR 72703.

Rachel S. Purvis, Research Associate, Office of Community Health and Research, University of Arkansas for Medical Sciences Northwest, 1125 North College Avenue, Fayetteville, AR 72703.

Karen H. Kim Yeary, University of Arkansas for Medical Sciences, 4301 West Markham Street, Little Rock, AR 72205.

Nia Aitaoto, University of Arkansas for Medical Sciences Northwest, 1125 North College Avenue, Fayetteville, AR 72703.

Peter O. Kohler, Vice Chancellor, University of Arkansas for Medical Sciences Northwest, 1125 North College Avenue, Fayetteville, AR 72703.

References