| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
Correspondence: Address correspondence to Sue Levkoff, Brigham and Women's Hospital: Department of Psychiatry, Aging Research and Education Center, 350 Longwood Avenue, Suite 201, Boston, MA 02115. E-mail: sue_levkoff{at}hms.harvard.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Minority • Recruitment • Retention • Elderly • Research
Essential to research in any field is the recruitment and retention of study participants. If recruitment lags, it is unlikely that results will be generalizable to the studied population; if retention lags, questions arise about the validity of conclusions drawn by study investigators. Thus, recruitment and retention pose specific threats to any studies involving minority elders. This article summarizes lessons learned about recruitment and retention of older ethnic minority individuals from a multisite national program to enhance research on minority aging and health promotion.
The Exploratory Centers for Minority Aging and Health Promotion (MAHP) were funded by the National Institute on Aging (NIA) and the Office of Research on Minority Health (National Institutes of Health, NIH) from 1993 through 1997. These studies were designed to conduct research and related activities aimed at improving the health status of older ethnic minority populations. The purpose of these Exploratory Centers was threefold: (a) to create an interdisciplinary coordinated program to stimulate research related to diseases, conditions, and other factors that lead to ill health and disability in older minority populations; (b) to support research on health behaviors and their relationship to health status; and (c) to develop and test interventions to improve health and functioning and programs of health education and outreach (Ory, Lipman, Barr, Harden, & Stahl, 2000). The Centers provided an environment in which experienced and junior investigators, and nonminority and minority investigators, interacted and directed research efforts focused on health and aging of ethnic minority populations. Each Exploratory Center included pilot research projects, small-scale intervention studies, and an administrative core that typically implemented a centralized recruitment and retention strategy. Table 1 presents an overview of the individual centers, their research projects, and the ethnic minority populations studied.
|
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
At the macro level of the researcher perspective, that of the academic institution, barriers also presented themselves. We discovered, not surprisingly, that our academic institution was perceived as part of an elite power structure not invested in the welfare of the ethnic minority community and therefore not to be trusted as a partner in research. There was also tension between other major academic centers wishing to expand as well as local agencies that perceived this as a threat to their patient base. Finally, competition among universities attempting to study the same groups created confusion for agencies in terms of who was doing what, with whom, and for what purpose. The result of this confusion was the feeling among agency administrators that the benefits of participation in research was outweighed by the costs (Levkoff et al., 2000).
A match in perspectives on the macro level occurred when both the community agency and the research organization were able to perceive a mutual benefit through collaboration. We offered a Chinatown community center and a Latino home care agency financial resources and research expertise to develop and implement a culturally informed dementia support group and Web site for Latina caregivers, respectively. It was essential to leave something behind in the community, rather than being perceived as just taking something away from the community. Sensitivity on the part of the research staff to agency concerns about the long-term commitment of our academic institution to their community was of critical importance. It was important that we become advocates for the community within our academic institution. We needed to make clear to agency administrators the benefits of participation in the long run, while setting up a collaborative, concrete exchange of services between the university and the community that would be beneficial in the short run (Levkoff et al., 2000).
Collaborator Research
All of the collaborating MAHP centers described the importance of overcoming potential distrust of local community groups in their academic institution. Sinclair and colleagues (2000) pointed out the failure of previous research to establish long-term relationships with community leaders as a major barrier to recruitment. They suggested that establishing this relationship requires a three-fold approach that focuses on respect for issues, needs, and "turf" politics of the target community. One way they suggested that this can be accomplished is to develop and collaborate in partnership with a local community advisory group composed of members who are key stakeholders of local community agencies, community leaders, and other key local gate-keepers. They described the various ways that a community advisory group can help the research project, including the following: helping to determine which community members to contact for recruitment, and the best methods for reaching these individuals; helping researchers to better understand local politics, which often shapes who to contact and how; helping to ensure the cultural competence of all components of the research process; and, reviewing research instruments and assisting in interpreting data. They concluded that research groups focusing on research alone, as opposed to research and community support through intervention, will run into barriers in the community. They stressed the need for researchers to become familiar faces in the community as they communicated the benefits of research for the community.
The value of developing a constituency-based recruitment effort was shared by each of the other MAHP investigators, who all stressed the importance of working with local community advisory groups, either ones developed expressly for the purpose of the proposed research, or better yet, a preexisting group that has credibility within the community of interest. The community advisory groups played different roles across the different centers and engaged in different activities throughout the research process at any one center. Collaboration with a community advisory group was the most commonly cited enabling factor for recruitment at the macro level mentioned across all centers. As Sinclair and colleagues (2000) noted, it is ironic that future efforts at recruitment in certain communities may be hindered by a community's lingering bad taste for researchers who promised in good faith that they would remain a visible presence in the community after the end of their research, but when faced with the realities of limited funding, pulled back from this commitment.
McNeilly and colleagues (2000) attempted to build trust for their study in the community from the onset of funding. The study was presented to the community as part of a larger university-wide effort that was already well established and respected by community members and called the Duke Center for Exploratory Studies in Older Minorities. Every effort was made to highlight the purpose of the study: to promote the health of area elderly minority communities. Linkages with recognized minority institutions were emphasized, including area Black churches and North Carolina Central University (an historically Black university). The principal investigators who directed and ran the research studies were minorities themselves and understood the local culture and priorities of the community. In an effort to build trust in the research staff, the Duke team hired recruitment and research staff who lived in the community and who shared a similar ethnicity, age, gender, and socioeconomic status with those they were attempting to recruit.
Hudson and colleagues (2000) described the usefulness of their center's newsletter, published twice a year, which helped to build community support for their research. The newsletter presented brief descriptions of study findings and implications for health, presented the results of national studies related to health promotion and disease prevention, and provided information on how their participation in research was helping students obtain master's and doctoral degrees. They felt strongly that the altruistic motives that many research participants hold provide a strong motivating factor leading to decisions to participate in research.
Facilitating access to the research site was also considered an important enabling factor increasing research participation. Sinclair and colleagues (2000) emphasized the importance of conducting the research in familiar settings, either at home or at a place conveniently located for the research participant. They provided transportation to the research site for study participants. The Harvard research team shared these concerns, and as a result, most interviews were conducted in the participant's home. For those caregivers that were working, interviews were also conducted in work cafeterias or local restaurants near the caregiver's place of work.
Although many researchers have described the church as a rich source of research participants for African American elders, Sinclair and colleagues (2000) cautioned that the entry phase of church-based recruitment efforts might not be cost effective because of the potential differing agendas between church leaders and researchers, with the former focusing on service outreach and the latter on research. The Harvard Center shared this view of the church based on our experiences with church pastors who were so overworked and understaffed that encouraging research participation was a low priority. In contrast, the African American church was the sole focus of recruitment for the University of Illinois Center for Health Interventions with Minority Elderly (CHIME), after efforts to establish their exercise program failed to attract participants in senior centers or senior housing sites. Prohaska and colleagues (2000) speculated that their success in recruitment within the church setting was based on the existing social support network within the organization, which included church directors and other networking tools, such as church bulletins, newsletters, and radio service programs that provided avenues for information dissemination and program awareness. They also found it helpful to have older parishioners in leadership roles in the church as recruiters, similar to the other centers that relied on recruiters sharing the same characteristic of the target population.
Mediator Level
Harvard Center Research
At the mediator level of the ethnic community group perspective, community gate-keepers had the power to provide the research team referrals for recruitment. Barriers arose for several reasons. In the Chinese American community, we found that gate-keepers often wanted to protect families from social stigma by avoiding officially labeling a client as demented. Others did not encourage families to participate because they saw no real benefit to individual families for participating and did not believe in the overall benefit of research participation (Guo, Levy, Hinton, Witzman, & Levkoff, 2000). Gate-keepers were also critical in assuring subjects that confidentiality of information would be maintained. The most cooperative gate-keepers were those who believed in research themselves and who were able to personally identify individual elders and caregivers that they believed would also view research as important. We also found a higher yield of referrals working from the bottom up, that is, recruiting through hands-on service providers who had an intimate knowledge of the elders they served, rather than dealing with higher level administrators who had little direct contact with families.
From the researcher perspective, barriers and enablers at the mediator level existed within the research team itself. Our multidisciplinary research team included individuals from many different academic backgrounds, including social gerontologists, medical anthropologists, psychologists, psychiatrists, ethnographers, and health services researchers. Differences in disciplinary perspectives (e.g. sociological, psychological, or anthropological) as well as research methodologies (e.g., quantitative or qualitative) resulted in pronounced differences in both the orientation and language used by research team members to describe research experiences and theoretical concepts. The research team was also multicultural and included individuals with different cultural backgrounds, beliefs, and social histories. These ideological and sociocultural differences, which began as barriers to effective communication among the research team, and thus to successful recruitment, were painstakingly transformed through open discussion and negotiation into enablers to recruitment.
Collaborator Research
Hazuda and colleagues (2000) described the importance of mediator-level factors that were crucial to successful recruitment and retention of Mexican American and European American elderly individuals for research in clinic and community agency settings. They also rated the relative importance of each factor according to whether they were essential, very important, or somewhat important to recruitment. They found it essential to match all aspects of their research to the culture of the targeted collaborators. They also found it essential that the design of their research activities would have negligible impact on regular staff routines or the conduct of other studies of the research collaborators. Very important approaches included having preestablished relationships with potential study collaborators and giving regular updates on research progress and results. Providing outreach training and/or educational programs for study collaborators was of somewhat importance.
In addition to rating the importance of recruitment strategies directed toward collaborators at the mediator level, Hazuda and colleagues (2000) also rated the importance of recruitment activities directed toward their collaborators' staff. All efforts directed at engaging collaborators' staff were of essential importance including the following: adapting research activities so as not to interfere with normal clinic/agency routine or staff priorities; tailoring study participation incentives, or as they refer to them, "reinforcements," to the specific culture of the office staff involved in the research (with some settings being receptive to special snacks or food, others valuing monetary contributions to the office, some preferring monetary incentives for individual staff referrals, and others valuing researcher participation in staff social events or regular updates of research progress and results); and, demonstrating sensitivity and genuine concern for the individuals referred.
Individual Level
Harvard Center Research
At the micro or individual level, a number of barriers existed from the ethnic minority group perspective. The legacy of Tuskegee remained an active voice explaining many African Americans' distrust of the research team. Many caregivers feared stigma to the family if a diagnosis of dementia were made public. Others feared a loss of services and did not trust in the confidentiality of the interviews. This last factor was particularly apparent among Latino participants, some of whom were not citizens and feared interview information might lead to deportation. Issues of confidentiality were also key for several of the Chinese American caregivers who refused to participate in the study for fear of individuals outside the family learning too much about their private family matters. Caregivers who agreed to participate expressed that they benefited from sharing their stories and from contributing to efforts that might improve the lives of other caregivers in similar situations (Levkoff et al., 2000).
One barrier at the level of the individual researcher was based on conflicts in the researchers personal goals. Some interviewers valued their own activist agendas over the study research agenda and felt ambivalent about taking on the role of researcher and abdicating their role of community advocate. Study investigators later found out that some of the researchers actually discouraged the potential participation of certain individuals that might have biased the generalizability of the responses received. Thus, many activist interviewers themselves became like community gate-keepers and raised barriers to recruitment. When it was possible to retain the activist-oriented researcher on the team, a barrier transformed itself into an enabling factor to recruitment because the activist-oriented researchers' previously established networks allowed us to better map out the local communities of interest.
Sensitivity to the degree of acculturation or assimilation of a potential research participant influenced greatly who should be approached for consent and how that approach would be interpreted. Flexibility in adapting the consent process to the particular family dynamic at hand was essential. In one case, a Chinese American participant would not give her consent to answer the caregiver survey without first talking to an uncle in California who was the eldest male in her family. Had we not respected her decision, which caused substantial delays in the research process, we would have probably caused her to refuse participation. Whereas consent is typically thought of as provided by one individual, for many participants, family members needed to be involved in the decision to participate, even though only one person ultimately signed the form.
Additional consent-related barriers at the individual level arose around language. We went through several iterations of consent forms before they were acceptable. Simple translation of forms was not adequate; substantial efforts were required through a careful translation and back-translation process in which attention to cultural idioms was also incorporated.
Linguistic matching was important throughout the research process. Even when a linguistic match was made, attention still had to be paid to the specific wording used. For example, the initial consent form stated, "We are talking to immediate family members of older persons with memory problems." This was an issue for the Chinese American participants, because married daughters are traditionally considered part of their husbands' family and not part of their family of origin. Furthermore, Chinese American daughters-in-law often were the de facto caregiver to the dementia-impaired elder, whereas the son of the affected elder was viewed as the "official" caregiver and needed to both provide consent and be present for the interview. It was essential to take these cultural factors into account, both in the wording of the Chinese American consent form and in the implementation of the Chinese American consent process. Members of the community advisory boards were key in making the language of consent forms and other instruments culturally appropriate.
A key enabling factor to successful recruitment and retention was the use of a mix of quantitative research with ethnographic and qualitative research, which allowed us to better understand the varying constructs of dementia and accompanying stigma held by ethnic minority families. The use of words that the potential study participants used to describe dementia, rather than the use of technical medical words, led to more effective communication with potential study participants and better recruitment and retention of study subjects.
Collaborator Research
On the individual or micro level, a number of additional barriers to recruitment and retention were identified. One barrier concerned potential study participants' concerns with regards to safety (Sinclair et al., 2000). Older African American participants have a greater fear of becoming a victim of crime than older White Americans (McAdoo, 1993). This can pose a significant problem for recruitment and retention. Sinclair and colleagues (2000) outlined several strategies to overcome these fears. Researchers can: conduct research via mail or over the telephone, without ever having to come face to face with the participant; notify respected community institutions, such as churches or local police stations prior to contacting potential participants to establish legitimacy in the community; and provide potential subjects with a name and phone number for any questions they may have. After repeated attempts to meet with potential participants in senior housing who refused to open their doors to the research staff, the Harvard Center had photo IDs made to show the elders that they were legitimate.
Although not specific to minority elders, Prohaska and colleagues (2000) noted that scheduling conflicts were a significant barrier to research participation for study subjects. They found that African American elders who did not consent to participation reported scheduling conflicts as the largest (58.2%) barrier to their participation. They recommended that researchers looking to have high retention or enrollment rates demonstrate flexibility in scheduling appointments with participants.
As Hudson and colleagues (2000) discussed, another significant barrier to research participation by ethnic minority elders is the researchers' selection of quantitative scales that may be inappropriate for the population being studied. If length and ease of comprehension of scales are not taken into account, elders may become confused or frustrated and either respond inaccurately or drop out of the study and refuse follow-up. Hudson and colleagues suggested that the length of the scales, as well as the redundancies of the measures, be taken into consideration when putting together study instruments. Sinclair and colleagues (2000) also discussed the importance of language in recruitment success. Written project materials needed to be clear and easy to read and sensitive to educational background. Effective wording of questions and piloting questions through open-ended interviews or in focus groups was helpful. Sinclair and colleagues also promoted offering potential participants the option of a verbal, rather than written, administration of the interview.
Ethics require that the research design be communicated in a clear manner. Sinclair and colleagues (2000) noted that recruitment may actually be assisted by this clarity. Focus groups with African American elders suggested that subjects' discomfort with the research process seemed to come from a lack of knowledge concerning the purpose of the research design. They suggested that if the research design were properly communicated, potential study participants would be more willing to enroll in the research study.
Though ethnic matching of project staff with the potential study participants has been discussed, another possible enabler for recruitment and retention in research is to seek ethnic matching between the principal investigator of the project to the studied population (Sinclair et al., 2000). Focus groups revealed that this was potentially of greater importance than mere ethnic matching of subjects to project staff. When the senior person responsible for the research was of the same ethnicity, potential subjects were likely to be more at ease.
Hazuda and colleagues (2000) rated the importance of a number of very specific enablers at the individual level that they found to be related to the recruitment of Spanish American elders. They discussed enabling factors and barriers oriented toward the family members of participants as well as those directed toward the individual participant. In terms of family members of the participants, they found it was very important to build rapport and interact with family members as joint decision makers, to accommodate Spanish-speaking family members, to adapt their approach to married female participants to meet their husbands' preferences (that is, secure husband's approval for wife's participation and use female interviewers), and to accommodate caretaking responsibilities for grandchildren. In terms of individual participation, they found it essential to demonstrate familiarity with the community and sensitivity to the participant's cultural background, to modify their research approach to match the worldview and life context of diverse population subgroups, to accommodate Spanish speakers, and to provide continued encouragement at follow-up visits. Though not essential, they found the following to be very important to assuring recruitment success and retention: emphasizing the importance of the study and reinforcing the elder's sense of altruism; emphasizing confidentiality, especially for such sensitive topics like mental health; conducting follow-up interviews at home or by telephone; sending thank-you notes and reminders; and providing monetary incentives for completed assessments.
Although to a large extent factors were rated similarly across research settings (e.g., the clinic vs. community agency) and ethnic groups (e.g., Mexican American vs. European American), Hazuda and colleagues (2000) noted several differences. The greatest difference between settings occurred at the mediator level and reflected the key role of collaborating physicians and their staff as gate-keepers of the clinic population. The key difference between Mexican American and European American groups occurred at the micro level and reflected cultural differences in language used (for some Mexican Americans, it was necessary to conduct all interviews in Spanish) and the role of family members in the decision to participate (with some Mexican American women having to secure their husband's approval for study participation).
Although Hazuda and colleagues (2000) found that an ethnic match between the interviewer and participant was not important, they had inconsistent findings about gender matching. Gender matching increased response rates among both Mexican American and European American women in a surgery study and among Mexican American women alone in a depression study. Hudson and colleagues (2000) also looked at the influence of gender on recruitment and retention separately for their studies in the community and clinic settings. They found that in the community setting slightly more men than women were recruited by personal contact, independent of churches or other organizations (80.8% vs. 69.9%), which understandably led to higher retention rates among men, because such contact typically increases follow-up. In contrast, retention rates in the clinic sample were higher for women than men (85.5% vs. 80.0%).
In Table 3 we summarize the lessons learned at each of the Centers on Minority Aging and Health Promotion.
|
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
We developed the matching model of recruitment in hopes that it might serve to enhance recruitment and retention efforts as well as further our theoretical understanding of recruitment and research participation. We plan to continue to build on this model and add additional levels (e.g., environmental) and perspectives (female/male) where appropriate. Formal evaluation studies are needed to compare the yield as well as the cost-effectiveness of the different approaches discussed.
|
Footnotes |
|---|
1Brigham and Women's Hospital: Department of Psychiatry, Aging Education and Research Center, Boston, MA. 
Received for publication December 31, 2001. Accepted for publication March 20, 2002.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. Davies, J. Larson, N. Contro, C. Reyes-Hailey, A. R. Ablin, C. A. Chesla, B. Sourkes, and H. Cohen Conducting a Qualitative Culture Study of Pediatric Palliative Care Qual Health Res, January 1, 2009; 19(1): 5 - 16. [Abstract] [PDF]
|
|
|
|
 |
|
 E. J. Samelson, J. L. Kelsey, D. P. Kiel, A. M. Roman, L. A. Cupples, M. B. Freeman, R. N. Jones, M. T. Hannan, S. G. Leveille, M. M. Gagnon, et al. Issues in Conducting Epidemiologic Research Among Elders: Lessons From The MOBILIZE Boston Study Am. J. Epidemiol., December 15, 2008; 168(12): 1444 - 1451. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Burns, A. C. M. Soward, A. H. Skelly, J. Leeman, and J. Carlson Effective Recruitment and Retention Strategies for Older Members of Rural Minorities The Diabetes Educator, November 1, 2008; 34(6): 1045 - 1052. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Thompson, T. Baranowski, K. Cullen, K. Watson, A. Canada, R. Bhatt, Y. Liu, and I. Zakeri Food, Fun and Fitness Internet program for girls: influencing log-on rate Health Educ. Res., April 1, 2008; 23(2): 228 - 237. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. G. Eakin, S. S. Bull, K. Riley, M. M. Reeves, S. Gutierrez, and P. McLaughlin Recruitment and retention of Latinos in a primary care-based physical activity and diet trial: The Resources for Health study Health Educ. Res., June 1, 2007; 22(3): 361 - 371. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. V. King, E. O. Burgess, M. Akinyela, M. Counts-Spriggs, and N. Parker "Your Body Is God's Temple": The Spiritualization of Health Beliefs in Multigenerational African American Families Research on Aging, July 1, 2005; 27(4): 420 - 446. [Abstract] [PDF]
|
|
|
|
 |
|
 L. B. Dunn and N. E. Gordon Economic Behavior and Informed Consent--Reply JAMA, May 4, 2005; 293(17): 2092 - 2092. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Pinquart and S. Sorensen Ethnic Differences in Stressors, Resources, and Psychological Outcomes of Family Caregiving: A Meta-Analysis Gerontologist, February 1, 2005; 45(1): 90 - 106. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
