Recruitment of the elderly into a pharmacologic prevention trial: The Ginkgo Evaluation of Memory Study experience

Contemporary Clinical Trials 27 (2006) 541 – 553 www.elsevier.com/locate/conclintrial

Recruitment of the elderly into a pharmacologic prevention trial: The Ginkgo Evaluation of Memory Study experience Annette L. Fitzpatrick a,⁎, Linda P. Fried b , Jeff Williamson c , Patricia Crowley d , Delilah Posey g , Linly Kwong e , Janet Bonk f , Roberta Moyer f , Joyce Chabot d , Lara Kidoguchi a , Curt D. Furberg g , Steven T. DeKosky h for the GEM Study Investigators a Department of Epidemiology, University of Washington, Seattle, WA, USA Departments of Epidemiology and Medicine/Geriatrics, Johns Hopkins University, Baltimore, MD, USA c Sticht Center on Aging, Wake Forest University School of Medicine, Winston-Salem, NC, USA d JHU Field Center, Johns Hopkins University, Hagerstown, MD, USA e Sacramento Field Center, University of California — Davis, Sacramento, CA, USA f Division of Geriatric Medicine, University of Pittsburgh, Pittsburgh, PA, USA Department of Public Health Sciences, Wake Forest University School of Medicine, Winston-Salem, NC, USA h Department of Neurology, University of Pittsburgh, Pittsburgh, PA, USA b

g

Received 3 August 2005; accepted 20 June 2006

Abstract The difficulty involved in recruiting healthy older adults into clinical trials, especially those involving pharmacologic agents, is an important issue in research. The Ginkgo Evaluation of Memory (GEM) Study, a double-blind, placebo-controlled randomized clinical trial evaluating Ginkgo biloba to prevent dementia, successfully recruited 3072 participants age 75 years and older at four U.S. sites from September 2000 through June 2002. Using targeted mailing lists, an estimated 243,400 study brochures were mailed out to potential participants. Subsequent attempts were made to reach 14,603 households by telephone, from which 12,186 (83.4%) successful contacts were made. Overall, telephone or in-person evaluations identified 2149 (17.6%) ineligible persons for cognitive (20.6%), medical (49.4%), or other (30.0%) reasons. A total of 6944 (57.0%) refused participation resulting in 3072 enrolled into the study, a recruitment rate of 25.2% based on telephone contacts made or 1.3% of all mailed brochures. Recruitment rates were stable over the 21-month enrollment period but were higher for the two urban centers than the two rural ones. Recruitment was dependent most on mailing lists available, density of older adults in the catchment areas, and Institutional Review Board restrictions. Men and persons under age 85 were more likely to enroll. Primary reason for refusals involved lack of interest (48.4%) or self-perceived poor health (16.2%). Over 9% were unwilling to give up current Ginkgo supplementation or would not accept assignment to placebo. An additional 7% did not want another medication and almost 4% had care-giving responsibilities which prevented involvement. Mass mailings were the most successful approach for recruitment at all four sites and the method

⁎ Corresponding author. Department of Epidemiology, University of Washington, Collaborative Health Studies Coordinating, Center Building 29, Suite 310 6200 NE 74th Street Seattle, WA 98115, USA. Tel.: +1 206 897 1903; fax: +1 206 616 4075. E-mail address: [email protected] (A.L. Fitzpatrick). 1551-7144/$ - see front matter © 2006 Published by Elsevier Inc. doi:10.1016/j.cct.2006.06.007

542

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

through which the vast majority of interviewees had learned about the study. Information on the experience of the GEM Study recruitment may be helpful to other clinical trials attempting to randomize older adults into prevention trials. © 2006 Published by Elsevier Inc. Keywords: Recruitment; Clinical trial; Ginkgo biloba; Dementia; Alzheimer's disease; Elderly; Aging

1. Background A large number of challenges face researchers planning studies of older adults, especially those focused on issues of dementia. While studies involving treatment of persons already showing symptoms of dementia have their own difficulties, [1–5] prevention trials recruiting non-demented older adults often face similar problems. The challenges faced can result in reduced response rates or the inability to reach study recruitment goals. Concerns have been expressed that low rates as well as differences between participants and non-participants may limit the generalizability of results to the general population or invalidate survey results [1,4]. In addition, obtaining a large enough sample to meet objectives is critical to assure that these studies are sufficiently powered to assess stated outcomes. Costs involved with the recruitment into studies of older populations can be extremely high and are also an important issue for investigators [6]. In prevention trials, an important issue of concern in recruitment of older adults involves their health [4]. Mortality is of great concern as enough individuals must remain in the trial for any treatment differences between treatment groups to become apparent. Functional abilities can also be a serious barrier to study success as adherence and retention may be affected. Cognitive difficulties and impairment may be a problem even for incidence studies. Scores on the MiniMental State Exam (MMSE) have been found to be a predictor of enrollment in studies of non-demented older adults after adjustment for education, alcohol use and other health problems with lower cognition correlated with lower recruitment [7]. Researchers in trials of the elderly must also be sensitive to different levels of sensory impairment that potential enrollees may have as vision, hearing and tactile limitations may alter test results so modifications to the protocol must be planned. Motivation for participation must also be considered [8]. Thus, the need for creative strategies in recruiting older adults into prevention trials is great. In this paper we describe the experience of the Ginkgo Evaluation of Memory (GEM) Study, a randomized, doubleblind, placebo-controlled clinical trial designed to evaluate Ginkgo biloba for the prevention of dementia [9]. This study successfully recruited 3072 adults age 75 years and older into the trial over a period of less than two years. We will describe the approaches utilized by the four study centers, evaluate differences in reasons for refusal by specific demographics, and examine how recruited individuals learned about the study. Finally, based on our experience, recommendations for others attempting to recruit older adults into similar types of studies are presented. 2. Methods Recruitment strategies across sites were initiated with the formation of a Recruitment Committee during the planning stages of GEM involving staff from all four clinical sites plus the Coordinating Center. Although all GEM sites had been involved in enrolling seniors age 65 years and older in the Cardiovascular Health Study (CHS) [10], it had been over ten years since this cohort was recruited. While CHS cohort members were approached for participation in the GEM Study, strategies for attracting additional members of each community were also needed. Conferences by telephone and meetings during GEM centralized training sessions identified standardized approaches that were deemed to be successful across sites, but also highlighted differences in sites due to Institutional Review Board (IRB) requirements and cultural subtleties of geographic areas. To address these, clinic recruitment staff were allowed to customize their own strategies based on differences. Telephone discussions throughout the recruitment period were utilized to review site-specific results, to allow recruitment coordinators to learn of successful approaches from one another, and to identify modifications in strategy that may be needed. Brochures were standardized but presented with each site's contact information. A coupon to allow recipients to either notify their respective site that they were interested in more information, or to state that they did not want to be contacted further about the GEM Study were included in the brochure. A press release by the study PI at the University of Pittsburgh opened recruitment in August, 2000.

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

543

All sites' IRBs required that responses from the brochure be respected. Only one site, Wake Forest University (WFU), was allowed to telephone both individuals who responded positively to the brochure mailing and to those who did not respond during all phases of recruitment. Two of the sites, University of California, Davis (UCD) and University of Pittsburgh (PIT) were restricted to calling only persons who responded positively to the mailing (i.e. sent back a coupon saying they were interested in the study). Researchers at Johns Hopkins University (JHU) were initially allowed to only contact those who requested more information, but the IRB modified its ruling and later allowed nonresponders to be called. Highlights of each site's strategies are briefly described below: 2.1. Wake Forest University (WFU) Strategies here relied primarily on mass mailings, using voter registration lists and InfoUSA (described below). A total of about 25,000 brochures were mailed. Some efforts were made recruiting at retirement centers, nursing homes, churches, fairs, and health centers. Posters were placed around town and half-page advertisements placed in the local paper. Realizing that recruitment goals could not be met from the Winston-Salem area alone, a second clinic was set up in Greensboro, NC, to attract potential enrollees from a new catchment area. Although Winston-Salem and Greensboro are urban centers, much of the geographic area targeted for recruitment is rural making this one of the more relatively rural of the GEM Study sites. 2.2. University of California, Davis (UCD) UCD enrolled participants from the city of Sacramento and surrounding areas. Initial recruitment efforts were made by mailing to the CHS cohort as well as to participants in other clinical studies inviting spouses, friends and family members to join the GEM Study. Mailing lists targeting over 20,000 seniors in Sacramento County were purchased with the assistance of the UC — Davis Public Affairs Department. An additional list of over 22,500 seniors in Sacramento and outlying areas was obtained to supplement the mass mailing effort. A repeat mailing of about 12,000 was also done here. Media awareness through newspapers, television and radio coverage was fairly successful. Limited recruitment was obtained through local community involvement and presentations. 2.3. Johns Hopkins University (JHU) The JHU clinic is located in Hagerstown, MD, a geographically rural and semi-rural area two hours west of Baltimore. Due to the small-town ambiance of this site, a variety of recruiting strategies was implemented which included newspaper advertisements, magazine and newsletter articles, radio and television public service announcements, community presentations and posters, as well as the brochure mailings. For mailings, JHU utilized its relationships with participants from another study, CLUE II (resulting in 2600 mailings) in addition to CHS. This clinic also sent out brochures using AmericaList to seniors within a 50-mile radius of the clinic (n = 12,600) and later, InfoUSA (n = 1000). 2.4. University of Pittsburgh (PIT) This urban site relied primarily on mass mailings with about 159,700 mailed out during the recruitment period. As only persons requesting more information could be contacted, 4532 persons were called as a result of all recruitment efforts. Of this number, about 94% were due to response cards in study brochures, while the remaining persons were identified from all other efforts, including television reports (3%), newspaper articles (2%) and other community contacts or friends of those already contacted (1%). In addition to the city of Pittsburgh, the catchment area included potential participants within a 35-mile radius of the clinic in Allegheny and four bordering counties. A national press release was organized by the site of the GEM PI at the University of Pittsburgh which included statements by the PI and National Center for Complementary and Alternative Medicine (NCCAM) officials of the National Institutes of Health (NIH). Local press releases were simultaneously coordinated at the other clinical centers. Other introductory efforts included radio, television, and newspaper coverage and informational presentations to local medical societies. These activities spear-headed recruitment efforts by introducing the study to the general public, providing positive public relations and increasing enthusiasm and community support.

544

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

Table 1 Approaches to recruitment utilized by individual field centers (sites); numbers indicate the level of effort for each, with 3 being the highest and 0 indicating no effort in this approach Method

Mailing lists Voter registration America List InfoUSA University lists Other studies Cardiovascular Health Study Other clinical studies On-site visits Retirement centers Nursing homes Churches/church functions Community functions Health fairs Newspaper ads Radio Television (public service announcements) Magazines/bulletins Posters Press release Other presentations Signage at clinic

Site⁎ WFU

UCD

JHU

PIT

3 0 2 0

0 0 2 2

0 3 1 0

3 0 1 0

1 0

1 2

1 2

1 1

1 1 1 1 1 2 0 0 0 1 1 0 0

2 0 0 3 3 2 1 1 2 1 2 1 1

3 0 2 2 2 2 2 2 1 2 1 2 1

1 0 0 0 0 2 0 1 1 1 2 0 0

⁎Site: WFU=Wake Forest University, UCD=University of California, Davis, JHU=Johns Hopkins University, PIT=University of Pittsburgh.

The GEM protocol required that each participant have a “proxy”, a person who was in close contact with the participant and who could answer questions about the participant's health as well as provide contact information. This made recruitment and obtaining informed consent from proxies a challenging yet important part of clinic procedures. A brochure targeting the information needs of the proxy was developed to enhance proxy recruitment. Proxies were asked to accompany participants to the clinic visits if possible. In these cases, transportation was offered for both the proxy and the participant. For proxies who could not come to the clinic, allowances were made to obtain informed consent by mail and to conduct a telephone interview for data needed by the proxy prior to the participant's first clinic visit. Proxies agreed to accompany the GEM participant to each subsequent clinic visit or to be contacted every six months by telephone to update information. Although no specific goals were set for minority recruitment, all sites developed some activities to attract non-whites into the study. Visits to churches were made to discuss the trial and community leaders were enlisted to introduce GEM coordinators. Advertising in church newsletters and minority neighborhood newspapers was also done, and posters were displayed at local community centers and offices utilized more frequently by non-whites. Efforts were increased during the latter half of the recruitment period after earlier statistics revealed that minority enrollment was lower than expected. After recruitment was about half over, it became apparent that additional persons would be needed to attain study goals. With the help of NCCAM, a mailing list was purchased from InfoUSA which included new geographic regions (not initially contacted by sites) and which targeted older adults and ethnic minorities. All sites utilized this list to complete recruitment. The screening protocol for GEM consisted of a three-stage system which included a telephone interview, a screening visit, and a baseline visit. At each step, eligibility and interest in joining the study were assessed. The telephone screening consisted of a computer-assisted telephone interview (CATI) programmed to assess potential participants' eligibility based on self-reported conditions and health parameters. Previously described in detail [9], eligibility requirements for the study included being free of dementia, 75 years of age or older, having English as one's primary language, and having the ability to find a proxy to represent him/her. Exclusion criteria included taking an anticoagulant, antipsychotic, or cognition medication, or more than 400 I.U. vitamin E per day; history of bleeding disorders, Parkinson's disease, cancer in the past five years, congestive heart failure with disability, hospitalized

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

545

Fig. 1. GEM study recruitment from mailings to final enrollment, all sites.

depression in the past year, and abnormal levels of thyroid hormones (TSH), serum B12, hematocrit, white blood cell count, or platelets. Individuals reporting these conditions by telephone were considered ineligible. If eligible, additional information on the study was provided and the respondent was invited to complete a screening examination at the clinic. During this examination, a neuropsychological test was administered to ascertain cognition, and phlebotomy was done for analyses of serum to determine other eligibility criteria. Individuals passing the screening

546 A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

Fig. 2. Rate of successful randomization of GEM Study participants by site over 21 months of recruitment (September 2000 through June 2002).

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

547

Table 2 Final eligibility status of potential participants into the GEM Study by site, gender and age group

Sitec

Genderd Age group (years)e

Total

WFU⁎ UCD JHU⁎ PIT Female Male b80 80–84 85+

Attempted contacta

Successful contact

3230 3406 5253 2714 7379 7182 7473 3950 1781 14,603

2631 3209 3705 2641 6320 5843 6622 3390 1367 12,186

Refusedb

Ineligibleb

Randomizedb

N (%)

N (%)

N (%)

1410 (53.6) 1736 (54.1) 2656 (71.7)⁎ 1162 (44.0) 3752 (59.4) 3192 (54.6) 3334 (50.3) 1952 (57.6) 929 (68.0) 6964 (57.1)

488 (18.5) 557 (17.4) 592 (16.0) 513 (19.4) 1149 (18.2) 998 (17.1) 1257 (19.0) 588 (17.3) 247 (18.1) 2150 (17.6)

733 (27.9) 916 (28.5) 457 (12.3) 966 (36.6) 1419 (22.4) 1653 (28.3) 2031 (30.7) 850 (25.1) 191 (14.0) 3072 (25.2)

a

Telephone contact attempted according to IRB requirements: after a positive response from brochure mailing (UCD, PIT) and after elimination of negative responses (WFU, JHU). b Percent based on individuals successfully contacted by telephone. c WFU=Wake Forest University, UCD=University of California, Davis, JHU=Johns Hopkins University, PIT=University of Pittsburgh. d Gender was unknown on 42 individuals. e Age was unknown on 1399 individuals. ⁎ A higher percentage of hang-ups on first telephone contacts (“cold calls”) occurred.

examination were invited to the baseline visit and to be randomized to 240 mg of G. biloba or placebo per day. At each step of the way, the potential participant was allowed to refuse to continue in the process. Ineligibility of each participant and reasons for refusal were tracked throughout all phases of the study. About six months into recruitment, a question was added to the telephone interview asking respondents to tell us how he/she had first heard of the study. Enrollment of 3000 randomized participants was needed to ensure adequate statistical power to determine a difference between G. biloba and placebo over the study period of five years [9]. Weekly numbers were set for each site in order to reach the final goal and data on success were carefully tracked throughout the recruitment period. Goals differed by site based on initial success and projections made using catchment area demographics. Recruitment was done from September 2000 through June 2002, a total of 21 months. Data on recruitment and randomization were entered and stored in a Structured Query Language (SQL) Server data base. Microsoft Access was used to extract data needed for this manuscript. Data shown here is primarily descriptive, with counts and percentages displayed. χ2 tests were done when appropriate to determine differences by site, gender and age group. SPSS, version 12.0, was used for all statistical analyses. 3. Results As described in Methods, a variety of approaches were utilized by the sites to inform potential participants about the GEM Study. A summary of the use of different tactics is shown in Table 1. In general, all sites relied heavily on mailing lists to send brochures to older adults living in each respective geographic region. The two more rural sites (WFU and JHU) spent more time on community recruitment efforts than did the more urban field centers (UCD and PIT). Staff in Pittsburgh focused their efforts almost entirely on mailings and sent out more than three times as many brochures as did the other sites. Both rural/semi-rural sites were limited by the number of seniors living within their respective catchment areas. For this reason, the enrollment goal for JHU was capped at 450 and researchers at WFU opened a second field center in another city (Greensboro) in order to reach their goal of 730 participants. Both urban centers (UCD and PIT) were able to compensate for the reduced enrollment at the rural sites by increasing their enrollment to over 900 individuals each. A flow chart of recruitment results, from initial mailing of brochures to randomization of participants, is shown in Fig. 1. Over 243,000 brochures were mailed out to inform potential participants about the GEM Study. However, only 47,329 persons comprised the true sampling frame from which telephone calls could be made due to IRB restrictions at three of the four sites. From this list, 14,603 attempted phone calls were made, and 83% of these attempts were successful. Of the 12,186 households reached, 4477 persons (37%) refused to begin or complete the telephone survey.

548

Reason for refusal

Gender a

Site WFU

UCD

JHU

PIT

Total N

1410 N (%)

1736 N (%)

2656 N (%)

1162 N (%)

Not interested/too busy Poor health/physical limitation Unwilling to take placebo/give up Ginkgo Don't want another med Caring for someone else Advised against participating Leaving town Transportation problems Other/Unknown

669 (47.4) 199 (14.1) 75 (5.3) 63 (4.5) 47 (3.3) 32 (2.3) 23 (1.6) 4 (0.3) 298 (21.1)

621 (35.8) 181 (10.4) 237 (13.7) 59 (3.4) 63 (3.6) 67 (3.9) 48 (2.8) 66 (3.8) 394 (22.7)

1084 (40.8) 457 (17.2) 135 (5.1) 225 (8.5) 89 (3.4) 88 (3.3) 36 (1.4) 27 (1.0) 515 (19.4)

352 (30.3) 76 (6.5) 63 (5.4) 54 (4.6) 12 (1.0) 15 (1.3) 20 (1.7) 9 (0.8) 561 (48.3)

a b c

pc

b.001 b.001 b.001 b.001 b.001 b.001 .002 b.001

Age group (years) b

Female

Male

3752 N (%)

3192 N (%)

1396 (37.2) 503 (13.4) 302 (8.0) 223 (5.9) 118 (3.1) 104 (2.8) 64 (1.7) 72 (1.9) 970 (25.9)

1318 (41.3) 410 (12.8) 208 (6.5) 168 (5.3) 93 (2.9) 98 (3.1) 63 (2.0) 34 (1.1) 800 (25.1)

Gender was unknown for 20 individuals refusing enrollment. Age was unknown for 749 individuals refusing enrollment. p-value based on chi-square tests comparing each specific reason (present/absent) by demographic category.

pc

.38 .003 b.001 .002 .12 .82 .44 b.001

b80

80–84

85+

3334

1952

929

1226 (36.8) 355 (10.6) 270 (8.1) 195 (5.8) 105 (3.1) 111 (3.3) 75 (2.2) 40 (1.2) 957 (48.7)

698 (35.8) 283 (14.5) 136 (7.0) 121 (6.2) 77 (3.9) 62 (3.2) 36 (1.8) 39 (2.0) 500 (25.6)

349 (37.6) 190 (20.5) 48 (5.2) 57 (6.1) 17 (1.8) 15 (1.6) 10 (1.1) 21 (2.3) 222 (223.9)

pc

b.001 b.001 .002 .07 .02 .50 .44 .002

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

Table 3 Reasons for refusing enrollment into the GEM Study by study site, gender and age group

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

549

Table 4 Enrollment status and reasons for refusals of 2409 potential GEM participants already enrolled in the Cardiovascular Health Study Enrollment Status

N (%)

Ineligible Refused initial contact Refused — after telephone call Reason for refusal Not interested Not feeling well enough/health problems Don't want another medication Don't want placebo/won't give up Ginkgo Not enough time/too busy Caring for sick spouse/friend/relative Too much trouble Transportation problems Planning to move Physician/friends/relatives advise against it Study concernsa Other Unwilling to provide a reason Randomized Total

335 (13.9) 166 (6.9) 1659 (68.9)

a

491 (29.6) 316 (19.0) 137 (8.3) 121 (7.3) 98 (5.9) 76 (4.6) 63 (3.8) 32 (1.9) 31 (1.9) 30 (1.8) 26 (1.6) 72 (4.3) 166 (10.0) 249 (10.3) 2409 (100.0)

Includes concerns about study protocol (too many visits, visits too long) or study medication (safety of Ginkgo biloba).

Of the 7709 individuals who completed the phone interview, 1707 (22%) were ineligible based on self-reported medical conditions or cognition problems discerned by results of the Telephone Interview for Cognitive Status (TICS) [11]. A little less than half of those who completed the telephone interview, 3756 individuals, were eligible and agreed to come into the clinic for a screening examination. At the screening visit, an additional 413 persons were found to be ineligible for cognitive, medical or other reasons, and 3133 agreed to return for the baseline clinic visit. Of these, a total of 3072 participants were randomized into the study. In summary, telephone or in-person evaluations identified 2149 (17.6%) ineligible persons for cognitive (20.6%), medical (49.4%), or other (30.0%) reasons. A total of 6944 (57.0%) refused participation, resulting in a recruitment rate of 25.20% based on telephone contacts made. Although recruitment rates varied by site, accrual was fairly stable over the 21 month enrollment period for each field center (Fig. 2). Rates were highest for the two urban centers (UCD and PIT) and somewhat lower for the two rural centers (WFU and JHU). By week 65, JHU had exhausted the mailing lists identifying seniors within their rural catchment area and recruitment ceased. UCD and PIT reached their enrollment goals by exceeding 900 participants each by week 76. WFU continued recruitment nine more weeks in order to maximize the potential of their second clinic site. Of the 3072 participants recruited into GEM, the percentage randomized based on telephone contacts ranged from 12% to 37% by site (Table 2). While eligibility rates were similar across sites, the site with the highest recruitment rate had the lowest refusal rate of the four centers. The site with the highest refusal rate was geographically the most rural, and a disproportionate number of refusals occurred up front at the first telephone contact with respondents refusing to talk to GEM recruiters (many of these were persons hanging up immediately after answering the telephone). This was one of the two sites that allowed calls to be made to participants if no response to the mailed brochure resulted. Eligibility was similar by gender and age group, while refusals tended to be higher for women and adults over age 85 years. Primary reason for refusals involved a lack of interest (48.4%) or self-perceived poor health (16.2%) (Table 3). Over 9% were unwilling to give up current Ginkgo supplementation or would not accept assignment to placebo. An additional 7% did not want another medication and almost 4% had care-giving responsibilities which prevented involvement. Reasons for refusing participation varied significantly by site. Individuals contacted by the two rural sites (WFU and JHU) reported being uninterested in the study or too busy, as well as not feeling well enough to participate. Residents of the Sacramento area were more than twice as likely to be unwilling to give up Ginkgo they were currently taking (or rejected being put on a placebo) and also reported more transportation problems. More women than men refused enrolling into the study due to poor health; they also were less willing to give up current use of Ginkgo. While “young–old” adults (under age 80 years) were busier and less interested in the study, those 85 and older were more likely to report heath problems as barriers to entering the study, although they met study eligibility criteria.

550

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

Fig. 3. Method of hearing about the GEM Study by site.

Initially, the study had been set up to recruit most participants from the Cardiovascular Health Study (CHS) [10]. Of the initial cohort of 5888 adults over 65 years of age at enrollment in either 1989/90 or 1992/93 [9], only 2409 were alive and met GEM eligibility criteria (Table 4). Of these, only 10% were ultimately randomized into this clinical trial. Compared to the community sample, CHS participants were less likely to cite level of interest as a barrier but more likely to believe their health was a factor in their decision. CHS participants were also less willing to take another medication and were more likely to be caring for a sick spouse, friend or relative. When asked how they had first heard of the GEM Study, the vast majority of respondents replied that the brochure mailed to them by the field center was their first information about the study (Fig. 3). Under 20% had heard of GEM through newsletters sent out by other studies or organizations (e.g. CHS, American Association of Retired Persons (AARP), local community newsletters). Very few persons (b 8%) had learned about GEM through newspapers, television or radio stories or advertisements, and even fewer became aware via community events. Sacramento County was most successful in using word-of-mouth to increase awareness about this trial. 4. Discussion In less than two years, the GEM Study successfully recruited over 3000 adults 75 years of age or older into a five year primary prevention trial to evaluate G. biloba for reducing the risk of dementia. Early in protocol development, it was acknowledged that recruitment in this age strata would be difficult, and a system to collect data on many aspects of recruitment was developed. From the initial mailing of brochures, through telephone and in-person screenings and ultimately to enrollment, the system tracked eligibility and refusal rates which allowed investigators to monitor recruitment and respond as needed. Reports in the literature have cited the number and type of exclusion criteria, and the perceived level of risk/benefit involved in the research as two important factors that influence recruitment [1]. Cohen-Mansfield et al. reported recruitment rates ranging from 1% to 82% in five of their studies of dementia, attributing some of the variability to

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

551

differences in these factors. They concluded that rates increased according to perception of risks and benefits as well as exclusion criteria. Pharmacologic studies had much lower recruitment rates than non-pharmacologic studies. However, many trials of older adults are designed as treatment studies and attempting to enroll persons with mild forms of dementia may add a far greater burden on recruitment than incidence studies. For example, a review of dementia prevalence studies concluded that important issues such as mortality, functional dependency and sensory impairment are seldom addressed in study design, resulting in under-representation of the cognitively impaired [4]. Likewise, even in studies in which nondemented individuals are targeted for incidence studies, cognition may play a role in who is eventually enrolled. In evaluating characteristics of participants and non-responders in a community survey of cognitive status in the elderly, Norton et al. reported that the MMSE score remained a predictor of participant enrollment after adjustment for education, place of birth, use of alcohol, and other health-related problems [7]. Lower cognition was associated with less successful enrollment. The need to communicate strategies that will best deal with the special challenges of this population is evident. While inclusion criteria in GEM primarily involved age and cognition, exclusions were made in efforts to assure that the participant would be able to remain in the study for five years. Thus, our sample may reflect the group that many other primary prevention studies of the elderly may desire: relatively healthy, non-demented older adults. The role of the caregiver is also important to consider when recruiting the oldest old for research. Several studies have undertaken the task of understanding the motivations of caregivers in facilitating a family member's participation in research studies. While we were conducting a primary prevention trial, we still required that a “proxy” be available to participate in GEM so that information could be acquired should the enrollee become cognitively or physically disabled. Motivations for proxies have been reported to be a belief that the study would help a family member feel better, contributions to medical science, and hope for a cure [8]. A qualitative study reported that families participated in research to help their care recipients and future generations, to receive support from clinical and research staff, and to obtain feedback about patient status and research results [2]. A total of 202 (1.5%) persons refusing enrollment stated that others had advised against it. While this is a very small percent, the need for proxy involvement emphasized the importance of communicating information about the study to persons other than the potential enrollee. A brochure addressing issues of concern to proxies and their role in GEM was developed to help facilitate proxy cooperation. Difficulty in recruiting ethnic minorities into medical research studies has been noted by several other investigators [12–15]. Issues of trust appear to be crucial to African-Americans who believe, more so than whites, that their physicians would be less likely to fully explain research participation and that their physician would expose them to unnecessary risks [16]. The GEM Study was only able to enroll a total of 139 (4.5%) non-white participants [9]. While community efforts and use of the InfoUSA mailing list were used to target more minorities, final results did not meet initial expectation. Our experience indicates that the mass mailing approach may not be the best to attract ethnic minorities into a study and that different strategies are necessary if these demographic groups are targeted. An additional challenge facing the GEM recruiters involved the general availability of the study medication as an over-the-counter supplement in many pharmacies, health food stores and nutritional centers. Because G. biloba has been advertised to enhance memory and reduce cognitive-related problems, older potential participants (especially at the Sacramento site) were unwilling to stop taking Ginkgo on their own. The best response to this reason for not participating in the trial was to highlight the high quality of the GEM Study medication and to note that over-thecounter formulas cannot be validated. Regardless, we could not assure that study participants would be randomized to the active treatment arm of the trial, and this remained a reason for over 500 potential participants to refuse enrollment. The belief systems held by individuals play an important role in their willingness to become involved in prevention research. Motivating factors for potential participants of clinical trials in the past have included curiosity, self-interest and altruism as well as having relatives/friends with the condition of interest, or having a favorable attitude toward the need for humans in medical research [17,18]. Although we collected information on reasons that eligible persons did not want to enroll, responses were often general or broad in scope. We would have benefited more from probing deeper for details and to identify some of the values behind the reasons provided. For example, the most frequently cited reason for refusing to enter the GEM Study concerned lack of interest. However, this information is too general to provide a good understanding of underlying issues involved. Similarly, we found geographic differences in-persons' unwillingness to accept a placebo or give up current use of Ginkgo. Whether the greater frequency of this reason at the Sacramento site was associated with trends for increased usage of dietary supplements noted by others in California [19,20] or involved the fact that health maintenance organizations are more likely to provide reimbursement for herbal medicine in California than in other parts

552

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

of the US [21] is unknown. In order to develop future recruitment strategies that address underlying reasons for refusals, gathering greater detail on refusal information is needed and warrants further investigation. Several of the aforementioned studies have resulted in recommendations for recruitment of elderly individuals into research studies. These have included (a) providing a wide range of recruitment strategies; (b) assessing and revising strategies as necessary during recruitment; and (c) a stepped screening approach [6]. Direct recruitment approaches, such as working with community leaders through community centers, have been found to be more successful in recruiting elders from ethnic minority groups while older adults from the dominant culture can be recruited indirectly and for about half the cost through newspaper announcements [12]. Clinics with experienced recruitment staff have also been found to provide higher recruitment rates than those with inexperienced recruiters [22]. The experience in the GEM Study recruitment has led us to emphasize some of these recommendations, and to add or modify others. In general, we believe the following guidelines may be of value to other studies attempting to enroll older adults into primary prevention trials: 1. Mass mailings using targeted addresses are by far the most cost-effective means to reach older adults from the general population. 2. While presentations and activities at retirement homes, community centers and other facilities for the elderly may increase awareness of the project among the general public and improve public relations, it is a very manpowerintense strategy and does not yield many participants. 3. As a caveat to [1] above, an exception should be made for recruitment of ethnic minorities. Approaching churches or organizations trusted by the group of interest may be necessary as mass mailing methods are encountered with greater suspicion by non-whites than their white counterparts. 4. An organized effort to open recruitment which highlights coordinated press releases and community presentations may enhance visibility and subsequent enrollment. 5. A study's potential for enrollment is related to the density of targeted individuals within its catchment area. It may be necessary to develop additional sites/clinical centers in different geographic regions in order to reach study goals. 6. In studies of older adults, strategies for recruitment of a proxy for each participant is critical. Mechanisms to identify proxies and receive informed consent, either through the potential participant or using other procedures such as mail and telephone contacts, should be addressed in all recruitment systems. 7. Experienced staff with the skills necessary to work with older adults is essential. The GEM Study sites had been working in studies of older adults for a minimum of 15 years each. The success of our recruitment is directly attributable to a loyal and dedicated recruitment staff. 8. Staff training and regularly scheduled conference calls for recruiters to discuss problems and provide support will keep staff motivated and enhance results. 9. While current enrollment in another observational study may not exclude participants from a prevention trial, older adults already enrolled in other efforts may be tired of such studies or feel that they have already contributed sufficiently to research. Although members of these studies may enhance recruitment, it is best to use approaches that will attract new community members rather than rely on other research lists. 10. Development of a system to document eligibility, refusal and enrollment rates during recruitment is critical for monitoring a study, to identify problems early, and to track success. The Ginkgo Evaluation of Memory Study has shown that it is possible to meet recruitment goals of the oldest old for trials of primary prevention. With careful planning and a systematic approach, we hope that other studies may have similar experiences as the need for studying this segment of the population grows increasingly important. Acknowledgment This publication was made possible by Grant Number 5 U01 AT000162 from the National Center for Complementary and Alternative Medicine (NCCAM) and support from the National Institute on Aging, National Heart, Lung, and Blood Institute, and National Institute of Neurological Disorders and Stroke. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NCCAM, or the National Institutes of Health. We gratefully acknowledge the contribution of Dr. Willmar Schwabe GmbH & Co. KG, Karlsruhe, Germany, for their donation of the Ginkgo biloba tablets and identical placebos, in blister packs, for the study.

A.L. Fitzpatrick et al. / Contemporary Clinical Trials 27 (2006) 541–553

553

References [1] Cohen-Mansfield J. Recruitment rates in gerontological research: the situation for drug trials in dementia may be worse than previously reported. Alzheimer Dis Assoc Disord 2002;16:279–82. [2] Connel CM, Shaw BA, Holmes SB, Foster NL. Caregivers' attitudes toward their family members' participation in Alzheimer disease research: implications for recruitment and retention. Alzheimer Dis Assoc Disord 2001;15:137–45. [3] Riedel-Heller SG, Schork A, Matschinger H, Angermeyer MC. Recruitment procedures and their impact on the prevalence of dementia. Results from the Leipzig Longitudinal Study of the Aged (LEILA75+). Neuroepidemiology 2000;19:130–40. [4] Riedel-Heller SG, Busse A, Angermeyer MC. Are cognitively impaired individuals adequately represented in community surveys? Recruitment challenges and strategies to facilitate participation in community surveys of older adults. A review. Eur J Epidemiol 2000;16:827–35. [5] Treves TA, Verchovsky R, Klimovitsky S, Korczyn AD. Recruitment rate to drug trials for dementia of the Alzheimer's type. Alzheimer Dis Assoc Disord 2000;14:209–11. [6] Rubin RR, Fujimoto WY, Marrero DG, et al. The Diabetes Prevention Program: recruitment methods and results. Control Clin Trials 2002;23:157–71. [7] Norton MC, Breitner JC, Welsh KA, Wyse BW. Characteristics of nonresponders in a community survey of the elderly. J Am Geriatr Soc 1994;42:1252–6. [8] Mastwyk M, Ritchie CW, LoGiudice D, Sullivan KA, Macfarlane S. Carer impressions of participation in Alzheimer's disease clinical trials: what are their hopes? And is it worth it? Int Psychogeriatr 2002;14:39–45. [9] DeKosky ST, Fitzpatrick AL, Ives DG, et al. The Ginkgo Evaluation of Memory (GEM) Study: design and baseline data of a randomized trial of Ginkgo biloba extract in prevention of dementia. Cont Clin Trials 2006;27:238–53. [10] Fried LP, Borhani NO, Enright P, et al. The Cardiovascular Health Study: design and rationale. Ann Epidemiol 1991;1:263–76. [11] Brandt J, Spencer M, Folstein M. The telephone interview for cognitive status. Neuropsychiatry Neuropsychol Behav Neurol 1988;1:111–7. [12] MacEntee MI, Wyatt C, Kiyak HA, et al. Response to direct and indirect recruitment for a randomised dental clinical trial in a multicultural population of elders. Community Dent Oral Epidemiol 2002;30:377–81. [13] Lee RE, McGinnis KA, Sallis JF, et al. Active vs. passive methods of recruiting ethnic minority women to a health promotion program. Ann Behav Med 1997;19:378–84. [14] Carter WB, Elward K, Malmgren J, Martin ML, Larson E. Participation of older adults in health programs and research: a critical review of the literature. Gerontologist 1991;31:584–92. [15] Lovato LC, Hill K, Hertert S, Hunninghake DB, Probstfield JL. Recruitment for controlled clinical trials: literature summary and annotated bibliography. Control Clin Trials 1997;18:328–52. [16] Corbie-Smith G, Thomas SB, St. George DM. Distrust, race and research. Arch Intern Med 2002;162:2458–63. [17] Tolmie EP, Mungall MM, Louden G, Lindsay GM, Gaw A. Understanding why older people participate in clinical trials: the experience of the Scottish PROSPER participants. Age Ageing 2004;33:374–8. [18] Trauth JM, Musa D, Siminoff L, Jewell IK, Ricci E. Public attitudes regarding willingness to participate in medical research studies. J Health Soc Policy 2000;12:23–43. [19] Gordon NP, Lin TY. Use of complementary and alternative medicine by adult membership of a large northern California health maintenance organization, 1999. J Ambul Care Manage 2004;27:12–24. [20] Schaffer DM, Gordon NP, Jensen CD, Avins AL. Nonvitamin, nonmineral supplement use over a 12-month period by adult members of a large health maintenance organization. J Am Diet Assoc 2003;103:1500–5. [21] Hughes A, Penner M. Reimbursement for complimentary/alternative medicine by California HMOs. Manag Care Q 2001;9:1–4. [22] Cosgrove N, Borhani NO, Bailey G, et al. Mass mailing and staff experience in a total recruitment program for a clinical trial: the SHEP experience. Systolic Hypertension in the Elderly Program. Cooperative Research Group. Control Clin Trials 1999;20:133–48.