Recruiting Participants for Community-Based Research:

The Diabetic Retinopathy Awareness Program

ELINOR R. SCHOENFELD, PhD, JUDITH M. GREENE, MPH, SUH-YUH WU, MA,
ERIN O’LEARY, MA, FRANCINE FORTE, AND M. CRISTINA LESKE, MD, MPH

PURPOSE: Recruiting participants is a major challenge for population studies. We present the recruit-
ment methods followed by the Diabetic Retinopathy Awareness Program (DRAP), a community-based,
randomized, masked, controlled trial to meet and exceed its sample size goals.
METHODS: A county-wide multi-media promotional campaign to recruit and enroll participants in
the trial was planned and executed from October 1993 through April 1994, with the assistance of the
local news media and community and professional groups. A toll-free 800 number recruitment line was
established, and postage-paid recruitment postcards distributed. The trial was designed to examine
whether a mailed educational intervention could increase compliance with vision care guidelines among
persons with diabetes in the community.
RESULTS: A total of 2308 persons with diabetes were interviewed for eligibility and 813 enrolled in
the intervention trial, exceeding the original recruitment goals of 1800 and 600, respectively. Those
who completed the enrollment interview reflected county demographics. During recruitment, newspaper,
television and radio stories featured the study; pharmacies and physician offices displayed study materials;
public service announcements appeared in local print and broadcast media. The largest single recruitment
response was a local television news report, followed by a newspaper story.
CONCLUSIONS: These experiences substantiate the need for a comprehensive coordinated approach,
using planned multiple sources, to achieve recruitment success. By engaging the lay and professional
communities along with the media, recruitment costs can be kept to a minimum. Participant costs can
be minimized by employing a toll-free number and eliminating study participant travel, thus allowing
for inclusion of traditionally underserved populations. This approach is applicable to other studies,
where community-based participation is desired.
Ann Epidemiol 2000;10:432–440.  2000 Elsevier Science Inc. All rights reserved.
KEY WORDS: Recruitment Methods, Community-Based Studies, Diabetic Retinopathy, Randomized
Controlled Trials.

INTRODUCTION Studies that have depended primarily on physician refer-

Recruitment is a major challenge for community-based stud-
ies. In recent years, as exemplified by the Women’s Health
Initiative (WHI) (1), recruitment sources for clinical re-
search have expanded beyond medical institutions into the
community. Because identifying potential participants from
a population has become more critical for the success of
clinical research, there is a need to optimize methods to
attract appropriate populations and sustain public awareness
of the study until the recruitment goal is achieved.
From the Department of Preventive Medicine, University Medical Cen-
ter at Stony Brook, Stony Brook, NY.
Presented at the 1998 American College of Epidemiology Annual Meet-
ing, San Francisco, CA.
Address correspondence to: Elinor R. Schoenfeld, PhD, Department of
Preventive Medicine, University Medical Center at Stony Brook, Z⫽8036,
Stony Brook, NY 11794-8036. Tel.: 631-444-2187; fax: 631-444-7525.
E-mail: eschoenfeld@uhmc.sunysb.edu
Received March 26, 1999; accepted November 1, 1999.
ral, hospital records, or clinical laboratories for recruitment
have reported difficulties in accruing sufficient numbers of
participants (2). Elements essential to successful recruitment
have included: awareness of the study—by both the public
and professional communities; extensive preplanning;
knowledge of the target population; recruitment monitor-
ing; and active involvement of trial investigators and staff
(2–9). Some studies that have provided details about recruit-
ment methodology have reported that the majority of parti-
cipants were referred through written and television news
stories (10–13). Other studies have employed various media
sources, including mass mailings, and television and radio
announcements to familiarize the community with the trial
before recruitment was to begin (14, 15). One possible limi-
tation of using the electronic and print media for recruit-
ment is that a high percentage of potential volunteers may
not meet entry criteria (16).
This report presents the methods of the Diabetic Retino-
pathy Awareness Program (DRAP), a community-based

 2000 Elsevier Science Inc. All rights reserved. 1047-2797/00/$–see front matter
655 Avenue of the Americas, New York, NY 10010 PII S1047-2797(00)00067-3
AEP Vol. 10, No. 7
October 2000: 432–440
trial, in identifying and enrolling participants and discusses
the numerous challenges to achieve and exceed sample size
goals. To achieve its aims, DRAP needed to identify and
recruit persons with diabetes from the community who were
non-adherent to current vision care guidelines. While striv-
ing to maintain masking of the trial’s main outcome mea-
sures, DRAP aimed to obtain a population sample that was
demographically representative with respect to age, gender,
race and ethnicity. DRAP’s experiences of establishing a
network of organizations and individuals to identify and
recruit persons with diabetes may be applicable to other
community-based programs or settings, which are similarly
interested in locating participants/patients from the gen-
eral population.
METHODS
Background
DRAP is a five-year controlled, community-based interven-
tion trial conducted in Suffolk County, New York. Through
a masked randomized double-arm design, the trial examined
the effects of an educational intervention to improve adher-
ence to vision care guidelines for persons with diabetes, as
measured by obtaining annual comprehensive eye examina-
tions during the study period. Individuals were eligible for
the trial if they were not adherent to American Academy
of Ophthalmology (AAO) and American Diabetes Associa-
tion (ADA) vision care guidelines (i.e., they did not have
a comprehensive eye examination with dilatation within
the year prior to their enrollment interview) (17, 18). The
main outcome of the trial was defined as obtaining annual
comprehensive vision care during the intervention period.
Participants enrolled in the trial were randomized to either
a multiple educational intervention or a control arm and
followed for at least two years, to assess adherence to vision
care guidelines. All recruitment and follow-up data collec-
tion was accomplished via telephone interviews and all
educational materials were distributed to participants
through the mail.
Recruitment Goals
Study recruitment was open to all county residents, age 18
or older, with physician-diagnosed diabetes, independent of
diabetes type, time since diagnosis, or diabetes management
regimen. The recruitment pool was therefore comprised of
an estimated 50,000 Suffolk County residents with known
diabetes (19–21). The original recruitment goal was to en-
roll a total of 600 participants into the trial, including 300
randomized into each of the two trial arms. To achieve
this goal, an estimated 1800 persons were to be identified,
screened and interviewed over a seven-month period (Octo-
ber 1, 1993–April 30, 1994). The study did not selectively
Schoenfeld et al. 433
DRAP: COMMUNITY RECRUITMENT
exclude non-English speaking individuals who met other
eligibility criteria, although the educational materials were
only available in English. A few telephone interviews were
conducted in Spanish and Italian by study staff to accommo-
date non-English speaking callers.
Recruitment into DRAP presented some challenges to
trial investigators. No organized or predefined network of
professionals existed to assist in participant identification.
In addition, much of the electronic and print media coverage
for Suffolk County overlaps with that of the larger tri-state
metropolitan region, which includes Long Island, New York
City, and other regions of New York State, northern New
Jersey and southern Connecticut. As a result, the challenge
was to target a message to persons with diabetes within
Suffolk County itself. Given the size of the county, which
covers 911.2 square miles and has a population of 1.3 million
(22), there was concern about overloading study resources,
since a large number of persons without diabetes or not
residing in Suffolk County might call for information about
diabetes or to join the study. Thus, it was necessary to select
a focused and targeted recruitment message.
Recruitment Strategies
During the start-up phase, the investigators conducted a
group discussion with nine residents from the adjoining
county, Nassau, who had diabetes of various types, duration,
and management regimens. This discussion group was con-
vened to evaluate the design and phrasing of promotional
materials, to discuss possible incentives to persuade potential
participants to contact the study, and investigate potential
educational materials for use during the study. The discus-
sion contributed valuable information to the process of study
promotion, recruitment, and selection of the types of diabe-
tes-related information to be distributed to study partici-
pants. Based upon the group’s recommendation, a toll-free
800 number was established, and postage-paid recruitment
postcards were provided to eliminate charges to potential
study participants, which could deter participation for senior
citizens and other individuals with fixed or low incomes. In
keeping with the study’s recruitment campaign slogan: “If
you have diabetes, you may not know all you should,” 1-800-
KNOW-MORE was selected as the recruitment telephone
number to provide an easily remembered mnemonic phrase.
The study investigators hired an advertising and graphic
design firm to create study posters, brochures and promo-
tional messages using suggestions synthesized during the
group discussion session. Discussion group participants em-
phasized the need to better understand medical terms related
to diabetes, which led to the acquisition and inclusion of
copies of the Diabetes Dictionary (23) among the literature
sent to all participants. In addition, a diabetes resource guide
was developed to provide a comprehensive compilation of
sources for diabetes products and services. This guide in-
434 Schoenfeld et al.
DRAP: COMMUNITY RECRUITMENT
cluded each organization’s complete address, telephone
number, and a contact person, in most cases. In addition
to educational materials, recruitment incentives included
coupons and samples of diabetes-related products, and a
chance to win a diabetes-related product as a gift. To this
end, manufacturers of such products and supplies were con-
tacted and asked to donate coupons, product samples and
larger items for a participant drawing.
The county-wide multimedia recruitment campaign was
executed between October 1, 1993 and April 30, 1994 to
introduce the study to the public and recruit participants.
During this phase, investigators employed a public relations
firm to disseminate information about the study to the media
with the publication of study announcements and news
stories in local newspapers, community and specialty news-
letters, coverage on local radio and television, and through
announcements in free local home-delivered weekly news-
papers. Prior to this launch, health care professionals and
community organizations that routinely provide diabetes-
related services were contacted by letter and telephone to
develop awareness of the study, assess general interest, and
build community support. The recruitment network was
subsequently broadened to include local businesses, addi-
tional health care professionals and facilities, the ADA, the
Juvenile Diabetes Foundation (JDF), the local public utility
company and local government agencies. A comprehensive
listing of primary care physicians who care for persons with
diabetes, and diabetes specialists on Long Island was created
using national and local medical society listings, and tele-
phone books; a list of other organizations that have contact
with persons with diabetes was developed in a similar man-
ner with the assistance of a certified diabetes educator from
the community and the public relations firms. These various
organizations and businesses were asked to display recruit-
ment posters and brochures, or place a study announcement
in publications to their respective members, customers, or
employees. Recruitment posters, fliers and brochures were
displayed throughout the recruitment period to promote the
study and encourage participation. For masking purposes,
the study was publicized in the community as the Suffolk
County Diabetes Study, with no specific mention of ocular
complications (the primary aim of the trial), in an attempt
to minimize selective inclusion into the study. The study
office was staffed between 8:30 a.m. and 7 p.m. weekdays to
take recruitment inquiries and complete entrance inter-
views. The office was equipped with a three-telephone rol-
lover system, with voice mail to eliminate the possibility of
a busy signal or unanswered calls. Inquiries received by
postcard were to be followed up by telephone the next
business day. All persons contacting the study who met age,
diabetes history and residency requirements were inter-
viewed over the telephone to obtain information on atti-
tudes, beliefs and practices with respect to diabetes manage-
ment and vision care. At the start of each interview, the
AEP Vol. 10, No. 7
October 2000: 432–440
interviewer read a list of potential recruitment sources and
asked the caller how they first learned about this study,
where else they might have heard about the study, and
which of all the sources mentioned specifically led them to
contact the study. All study recruitment brochures were
pre-numbered, and a distribution log was maintained, to
track the source of returned postcards, which was factored
into the analysis of recruitment source.
RESULTS
Study Promotion
During the seven months of recruitment and enrollment,
promotional information about the study appeared via pub-
lic service announcements in 44 different publications; nine
feature stories were printed in local/regional newspapers;
two news stories were aired on local television; four inter-
views were broadcast on radio; and repeated public service
announcements (PSAs) appeared on television and radio.
During all radio and television interviews, special emphasis
was placed on discussing the risk of diabetes among African
Americans, Hispanics and Native Americans and the inter-
est of the study to recruit members of these ethnic minorities.
Study posters were displayed in many locations throughout
Suffolk County. These included: all 252 pharmacies, 47
food stores/supermarkets; 45 health care facilities, 104 pri-
mary care physicians’ and endocrinologists’ offices, and all
eight county health clinics. Another 30 posters were on
display at various other locations within the county. Study
staff attended local health fairs to distribute study brochures
and screened inquirers for potential eligibility. The ADA
and JDF printed a request for participants in their respective
membership newsletters. The local public utility company
included a study announcement in two billing cycles mailed
to the 660,000 households it services. The study also enlisted
the assistance of the county government and the Suffolk
County Office on Aging to promote the study to senior
citizens through nutrition and senior centers, Meals-on-
Wheels and heat assistance program recipients, as well as
to county employees and users of county-administered medi-
cal services.
Recruitment and Enrollment
The recruitment goal of identifying potential participants
through community-wide involvement was attained and
sample size (1800 interviewed, 600 randomized) was ex-
ceeded by the middle of the fifth month (February 1994)
of recruitment. Consequently, upon mutual agreement be-
tween the funding agency and study investigators, enroll-
ment was extended to accommodate the extraordinary vol-
ume of interested individuals. A total of seven months was
originally allotted for the recruitment effort, thus despite
AEP Vol. 10, No. 7
October 2000: 432–440
FIGURE 1. Distribution of inquiries to DRAP during the recruitment period October 1, 1993–April 31, 1994 (n ⫽ 3424 inquiries).
the extension, not everyone interested could be interviewed
by the close of the recruitment/enrollment period on April
30, 1994.
Throughout study recruitment, the investigators mon-
itored the distribution of study participants by age, gender
and ethnicity. In so doing, if at any time a focused effort
was needed to increase recruitment within a select group, it
could be accomplished by refocusing the publicity campaign.
For example, during the February 1994 television news seg-
ment, the risk of diabetes and diabetic complications among
African Americans, Hispanics and Native Americans was
discussed and minority participation in the study was re-
quested. As a direct result of this interview, the number of
minority participants increased. Because gender and eth-
nicity were not determined until the interview, such infor-
mation was not available for the excess of interested recruits
who could not be interviewed. Thus the true impact of
specifically inviting minorities to participate in the trial
could not be fully determined.
A total of 3424 inquiries about study participation were
received during the enrollment period. Figure 1 shows the
distribution of inquiries regarding study participation by
week. The greatest recruitment activity occurred during
February 1994. The first flurry of recruitment activity oc-
curred the week of February 6 with a total of 561 calls
logged during that time period. The majority of the first
Schoenfeld et al. 435
DRAP: COMMUNITY RECRUITMENT
peak (522 of the 561 calls) can be attributed to a local
television news story about the study, which appeared every
hour on the half hour between 9:30 a.m. and 10:30 p.m. on
February 9, on the news round-up the following morning,
and on the weekly round-up the following weekend. In the
24 hours following the February 9 broadcast, 522 telephone
inquiries were received. The second recruitment peak oc-
curred the week of February 13 after a news article about
the study appeared on February 15 in the only regional
daily newspaper with a Suffolk County edition. Following
publication of the newspaper article, 454 inquiries were
processed in the subsequent 24-hour period. The remaining
129 inquiries logged in during this time were attributed to
the television news story from the previous week, in addition
to other regular sources of participant recruitment (e.g.,
local newspaper PSAs and pharmacies).
Figure 2 presents the recruitment outcomes. Of the 3424
inquiries logged in during the enrollment period, 2560 in-
quiries came from individuals potentially eligible for inclu-
sion in the randomized educational intervention trial; an
additional 395 persons who inquired about the study were
ineligible for an enrollment interview and ineligible for
inclusion in the trial. Specifically, 107 (4%) could not be
recontacted for interview due to disconnected, wrong or
missing telephone numbers, having moved prior to re-
contact; or remained uncontacted after 10 attempts at vary-
436 Schoenfeld et al.
DRAP: COMMUNITY RECRUITMENT
FIGURE 2. Results of participant recruitment.
ing times of the day and days of the week; 9 (⬍1%) were
too ill to participate; and 279 (10%) were ineligible because
they did not meet age, residence or history of diabetes re-
quirements. A total of 469 (14%) contacts could not be
interviewed due to the suspension of interviewing once the
enrollment goal (sample size) was exceeded. These individu-
als were recontacted, thanked for their interest in the study,
and if they agreed, placed on a list for future contact. Of
the 3424 inquiries the majority were received via the 800
number recruitment line (77%, n ⫽ 2624). Because it was
possible for one individual to inquire multiple times prior
to interview, the figures in the top half of the flow chart
are inquiry-based and person-based on the bottom half.
Overall 2308 persons (84% of all inquiries) were inter-
viewed for potential enrollment in the intervention trial,
representing 2477 inquiries (169 inquiries from this group
AEP Vol. 10, No. 7
October 2000: 432–440
were repeat contacts), while another 3% (n ⫽ 83) met
eligibility criteria, but declined to participate. Therefore,
among those eligible to participate, the overall participation
rate was 96.5% (total persons interviewed/total persons eli-
gible for interview ⫽ 2308/2308 ⫹ 83).
Those who completed the enrollment interview (n ⫽
2308) ranged in age from 18 to 91 years (median ⫽ 55
years), with a duration of diabetes of less than one week to
70 years (median ⫽ 6 years) (Table 1). While slightly more
women completed the baseline interview (54%), more men
became eligible for the trial because of their lack of a com-
prehensive eye examination within the 12 months prior to
interview (51%, p ⬍ 0.001). The distributions of the study
population by ethnicity (13% minorities) and educational
status (45% with at most a high school diploma) were similar
to those reported in the 1990 census for Suffolk County
AEP Vol. 10, No. 7 Schoenfeld et al. 437
October 2000: 432–440 DRAP: COMMUNITY RECRUITMENT

TABLE 1. Demographic Characteristics (n ⫽ 2308)a TABLE 2. Self-reported recruitment source (n ⫽ 2906)a

Demographics Percentage Recruitment Source n (%) n (%)

Age (years) Newspaper/magazine 877 (30.2)

Mean 54.1 ⫾ 14.3 Regional daily newspaper 410 (14.0)
Median (range)a 55 (18–91) Free home delivered local newspapers 289 (9.9)
Duration of diabetes (years) Other local newspapers 178 (6.1)
Mean 9.6 ⫹ 9.7 Health-related 651 (22.4)
Median (range)a 6 (1 day ⫺ 50 ⫹ years) Pharmacy 344 (11.8)
Gender Physician/hospital 191 (6.6)
Male 45 Others (ADA, etc.) 116 (4.0)
Race Friend/family 456 (15.7)
Native American 1 Newspaper/magazine 95 (3.3)
African American/Black 6 Health related 64 (2.2)
Hispanic 5 Television 50 (1.7)
Caucasian 87 Work 26 (0.9)
Other 1 Radio 25 (0.9)
Education Mailed announcements 21 (0.7)
⬍ 8th grade 1 Government 9 (0.3)
Some high school 10 Poster displays 8 (0.3)
High school graduate 34 Does not recall/not specified 158 (5.4)
Some college 27 Television 442 (15.2)
College graduate 14 Radio 129 (4.4)
Beyond college 14 Mailed announcement
(e.g., utility company) 95 (3.3)
a
n ⫽ total number of participants.
Government 81 (2.8)
Poster displays
(e.g., supermarkets, library) 57 (2.0)
(19). A total of 813 persons interviewed (35%) were found Work 49 (1.7)
Do not recall/not specified 69 (2.4)
eligible for inclusion in the trial. These 813 participants
were randomized to an intervention (n ⫽ 407) or noninter- a
Total count of all responses from 2308 participants. Each participant may
vention (n ⫽ 406) trial arm. have had more than one response for recruitment source.

Recruitment by Source
Table 2 summarizes the sources through which participants
stated they learned about the study. Because each partici-
pant could have been exposed to more than one recruitment
source, numbers presented in this table are based on the
total number of recruitment sources (n ⫽ 2906) for the
2308 study participants. The recruitment source most often
reported by the participant was reading about the study in
a newspaper or magazine (n ⫽ 877, 30.2%); 14% in the
regional daily newspaper; 10% in free home-delivered
weekly papers; and 6% in other local weekly papers. Of the
651 (22.4%) participants who reported having learned about
the study from health-related sources, 11.8% found study
recruitment materials in a pharmacy, 6.6% in a physician’s
office or hospital and 4.0% from other health-related
sources, such as the ADA. The third most frequent source,
15.7%, resulted from receiving the study’s 800 number from
friends or family members. Of this group, most respondents
reported not remembering or knowing where their friends
or family members heard about the study (5.4%). Partici-
pants who did recall where their friends or family members
heard about the study cited newspapers most frequently
(3.3%), followed by health-related facilities (2.2%) and tele-
vision (1.7%). An additional 15.2% saw the local television
news story about the study, and the remaining reported
sources which included: radio, mailed announcements, the
local government, supermarkets, libraries and work.
We were interested in determining which source(s) most
influenced participants to contact the study. On interview,
a total of 545 (23.6%) people mentioned that they learned
about the study from more than one source, with the combi-
nation of television and newspaper most often cited. For
this group of persons, we determined which source they
believed led them to contact the study. To determine a
source’s potential influence, we used the lead response for
those reporting more than one recruitment source and for
those reporting only one recruitment source, we used that
source as the lead source. The results of this analysis appear
in Table 3.
Most frequently mentioned as lead source was friends or
family (15.3%), followed by television news story (15.2%),
announcements in area pharmacies (12.8%), the regional
daily newspaper story (12.5%) and stories/announcements
in the free home-delivered local newspapers (10.0%). These
five sources led 65.8% of participants to contact the study.
Only 44 persons responded that multiple sources triggered
their contacting the study. No particular source was found
to have had a higher yield of participants by age, duration
of diabetes, sex, race, or eligibility for enrollment in the
intervention trial.
438 Schoenfeld et al. AEP Vol. 10, No. 7
DRAP: COMMUNITY RECRUITMENT October 2000: 432–440

TABLE 3. Sources leading to participant contact (n ⫽ 2308)a While a large investment of time and resources is nec-
Led to contact essary to develop community relationships and trust, it also
Source n (%) takes a great deal of effort to maintain positive relationships
once study recruitment ceases. For DRAP, community in-
Friend/family 354 (15.3)
TV 350 (15.2) quiries continued even after the close of recruitment, with
Pharmacy 295 (12.8) over 450 inquiries documented, presenting multiple chal-
Regional daily newspaper 288 (12.5) lenges. It was necessary to develop a referral network to provide
Free home-delivered local newspapers 231 (10.0) information on community diabetes-related resources, yet
Physician/hospital 143 (6.2) at the same time leaving the door open for possible future
Other local newspapers 122 (5.3)
Mailed announcement (e.g., utility company) 86 (3.7) recruitment efforts. Though infrequent, study inquiries are
Radio 81 (3.5) still received, even though more than four years have elapsed
Other health-related source (ADA, etc.) 81 (3.5) since the close of recruitment. This problem of shutting off
Government 80 (3.5) the flow of inquiries has not been widely discussed by previ-
Any combination 44 (1.9) ous investigators and thus warrants further consideration.
Work 40 (1.7)
Poster displays (supermarket, library, etc.) 38 (1.6) The diabetes discussion group undertaken by DRAP
Don’t remember/NOS 75 (3.2) proved very helpful in the final selection of recruitment
a
n ⫽ total number of participants. materials, promotional messages and incentives, and educa-
tional materials selected for distribution to study partici-
pants’. Similar to the discussion session experiences of the
Sixty Something . . . project (26), and of other clinical
DISCUSSION studies cited in the review article by Lovato et al. (3), the
Recruiting a sufficient number of eligible study participants DRAP discussion session provided well-defined recommen-
from the general population requires considerable prepara- dations for the recruitment process, and elucidated which
tion and effort. Success in clinical trial recruitment from educational materials would be of most interest to those
community sources requires an advance understanding of with both newly diagnosed and longer-term diabetes.
the target population, establishing community networks, as Effective use of the news media to assist in study recruit-
well as the ability to convey a focused and evocative message ment has been reported by several other diabetes studies as
to interest potential participants. The recognition of the well. The Sixty Something . . . Diabetes Self Care Program
need for a development and planning phase, as emphasized found that participants recruited through the media were
by other investigators (3, 6, 9, 24), proved to be a critical twice as likely to participate in the study as compared to
component of DRAP’s recruitment success. During the one- those who learned about the study through passive means
year lead time prior to recruitment, investigators met with (i.e., project initiated recruitment methods), and that re-
community and professional groups to establish a coopera- cruitment reflected the demographic complexity of persons
tive network, while heightening awareness about diabetes with diabetes in their general target population (26). Simi-
and DRAP in the community. Similar to the experiences larly, the Diabetes Care for Older Adults project reported
reported by Fitzgibbon et al. (6), DRAP became successful that about three-fourths of its inquiries were generated as
in disseminating information about the study and recruiting a result of press releases and newspaper advertisements. An-
participants only after establishing a relationship with com- derson et al., found that these advertisements were one of
munity leaders and the media. the least expensive forms of recruitment and accounted for
The recruitment campaign was designed with promo- 80% of those enrolled in their study (27). As exemplified
tional messages targeted only to persons with diabetes who in Table 2, DRAP also found the greatest response from
resided in the county of interest. To maximize the use of the the media, with about 50% of participants being recruited
limited resources available for study promotion, it became from print and electronic news stories and PSAs. In contrast
necessary to develop a strategy for reaching the largest num- to the projects cited above, however, media coverage for
ber of potentially eligible individuals, while minimizing con- DRAP recruitment was at no cost to the study.
tacts from ineligible persons and sustaining the message over Reports from clinical trials of various other diseases and
the entire seven-month recruitment period. The use of local conditions further support the successful role of the media
advertising and public relations firms to develop the promo- in participant recruitment. One of the earliest trials to report
tional messages and attract media attention proved a cost- recruitment statistics by source was the Coronary Drug Proj-
effective investment to promote the study. Similar experi- ect (CDP). By the midpoint of the study, only 23% of the
ences have been documented by other researchers (3, 8, recruitment goal had been achieved, but was later reached
25). No additional costs were incurred to promote the study after initiating a mass media campaign, while sustaining all
since the toll-free recruitment number was included in all the earlier recruitment strategies (16). Similarly, Brecken-
study-related news stories and PSAs. ridge et al. reported that a single newspaper story about the
AEP Vol. 10, No. 7
October 2000: 432–440
study accounted for 67% of recruits (total recruitment of
71% through news media sources) (12). They concluded
that the public seemed to prefer a lengthy description of
the trial in writing before deciding whether to contact inves-
tigators. Recruitment for the Aspirin Myocardial Infarction
Study (AMIS) was targeted at the general public from the
start, including mass mailings in utility bills, public rallies,
radio and television announcements and programs, repeated
newspaper articles, and paid advertisements. Even with
strict inclusion criteria, one-third of the patients were re-
cruited directly through the mass media (16, 28). Lee et al.
(25) concluded that passive methods of participant recruit-
ment, which included use of the media, fliers and brochures,
were found to be both more effective and more efficient
than active recruitment methods. Similarly, DRAP’s use
of a focused mass media campaign with strict disease and
residency requirements proved highly effective in reaching
the target audience, as reflected by the fact that 84% of
inquiries resulted in a completed interview and the partici-
pation rate of 96%. This success was translated into eco-
nomic savings, in that we did not have to screen a significant
number of ineligible individuals. In a clinical trial of early
Parkinson’s disease (1989), the primary referral sources in-
cluded the news media (56.9%), physicians (27.2%), and
other referrals (15.9%). A central toll-free number for infor-
mation and referral for study participation was found effec-
tive (10, 25). In DRAP, the use of a toll-free telephone line
eliminated two potential barriers to participation, namely
cost and distance. These are potential obstacles for any
population-based study with a large catchment area or which
includes older, homebound or disabled persons, as might be
expected in a population with diabetes, who may be unable
to travel distances or afford the cost of a lengthy tele-
phone call.
Flexibility and a readiness to change recruitment strat-
egies during the recruitment period were found to be critical
to DRAP’s enrollment success. The experiences of DRAP
support the recommendations from previous studies to closely
monitor the demographic composition (e.g., age, gender, race/
ethnicity) of participants as they are recruited, and to be ready
to modify recruitment messages, focus and targets as needed
to achieve the study’s recruitment goals (7, 8).
Following the example of the Lipid Research Clinics
Coronary Primary Prevention Trial (LRC-CPPT) (2, 5, 29),
DRAP sought to achieve and attained geographic represen-
tativeness by using multiple recruitment sources to dissemi-
nate the study message to potential participants from many
different sociodemographic, cultural and ethnic groups
throughout the region. Although this goal was met and the
composition of the DRAP population was similar to the
Suffolk County population, it had a lower proportion of
minorities than the diabetes population. On the other hand,
DRAP might have achieved a higher representation of mi-
norities if recruitment had not been stopped early. The
Schoenfeld et al. 439
DRAP: COMMUNITY RECRUITMENT
LRC-CPPT reported that characteristics of the participants
differed by referral source. For example, the highest percent-
age of professionals were recruited through the news media,
medical referrals, and clinical studies; recruitment of white
collar and skilled workers was highest from mass mailings
and community sources (2, 5, 29). The use of a variety of
recruitment methods to achieve ethnic and gender diversity
in participation has since been documented by other studies
(3, 7, 24, 30, 31).
CONCLUSIONS
To recruit a sufficient number of people from the community
for clinical research, solid community support and diverse
resources must be used to promote a study and attract com-
munity participants. The experiences of DRAP further sub-
stantiate the importance of the mass media (e.g., newspapers
and television) and health-related services (e.g., physicians,
pharmacies) to disseminate health-related information
within the community. By engaging the lay and professional
communities along with the media, recruitment costs can
be kept to a minimum. Employing a toll-free study number,
and minimizing or eliminating travel by study participants,
costs to participants can be lowered or eliminated thus
allowing for inclusion of traditionally underserved popula-
tions (e.g., the elderly, homebound, homeless). As in any
study drawing participants from the community, an ongoing
issue is to maintain good relations with the community once
enrollment in the study has closed, and to provide interested
inquirers with further information about possible commu-
nity resources. Finally, DRAP has shown that there is ade-
quate community support for community-based research and
that similar methods may be employed to other diseases
and other settings where a community-based approach to
participant identification is desired.
Supported by National Eye Institute Grant #EY09891. We thank George
Gerard Associates and Pamela Cordes for their valuable assistance in the
media campaign and to the many manufacturers who donated diabetes-
related products. In addition, many thanks to Karen Glanz, PhD, MPH,
for her assistance in designing the surveys, Joia DiStefano, MPH, Jean
Straub, MS, and Susan Wilk for their assistance in interviewing participants
and Dr. Terry Gray for her assistance in evaluating and selecting educa-
tional materials for the study.
REFERENCES
1. The Women’s Health Initiative Study Group. Design of the Women’s
Health Initiative clinical trial and observational study. Controlled
Clin Trials. 1998;19(1):61–109.
2. Hunninghake DB, Darby CA, Probstfield JL. Recruitment experience
in trials: Literature summary and annotated bibliography. Controlled
Clin Trials. 1987;8(Suppl. 4):6S–30S.
3. Lovato L, Hill K, Hertert S, Hunninghake DB, Probstfield JL. Recruit-
ment for controlled clinical trials: Literature summary and annotated
bibliography. Controlled Clin Trials. 1997;18:328–357.
440 Schoenfeld et al.
DRAP: COMMUNITY RECRUITMENT
4. Spilker B, Cramer J. Patient Recruitment in Clinical Trials. New York:
Raven Press; 1992.
5. Bradford RH. Participant recruitment to the Lipid Research Coronary
Primary Prevention Trial. Controlled Clin Trials. 1987;8:31S–40S.
6. Fitzgibbon ML, Prewitt TE, Blackman LR, Simon P, Luke A, Keys
LC, et al. Qualitative assessment of recruitment efforts for prevention
trials in two diverse black populations. Preventive Medicine. 1998;
27:838–845.
7. Halbert JA, Silagy CA, Finucane P, Withers R, Hamdorf PA. Recruit-
ment of older adults for a randomized, controlled trial of exercise advice
in a general practice setting. J Am Geriatr Soc. 1999;47:477–481.
8. Cosgrove N, Borhani N, Baily G, Borhani P, Levin J, Hoffmeier, et
al. Mass mailing and staff experience in a total recruitment program
for a clinical trial: The SHEP experience. Controlled Clin Trials.
1999;19:133–148.
9. Nathan RA. How important is patient recruitment in performing
clinical trials? J Asthma. 1999;36(3):213–216.
10. Parkinson Study Group. DATATOP: A multicenter controlled clinical
trial in early Parkinson’s disease. Arch Neurol. 1989;46:1052–1060.
11. Arnold A, Johnstone B, Stoskopf B, Skingley P, Browman G, Levine
M, et al. Recruitment for an efficacy study in chemoprevention—the
Concerned Smoker Study. Prev Med. 1989;18:700–710.
12. Steinmetz Breckenridge JS, Zeiss AM, Breckenridge JN, Gallagher
D, Thompson LW. Solicitation of elderly depressives for treatment
outcome research: A comparison of referral sources. J Consult Clinical
Psychol. 1985;53:552–554.
13. Levenkron JC, Farquhar JW. Recruitment using mass media strategies.
Circulation. 1982;66(Suppl. 4):IV-32–IV-36.
14. Baines CJ. Impediments to recruitment in the Canadian National
Breast Screening Study: Response and resolution. Controlled Clinical
Trials. 1984;5:129–140.
15. Petrovitch H, Byington R, Bailey G, Borhani P, Carmody S, Goodwin L,
et al. Systolic Hypertension in the Elderly Program (SHEP): Part 2:
Screening and recruitment. Hypertension. 1991;17(Suppl. 3):II-16–II-23.
16. Schoenberger JA. Recruitment in the Coronary Drug Project and the
Aspirin Myocardial Infarction Study. Clin Pharmacol Ther. 1979;
25:681–684.
AEP Vol. 10, No. 7
October 2000: 432–440
17. American Academy of Ophtalmology (AAO) Newsletter, October
30, 1989.
18. Position Statement: Eye care guidelines for patients with diabetes
mellitus. Diabetes Care. 1997;20(1S):S28–30.
19. United States Census 1990.
20. Centers for Disease Control: Prevalence and incidence of diabetes
mellitus—United States 1980–87. MMWR. 1990;39(45):809–812.
21. Centers for Disease Control: Regional variation in diabetes mellitus
prevalence—United States 1988–89. MMWR. 1990;39(45):805–809.
22. Long Island Almanac. Long Island Business News, LI, NY 1996.
23. Diabetes Dictionary. Washington, DC: US Government Printing Of-
fice. 1989;DHEW publ. no. NIH 89-3016.
24. Patrick JH, Pruchno RA, Rose MS. Recruiting research participants:
A comparison of the costs and effectiveness of five recruitment strate-
gies. The Gerontologist. 1998;38(3):295–302.
25. Lee RE, McGinnis KA, Sallis JF, Castro CM, Chen AH, Hickmann
SA. Active vs. passive methods of recruiting ethnic minority women
to a health promotion program. Ann Behav Med. 1997;19(4):378–84.
26. Glasgow RE, Toobert D, Hampson S. Participation in outpatient diabe-
tes education programs: How many patients take part and how repre-
sentative are they? The Diabet Educat. 1991;17(5):376–380.
27. Anderson L, Fogler J, Dedrick R. Recruiting from the community:
Lessons learned from the Diabetes Care for Older Adults Project. The
Gerontologist. 1995;35(3):395–401.
28. Schoenberger JA. Recruitment experience in the Aspirin Myocardial
Infarction Study. Controlled Clin Trials. 1987;8(Suppl. 4):74S–78S.
29. Hunninghake DB, LaDouceon M, Peterson F. Population characteris-
tics according to recruitment source. Circulation. 1982;66(Suppl. IV):
46–48.
30. Lewis CE, George V, Fouad M, Porter V, Bowen D, Urban N for
the WHT:FSMP Investigators Group. Recruitment strategies in the
Women’s Health Trial: Feasibility study in minority populations. Con-
trolled Clin Trials. 1998;19:461–476.
31. Boult C, Boult L, Morishita L, Pirie P. Soliciting defined populations
to recruit samples of high-risk older adults. J Gerontology. 1998;53A(5):
M379–84.