Recruiting the oldest-old for clinical research

Abstract

The oldest-old, those 90 years and older, are the fastest growing segment of the population and the number of dementia cases at these ages will steadily increase over time. It is therefore critical to include this population in clinical research. Evidence to guide recruitment of this group is scarce. We report our experience recruiting the oldest-old to a longitudinal study of aging and dementia. Recruitment activities were grouped into four strategies: direct mailing of recruitment brochures, community outreach, earned media, and referrals. Recruitment sources were recorded based on enrollees’ self-report. Cost was estimated based on personnel time and materials. One hundred forty five new participants were enrolled over 40 months. Community outreach produced the most recruitment (33.8%) followed by earned media (21.4%), direct mail (16.6%) and referrals (15.8%). Earned media and direct mailing were most cost-effective. Local media produced more enrollment and was more cost-effective than national media.

Introduction

The oldest-old (defined as people age 90 and older) are the fastest growing segment of the population, are at increased risk for dementia, and are distinct from older adults below the age of 90 with respect to clinical and pathological characteristics¹. Therefore to inform clinical care reflective of societal demographics it is important to include the oldest-old in research. Yet, this group is consistently underrepresented in research² and little guidance is available to instruct recruitment.

Initiated in January 2003, The 90+ Study is a longitudinal study of aging and dementia in the oldest-old. It includes comprehensive semiannual neurological and neuropsychological examinations, as well as blood draws, brain imaging, and neuropathological examination¹. The 90+ Study originally enrolled oldest-old participants from the Leisure World cohort study (LWCS) ¹. LWCS is a health survey study of the residents of the Leisure World (now known as Laguna Woods Village), a large retirement community in Orange County, California. Over time open recruitment beyond the LWCS was initiated due to natural dwindling of the cohort. The aim of this brief report is to describe outcomes and costs of the recruitment strategies used in this open recruitment.

Methods

Participants

Inclusion criteria for newly recruited participants are: age 90 or older, residing within an hour drive of the study location, no or mild dementia, and no contraindications to brain imaging. Participants must be willing to undergo semi-annual neurocognitive evaluations and consider brain donation.

Recruitment Strategies and Activities

We report recruitment performed between January 2014 and April 2017. Recruitment activities were categorized into four strategies: community outreach, earned media, direct mailings, and referrals.

(1). Community outreach.

(a) During the recruitment period the study team delivered 46 community presentations on longevity and cognitive health highlighting findings from The 90+ Study (10–22 presentations per year), open to the general public. Four presentations were given in two retirement communities near the study location, others at various community organizations in the area. At each presentation, recruitment brochures were distributed enabling attendees to contact the study team if interested in enrolling.

(b) The study team organized one recruitment event in the Laguna Woods Village retirement community. Invitations were sent to 1,733 individuals age 90 and older living in Laguna Woods Village and two neighboring zip codes. Addresses were rented from a contracted mailing house. Oldest-old individuals were invited to learn about the study and consider participation. The event, held in September of 2016, featured study investigators, who presented the study and answered questions. Sixteen current study participants and relatives shared their experiences. Fifty-four potentially eligible individuals (3% of those who received invitations) attended the event. Attendees interested in joining the study were given information packets including consent forms and other study materials.

(c) Study personnel distributed 25–50 brochures to each of 10 local retirement communities and senior centers.

(2). Earned media.

The study was featured in the media 9 times during the reporting period. We categorized media as (a) national, including CBS 60 minutes (aired twice), The Los Angeles Times, and a web podcast, or (b) local, including television, radio, and newspaper. Local media coverage included television and newspaper outlets operated by Laguna Woods Village for its residents. In most cases, coverage was initiated by the media outlet following public presentations or press releases of research publications. An exception was coverage by the Laguna Woods Village outlets, which was initiated by the study team.

(3). Direct mailings.

Three direct mailings, separate from mailed invitations to the recruitment event, were performed through a contracted mailing house. Mailings were sent to 16 different zip codes within an hour drive of the study location. A first mailing, in August 2015, was sent to the households of 389 individuals age 96 and older. Subsequent mailings targeting those 90–95 years old were sent in December 2016 to 2,250 households and in March 2017 to 2,350 households. Mailed materials included a study brochure with a postage-paid contact information card.

(4). Referrals.

Active study participants and collaborating investigators of the University of California, Irvine Alzheimer’s Disease Research Center (ADRC) and Willed Body Program referred 90+ individuals to the study.

Data Collection and Analysis

Recruitment sources were captured though participant self-report at enrollment. To estimate the total cost for each recruitment activity, we summed personnel and material expenses. The cost of personnel time was calculated as estimated staff or faculty time performing target activities (delivering presentations, being interviewed for a show, answering reporters’ questions) multiplied by standard university hourly rates ($128 for faculty and $47 for staff, including benefits). Staff time organizing recruitment activities was not included in personnel cost. Material costs were actual dollar amounts for printing, postage, food and beverages, and venue rental. Earned media appearances were free of charge; the only cost was personnel time. Average cost per recruited participant was calculated by dividing total cost of recruitment activity by the number of participants recruited through that activity.

Results

A total of 145 new participants were enrolled over 40 months (mean 3.6 participants/month, range 0–15) (Figure). Mean age of the enrolled participants was 93.1±2.9 years (range: 90–110). Most participants were female (59.2%) and Caucasian (96%). Three percent were Asian and 1.5% were Hispanic/Latino. Most of the participants (53%) had completed college, 31% had attained a post-baccalaureate education, and 16% had not attended college.

Figure.

Number of recruited participants by month (solid line) and cumulatively (dotted line). Letters indicate timing of major recruitment activities: CBS - CBS 60 minutes (aired twice), NP – retirement community paper (two publications), TV – retirement community TV, DM – direct mailing (three mailings), RE – recruitment event. Between January 2014 and peak 1: 6 presentations, 1 earned media (7 activities total). Between peaks 1 and 2: 16 presentations, 4 earned media (20 activities total). Between peaks 2 and 3: 19 presentations, 1 recruitment event, 1 mailing, 4 earned media (25 activities total). Between peaks 3 and 4: 2 presentations, 2 mailings, 2 earned media (6 activities total).

Data on recruitment source was available for 87% of new enrollees; two participants (1.4%) reported more than one source. These individuals were categorized based on the recruitment activity that immediately preceded their enrollment.

Each strategy resulted in recruitment (Table). Among the four recruitment strategies community outreach produced the highest yield (49 participants, 33.8%) followed by earned media (32, 22.1%) and direct mailings (24, 16.6%). Among specific recruitment activities, community presentations (24, 16.6%), mailings (24, 16.6%), and local earned media (24, 16.6%) had highest yields. Among those recruited through earned local media, each participant reported a local retirement community newspaper as the source of recruitment. Similarly, among those recruited through earned national media, only one outlet (CBS 60 minutes) was reported as a source. The direct mailing to the older age group (96+) produced only one enrollee (0.2% of households receiving the mailings), subsequent mailings to younger participants yielded 23 enrollments (0.5% of households).

Table.

Number Recruited and Cost-Effectiveness by Recruitment Strategies and Activities

Recruitment Strategies and Activities	Number of Events	Participants Enrolled N (%)	Total Cost (USD)	Cost per Participant (USD)
1. Community outreach		49 (33.8)	14,786	261
(a) Presentations	46	24 (16.6)	7,988	333
(b) Recruitment event	1	20 (13.8)	6,063	303
(c) Brochures at organizations	10	5 (3.4)	735	147
2. Earned media		31 (21.4)	4,093	132
(a) Local media	6	24 (16.6)	734	31
(b) National media	8	7 (4.8)	3,359	480
3. Direct mailing of recruitment brochures	3	24 (16.6)	6,191	258
4. Referrals		23 (15.8)	-	-
By The 90+ Study participant/staff	-	17 (11.7)	-	-
Clinical referrals	3	5 (3.4)	-	-
The 90+ Study web-site	-	1 (0.7)	-	-
Unknown	-	18 (12.4)	-	-
Total	-	145 (100)	25,070	173

The total estimated cost of this recruitment campaign was $25,070; the mean cost per enrollee was $173 (Table). Among strategies associated with cost, earned media ($132) was most cost-effective; community outreach was least cost-effective ($261). Local earned media ($31) and distributing recruitment brochures at senior organizations ($147) were the most cost-effective recruitment activities; national earned media had highest cost per participant ($480).

Discussion

This report demonstrates the feasibility of recruiting individuals age 90 and older into longitudinal clinical research and provides preliminary guidance related to the effectiveness of recruitment strategies. The recruited sample is representative of the US oldest-old with respect to gender but is more educated,³ much like other research in this age group⁴. The ethnic composition resembles 2014 census projection for the oldest-old in Orange County, California (74%)⁵, though the percentage of Caucasians may be higher due to the ethnic composition of the specific recruitment areas and other challenges related to underrepresented ethnic groups⁶.

Some recruitment strategies were more effective than others. Consistent with survey research in older adults⁷, community outreach produced the highest number of new enrollees, but at the greatest cost. Recruitment activities with age-appropriate individuals yielded the most enrollees. Thus, a single targeted recruitment event resulted in nearly as many participants as did 46 general public education presentations, and at a lower cost.

After community outreach, earned media produced the most enrollments and did so at lower cost. Local media had greater yield than national media and were associated with less personnel time. National media, however, may have allowed for widespread dissemination of information and initiated trust building for the study, as 6.9% of new enrollees indicated that they first learned about the study through national media but enrolled after a subsequent local recruitment activity (mailing, retirement community newspaper or invited presentation). This “two-hit” model of recruitment⁸ may be critical to establishing trust and overcoming skepticism, which may be key to successful recruitment of this population⁹. Perhaps supporting this notion and in contrast to some experiences recruiting to Alzheimer’s disease prevention clinical trials¹⁰, direct mailings were not a cost-effective recruitment strategy when targeting the oldest old. Direct mail was least effective with those age 96 and older.

Several limitations should be noted. This was not a planned study of recruitment; recruitment activities were not balanced for number, cost, or time. The 90+ Study is well established, highly recognizable, and has the opportunity to recruit from nearby retirement communities with large 90+ populations, which may limit the generalizability of our findings. Twelve percent of enrollees did not remember the recruitment source. There was no way for the study team to assess the validity of participant reports. Nevertheless, this report provides guidance for investigators considering recruitment of 90+ individuals to aging research. Targeted community outreach yielded the most new participants but a comprehensive and multifaceted marketing and recruitment plan may be essential to maximizing enrollment of the oldest old in clinical research.