Facebook advertising for participant recruitment into a blood pressure clinical trial

Abstract

Objectives:
Recruitment of sufficient sample size into clinical trials is challenging. Conventional advertising methods are expensive and are often ineffective. The effectiveness of Facebook for recruitment into blood pressure clinical trials of middle-to-older-aged people is unknown. This study aimed to assess this by comparing Facebook advertising with conventional recruitment methods from a retrospective analysis within a clinical trial.

Methods:
Conventional advertisements (newspaper, radio and posters) were employed for the first 20 months of a randomized controlled clinical trial conducted in three Australian capital cities from Tasmania, Queensland and the Australian Capital Territory. With dwindling participant recruitment, at 20 months a Facebook advertising campaign was employed intermittently over a 4-month period. Recruitment results were retrospectively compared with those using conventional methods in the previous 4 months.

Results:
Compared with conventional recruitment methods, Facebook advertisement was associated with a significant increase in the number of participants recruited in the Australian Capital Territory (from an average 1.8-7.3/month; P < 0.05). There was also an increase in Tasmania that was of borderline significance (from 4.0 participants recruited/month to 9.3/month; P = 0.052). However, there was no effect in Queensland (from 6.0 participants recruited/month to 3.0/month; P = 0.15). Facebook advertisement was associated with a significant decrease in the age of participants enquiring into the study (from 60.9 to 58.7 years; P < 0.001).

Conclusion:
Facebook advertising was successful in helping to increase recruitment of middle-to-older aged participants into a blood pressure clinical trial, although there may be some variability in effect that is dependent on location.

Full text

CE: Swati; JH-D-17-00388; Total nos of Pages: 5;
JH-D-17-00388
Facebook advertising for participant recruitment into
a blood pressure clinical trial
Erin L. Nash
a
, Deborah Gilroy
b
, Wichat Srikusalanukul
c
, Walter P. Abhayaratna
c
, Tony Stanton
d,e
,
Geoffrey Mitchell
e,f
, Michael Stowasser
b
, and James E. Sharman
a
Objectives: Recruitmentofsufficientsamplesizeinto
clinical trials is challenging. Conventional advertising
methods are expensive and are often ineffective. The
effectiveness of Facebook for recruitment into blood
pressure clinical trials of middle-to-older-aged people is
unknown. This study aimed to assess this by comparing
Facebook advertising with conventional recruitment
methods from a retrospective analysis within a clinical
trial.
Methods: Conventional advertisements (newspaper, radio
and posters) were employed for the first 20 months of a
randomized controlled clinical trial conducted in three
Australian capital cities from Tasmania, Queensland and
the Australian Capital Territory. With dwindling participant
recruitment, at 20 months a Facebook advertising
campaign was employed intermittently over a 4-month
period. Recruitment results were retrospectively compared
with those using conventional methods in the previous 4
months.
Results: Compared with conventional recruitment
methods, Facebook advertisement was associated
with a significant increase in the number of
participants recruited in the Australian Capital Territory
(from an average 1.8–7.3/month; P<0.05). There was
also an increase in Tasmania that was of borderline
significance (from 4.0 participants recruited/month to
9.3/month; P¼0.052). However, there was no effect in
Queensland (from 6.0 participants recruited/month to
3.0/month; P¼0.15). Facebook advertisement was
associated with a significant decrease in the age of
participants enquiring into the study (from 60.9 to
58.7 years; P<0.001).
Conclusion: Facebook advertising was successful in
helping to increase recruitment of middle-to-older aged
participants into a blood pressure clinical trial, although
there may be some variability in effect that is dependent
on location.
Keywords: direct-to-consumer advertising, direct-to-
consumer marketing, research subject selection, social
mediums, web 2.0
Abbreviations: ACT, Australian Capital Territory; QLD,
Queensland; TAS, Tasmania
INTRODUCTION
For the appropriate conduct of clinical trials, inves-
tigators must demonstrate that the size and profile of
the study population is adequate to answer the study
question [1] and the potential to recruit the required number
of participants within the recruitment period [2]. Even with
the apparent availability of eligible study participants, the
recruitment of sufficient study sample in a timely manner
often presents a major challenge. Indeed, less than 10% of
clinical studies are completed on time due to poor partici-
pant recruitment and retention [3]. The use of conventional
recruitment methods, in particular press advertisements
and mail outs, can be expensive with no guarantee of
sufficient recruitment. With increasing popularity of online
media [4], web-based social platforms present a potential
opportunity for participant recruitment.
Facebook (Menlo Park, California, USA) is a form of
online social media, in which users are able to interact with
other users by sharing personal information into a news-
feed. As of January 2017, there were more than 1.86 billion
active monthly users of Facebook worldwide, representing
an annual increase of 17% [5]. In 2016, Facebook was the
most frequented website in Australia with users spending
approximately 12.5 h/week visiting the site [6]. Facebook
has been used successfully for recruitment of young par-
ticipants in online surveys [7], longitudinal studies [8,9] and
randomized controlled trials [10], but to our knowledge,
there are no reports on the value of Facebook for recruit-
ment in cardiovascular-related trials of older people. In this
article, we report the results of a retrospective analysis in
which Facebook advertising was included after conven-
tional recruitment methods had become exhausted in a
Journal of Hypertension 2017, 35:000–000
a
Menzies Institute for Medical Research, University of Tasmania, Hobart, Tasmania,
b
Endocrine Hypertension Research Centre, University of Queensland School of Medi-
cine, Princess Alexandra Hospital, Brisbane, Queensland,
c
Australia National Universi-
ty, Canberra Hospital, Canberra, Australian Capital Territory,
d
Sunshine Coast
University Hospital, Birtinya,
e
School of Medicine, University of Queensland, Brisbane
and
f
Limestone Medical Centre, Ipswich, Queensland, Australia
Correspondence to James E. Sharman, PhD, Menzies Institute for Medical Research,
University of Tasmania, Private Bag 23, Hobart 7000, TAS, Australia. Tel: +61 3 6226
4709; fax: +61 3 6226 7704; e-mail: James.Sharman@utas.edu.au
Received 3 April 2017 Revised 5 June 2017 Accepted 21 June 2017
J Hypertens 35:000–000 Copyright ß2017 Wolters Kluwer Health, Inc. All rights
reserved.
DOI:10.1097/HJH.0000000000001477
Journal of Hypertension www.jhypertension.com 1
Original Article
Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

CE: Swati; JH-D-17-00388; Total nos of Pages: 5;
JH-D-17-00388
randomized controlled trial among middle-to-older aged
people treated for high blood pressure (BP).
METHODS
Study participants
The randomized controlled trial was entitled Tar-
geted LOWering of Central Blood Pressure Trial
(ACTRN12613000053729) and required recruitment of
300 participants from three study sites located in Australian
capital cities within Tasmania (TAS), the Australian Capital
Territory (ACT) and Queensland (QLD). Inclusion criteria
included men and women with high BP aged 18–69 years
that were currently taking medication for high BP. Exclu-
sion criteria included if participants were taking more than
three BP medications, had a clinical history of cardiovascu-
lar disease, or were taking medications or supplements that
were contraindicated with the study medication.
Study protocol
The study was approved by local Human Research Ethics
Committees and participants provided informed consent.
At the start of the study, participant recruitment was under-
taken through conventional methods of advertising that
included posters in hospitals and doctors’ clinics, local
newspaper advertisements and inclusion of study informa-
tion slips within antihypertensive prescriptions distributed
at pharmacies. Study information was also made available
via a webpage linked to the TAS study site. In addition,
there were calls for study participants made via several
sources, including radio interviews, local publicly available
seminars and community events. Conventional recruitment
campaigns ran for the first 20 months of the trial, by which
time it became clear that recruitment options were becom-
ing exhausted as the number of enquiries had diminished
significantly. At this point, the Facebook campaign was
initiated at all study sites as an add-on measure to try and
increase recruitment. However, an additional source of
recruitment was also made available at the QLD site in
the months preceding the Facebook campaign, involving
identifying potential participants at a regional general prac-
tise located in Ipswich, 40 km from Brisbane city. The
practise database of approximately 5000 patients was
screened for potentially eligible participants and letters
were sent via the attending general practitioners with a
request to contact the practise, if interested in study partici-
pation.
Facebook advertisements
Intermittent broadcast of Facebook advertisements were
initiated over a 4-month period. Two forms of advertise-
ment were used – first, those placed in the right-side panel
of the Facebook newsfeed and; second, those placed
directly into the newsfeed through the University of TAS
Facebook page. Right-side panel advertisements included
an image (up to five images could be used), a 20-character
title and 95-character text body. The newsfeed advertise-
ments allowed more text to be included in the title and body
sections (refer to Fig. 1 for examples). If people clicked
onto either form of advertisement, they were redirected to
the study website, which contained additional study infor-
mation and site coordinator contact details. Advertisements
were broadcast to a radius of 50 miles (80 km) around the
study sites in TAS and ACT, and a 30-mile radius (48 km)
around the QLD study sites (one each in Brisbane and
Ipswich; lower radius used due to much higher population
in QLD). Advertisements were able to be targeted towards
men and women aged between 35 and 69 years as these
details were included as part of registering for Facebook.
The daily cost for advertisements was capped at Australian
dollars (AUD$)25/day for newsfeed advertisements and
AUD$50/day for right-side panel advertisements in TAS
and ACT, and AUD$30/day for newsfeed advertisements
and AUD$100/day for right-side panel advertisements in
QLD. Once the daily cost had been reached, advertisements
would cease appearing for that day. Cost was calculated per
click onto advertisement which varied from AUD$0.20 to
AUD$1.20/click.
Statistical analysis
Success of Facebook advertising was gauged by a change in
the number of new enquiries and number of new partic-
ipants recruited at each site in the 4-month period using
Facebook advertising compared with the 4 months preced-
ing Facebook advertising. With respect to age of partici-
pants, we compared the values for the 20-month
conventional recruitment period with the values for the
Facebook recruiting period. Differences between groups
were assessed using Student’s two-tailed ttests for contin-
uous variables and independent two-tailed ttest was used
on age related data. Data has been presented as mean
standard deviation unless otherwise specified and statistical
significance was denoted by Pvalue less than 0.05.
RESULTS
Table 1 presents Facebook statistics related to the 4-month
advertising period. Figures 2 and 3 present site-specific
comparisons on the number of enquiries and participants
recruited related to Facebook advertising. In ACT, this was
associated with a significant increase in the average number
of enquiries to participate in the study (from 6.5 to 22.0
enquiries/month; P<0.05) and average number of particip-
ants recruited (from 1.8 to 7.3 participants recruited/month;
P<0.05) when compared with the previous 4-month peri-
od only using conventional advertising methods. In TAS,
there was a significant increase in the average number of
enquiries to participate in the study (from 12.5 to 28.5
enquiries/month; P<0.05) and a trend towards an increase
in average number of participants (from 4.0 to 9.3 partic-
ipants recruited/month; P¼0.052). However, in QLD, there
was a nonsignificant decrease in the average number of
enquiries from pre to post Facebook advertising (from 22.8
to 11.0 enquiries/month; P¼0.23; Fig. 2). Similarly, for
QLD, there was a nonsignificant reduction in the number
of participants recruited from pre to post Facebook adver-
tising (from 6 to 3/month; P¼0.23; Fig. 3). The reversal in
trends at the QLD site appears to be related to the higher
average numbers of enquiries (Fig. 2) and participants
recruited (Fig. 3) in the 4 months preceding initiation of
Facebook. This period also coincided with the additional
Nash et al.
2www.jhypertension.com Volume 35 Number 1 Month 2017
Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

CE: Swati; JH-D-17-00388; Total nos of Pages: 5;
JH-D-17-00388
source of recruitment at the regional general practise
in QLD. The average monthly ratio of the number of
people recruited per enquiries made to study staff was
not significantly different from preinitiation to postinitia-
tion of Facebook at any site (ACT 0.20 vs. 0.31; P¼0.09,
TAS 0.36 vs. 0.26; P¼0.13, QLD 0.44 vs. 0.37; P¼0.38).
Compared with the first 20 months using conventional
recruitment methods, the average age of participants
contacting coordinators at all sites was significantly re-
duced after Facebook advertising (from 60.9 9.4 to
58.7 8.8 years; P<0.001).
DISCUSSION
To our knowledge, this study is the first to report on the
effect of Facebook advertising to engage and enhance
participant recruitment in a BP randomized controlled trial
of middle-to-older aged people. This was a retrospective
analysis from within an existing trial. The main findings
were that Facebook advertising was successful at increasing
the number of enquiries and participants recruited at two of
three national clinical trial study sites. It is also of note that
Facebook was successful in recruiting a cohort of particip-
ants in an older age bracket typical of cardiovascular-
0
5
10
15
20
25
30
35
40
TAS ACT QLD
Average number of enquiries /month
Convenonal adversements Convenonal + Facebook adversements
*
*
p= 0.16
FIGURE 2 Comparison of the average (SEM) number of enquiries from potential
study participants in the 4-month period pre (black bars) and post (open bars)
Facebook advertising in Tasmania, Australian Capital Territory and Queensland
study sites.

P<0.05.
TABLE 1. Statistics from the 4-month period where Facebook
advertisements were available intermittently at
clinical trial sites in Tasmania, Australian Capital
Territory and Queensland
Variable TAS ACT QLD
Days advertisements run, n80 55 49
Potential viewers, n49 912 23 781 154 477
Clicks onto advertisement, n3436 2521 1970
Enquiries to study staff, n114 88 44
Randomized participants, n30 26 11
Average cost per click (AUD) 0.80 0.43 0.98
Total cost of all advertisements (AUD) 2527 1174 1936
Average cost/enquiry to study staff (AUD) 22.17 13.34 44.00
Average cost/participant randomized (AUD) 84.23 45.15 176.00
ACT, Australian Capital Territory; AUD, Australian dollars; QLD, Queensland; TAS,
Tasmania.
FIGURE 1 Examples of Facebook advertisements placed on right-side panel to the Facebook newsfeed (a) and advertisement placed directly into newsfeed from the
University of Tasmania Facebook page (b).
Facebook and trial recruitment
Journal of Hypertension www.jhypertension.com 3
Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

CE: Swati; JH-D-17-00388; Total nos of Pages: 5;
JH-D-17-00388
related trials, although the average age was lower than
participants recruited prior to Facebook advertising. These
observations suggest that Facebook may be a useful tool to
enhance clinical trial recruitment.
Initiation of Facebook advertising was associated with
improved recruitment at two study sites (TAS and ACT), but
not in QLD, where there was a nonstatistically significant
trend for a reduction in the number of enquiries to study
staff as well as the number of participants recruited.
Although we cannot definitively identify the reasons for
this discrepancy, there are two factors that are likely to have
been contributory. First, a new source of recruitment was
opened at the QLD site in the months prior to Facebook
initiation, and this resulted in a boost in both enquiries and
number of participants recruited. Thus, the baseline level
for these metrics was higher at QLD compared with the
other study sites (Figs. 2 and 3). Second, the QLD site was a
larger metropolitan centre with a greater population than
the other study sites, and location differences could have
played a role. Indeed, recruitment rates vary significantly
between regions and medical practises, with lower rates
reported in a large capital city (London) compared with a
smaller city (Edinburgh) [11]. A key barrier for participation
in large cities is greater burden from long journey times [12],
and our study design involved many visits for participants.
Thus, despite QLD having a much larger population com-
pared with the other sites, it may be that Facebook adver-
tising cannot overcome difficulties of recruiting participants
in large capital cities.
The average age of study participants recruited after
initiation of Facebook advertising was statistically younger
(albeit only 2 years lower on average) than participants
recruited via conventional advertising. Although this
result was among a population of middle-to-older aged
participants, similar observations have been reported in
the recruitment of younger people into a clinical trial of
smoking cessation, whereby online-recruited participants
were aged 38.8 12.1 years compared with 45.6 13.5
years for traditional advertising. However, there were no
other differences in terms of socio-economic variables [13].
According to 2016 statistics, irrespective of age, Facebook
was the most popular social networking site, and older
people generally had the same reasons as young people for
using social networking (e.g. catching up with family and
friends, sharing photographs, getting information on news
and current events) [5]. Significantly, over the previous year
from 2016, the time spent on social networking had in-
creased to about the same level among people aged
65 years or older (23%) compared with those aged 18–
29 (22%), 30–39 (27%) and 40–49 years (26%) [6]. Social
media platforms are increasingly used for medical market-
ingtoprovideinformationtoendusersonhealthand
disease, with pharmaceutical companies, hospitals, medi-
cal research institutes and professional societies (e.g. dia-
betes education) participating in this arena [5]. Altogether
these data appear to indicate an increase in the opportu-
nities available to recruit older participants (in good
health or with existing disease) into clinical studies
using Facebook.
Limitations
The study presents findings from a retrospective observa-
tional analysis within a randomized controlled trial rather
than being specifically designed to determine the effects of
Facebook advertising on recruitment for a BP-related trial.
Thus, we cannot make definitive conclusions on the effects
of Facebook advertising, and more rigorous studies are
needed to determine if the concept is more generalizable. A
complete cost breakdown for the conventional recruitment
approach was not available due to the retrospective nature
of the study, and as such we cannot compare the cost
differences with Facebook advertising. This would also
need to be determined in future studies.
In conclusion, the current study found that using Face-
book advertising was effective for increasing recruitment of
older aged participants into a national BP-related study.
Effectiveness of recruitment using Facebook appeared to
be location dependent, with greater success among the
smaller, regional centres. Overall, Facebook seems to
reach a proportion of the population potentially interested
in clinical trial participation that is above and beyond
the attainable using conventional advertising methods
alone.
ACKNOWLEDGEMENTS
We are very grateful to Raylene Steinhardt (Practice Man-
ager, Limestone Medical Centre) for help in recruiting
participants for this study.
J.E.S. was supported by a National Health and Medical
Research Council (NHMRC) Career Development Award
(reference 569519) and the study was supported by an
NHMRC Project Grant (reference 1044551).
Conflicts of interest
There are no conflicts of interest.
REFERENCES
1. National Statement on Ethical Conduct in Human Research 2007
(Updated December 2013). Canberra, ACT: National Health and Medi-
cal Research Council, the Australian Research Council and the Austra-
lian Vice-Chancellors’ Committee; 2015.
2. Commonwealth Department of Health and Aged Care. Note for guid-
ance on good clinical practice (CPMP/ICH/135/95), annotated with
TGA comments. Melbourne: Commonwealth Department of Health
and Aged Care; 2000.
0
2
4
6
8
10
12
14
TAS ACT QLD
Average number of parcipants recruited/month
Convenonal adversements Convenonal + Facebook adversements
*
p=0.052
p=0.23
FIGURE 3 Comparison of the average (SEM) number of participants recruited
into the study in the 4-month period pre (black bars) and post (open bars) Face-
book advertising in Tasmania, Australian Capital Territory and Queensland study
sites.

P<0.05.
Nash et al.
4www.jhypertension.com Volume 35 Number 1 Month 2017
Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.

CE: Swati; JH-D-17-00388; Total nos of Pages: 5;
JH-D-17-00388
3. Tufts Center for the Study of Drug Development. 89% of trials meet
enrollment, but timelines slip, half of sites under-enroll. Impact report
volume 15. Jan/Feb 2013.
4. Austin A, Barnard J, Hutcheon N. Media consumption forecasts 2015.
South Brisbane: ZenithOptimedia; 2015; Available from: http://zenith
media.se/wp-content/uploads/2015/05/Media%20Consumption%
20Forecasts%202015.pdf. [Accessed 16 March 2017].
5. Zephoria Digital Marketing. The top 20 valuable Facebook statistics –
updated May 2017. Zephoria Digital Marketing; 2017; Available from:
https://zephoria.com/top-15-valuable-facebook-statistics/. [Accessed
2017 16 May 2017].
6. Sensis. Sensis social media report 2016. How Australian people and
businesses are using social media. Sensis Pty Ltd; 2016; Available from:
https://www.sensis.com.au/asset/PDFdirectory/Sensis_Social_Media_
Report_2016.PDF. [Accessed 17 March 2017].
7. Ramo DE, Prochaska JJ. Broad reach and targeted recruitment using
Facebook for an online survey of young adult substance use. J Med
Internet Res 2012; 14:e28.
8. Jones L, Saksvig BI, Grieser M, Young DR. Recruiting adolescent girls
into a follow-up study: benefits of using a social networking website.
Contemp Clin Trials 2012; 33:268–272.
9. Ramo DE, Rodriguez TM, Chavez K, Sommer MJ, Prochaska JJ. Face-
book recruitment of young adult smokers for a cessation trial: methods,
metrics, and lessons learned. Internet Interv 2014; 1:58–64.
10. Raviotta JM, Nowalk MP, Lin CJ, Huang HH, Zimmerman RK. Using
Facebook to recruit college-age men for a human papillomavirus
vaccine trial. Am J Mens Health 2016; 10:110– 119.
11. Gilbert H, Leurent B, Sutton S, Morris R, Alexis-Garsee C, Nazareth I.
Factors predicting recruitment to a UK wide primary care smoking
cessation study (the ESCAPE trial). Fam Pract 2012; 29:110–117.
12. Newington L, Metcalfe A. Factors influencing recruitment to research:
qualitative study of the experiences and perceptions of research teams.
BMC Med Res Methodol 2014; 14:10.
13. Frandsen M, Walters J, Ferguson SG. Exploring the viability of using
online social media advertising as a recruitment method for smoking
cessation clinical trials. Nicotine Tob Res 2014; 16:247–251.
Reviewer’s Summary Evaluation
Reviewer 1
Participant recruitment is often the primary reason
for slower than optimal progress in the conduct of
clinical trials. This study suggests that Facebook adver-
tising could be a useful approach to recruiting parti-
cipants. More efficient recruitment would save time
and money.
Facebook and trial recruitment
Journal of Hypertension www.jhypertension.com 5
Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.