Efficacy of Interventions Promoting Blood Donation:
A Systematic Review
Gaston Godin, Lydi-Anne Vézina-Im, Ariane Bélanger-Gravel, and Steve Amireault
Findings about the efficacy of interventions promoting
blood donation are scattered and sometime inconsistent. The aim of the present systematic review
was to identify the most effective types of interventions and modes of delivery to increase blood
donation. The following databases were investigated:
MEDLINE/PubMed, PsycINFO, CINAHL, EMBASE, and
Proquest Dissertations and Theses. Additional studies
were also included by checking the references of the
articles included in the review and by looking at our
personal collection. The outcomes of interest were
either blood drive attendance or blood donations. A
total of 29 randomized controlled trials or quasiexperimental studies were included in the review,
detailing 36 interventions tested among independent
samples. Interventions targeting psychosocial cognitions (s = 8, s to represent the number of
independent samples; odds ratio [OR], 2.47; 95%
confidence interval [CI], 1.42-4.28), those stressing
the altruistic motives to give blood (s = 4; OR, 3.89;
95% CI, 1.03-14.76), and reminders (s = 7; OR,
1.91; 95% CI, 1.22-2.99) were the most successful
in increasing blood donation. The results suggest that
motivational interventions and reminders are the most
effective in increasing blood donation, but additional
studies are needed to evaluate the efficacy of other
types of interventions.
© 2012 Elsevier Inc. All rights reserved.
LOOD AGENCIES ARE investing much
effort to recruit new blood donors and to
ensure that those who have donated return for
additional donations. This preoccupation is wellfounded given that, according to the American Red
Cross, only 3% of the population gives blood [1].
To further complicate matters, blood donation is a
cyclic process in which once an individual has
donated blood, he/she must wait a certain time (eg,
56 days) before being eligible to give blood again
[2]. Moreover, many medical and nonmedical
reasons such as foreign traveling can prevent
someone from successfully giving blood [2]. This
is done to ensure donor and recipient safety [3].
Effective interventions to retain blood donors and
recruit new donors are thus imperative to maintain a
safe and sufficient level of blood supply.
In a previous review on the motives and the
characteristics of blood donors, Piliavin [4] made
an inventory of potential successful approaches to
increase blood donation, particularly among first-
time donors. For instance, this author suggested that
an increase in the perception for the need of blood
in the community and in the perception of the
adequate support for blood donation provided in
these communities as well as the use of modeling
were promising approaches to increase blood
donation. The author also observed that face-toface solicitations, the use of personal reminders,
and the foot-in-the-door technique (consisting of a
small request followed by a bigger one) appeared to
be effective techniques to motivate individuals
to give blood, given that individuals are sensitive to
social pressure. These observations were also
supported in a more recent review by Ferguson
et al [5], in which the use of reminders and the footin-the-door technique were mentioned as effective
strategies to favor blood donation.
Mixed evidences were observed concerning the
efficacy of interventions aimed at increasing
altruistic motives vs the use of different forms of
incentives [4]. However, as pointed out earlier by
Piliavin [4], the use of incentives might undermine
the development of an altruistic motivation and
provoke a backfire effect on future donations
among active blood donors. Moreover, in a recent
review of the literature among active blood donors,
Ringwald et al [6] mentioned that the use of
incentives might not be the best strategy to keep
long-term and committed blood donors given that
as a donor advances in his/her career, his/her
motives for giving blood become more internalized.
The role of altruism also needs some consideration.
According to a deeper analysis by Ferguson et al [5]
B
From the Canada Research Chair on Behaviour and Health,
Laval University, Québec, Canada; Faculty of Nursing, Laval
University, Québec, Canada; and Department of Social and
Preventive Medicine, Laval University, Québec, Canada.
The authors declare that they have no conflict of interest.
Address reprint requests to Gaston Godin, PhD, Canada
Research Chair on Behaviour and Health, Laval University,
Québec, Canada G1V 0A6.
E-mail: gaston.godin@fsi.ulaval.ca
0887-7963/$ - see front matter
© 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.tmrv.2011.10.001
224
Transfusion Medicine Reviews, Vol 26, No 3 (July), 2012: pp 224-237.e6
INTERVENTIONS PROMOTING BLOOD DONATION
of previous studies aimed at promoting blood
donation with the use of the altruism concept,
results seem to provide more support for the use of
the moral norm concept or the benevolence
response rather than pure altruism. Consequently,
the role of altruism in the promotion of blood
donation has to be more clearly defined to further
increase the efficacy of such interventions.
Bednall and Bove [7] also did a review of selfreported motivators and deterrents of blood donation. In particular, they verified that the motives
behind blood donation differed among different
types of donors (first-time, repeat, lapsed, apheresis, and eligible nondonors). They concluded that,
among first-time and repeat donors, the most
common motivators were convenience (eg, blood
drive nearby), prosocial motivation (eg, altruism),
and personal values (eg, moral norm). They also
identified several deterrents among which the most
frequently mentioned barrier was low self-efficacy
to donate (eg, ability to overcome barriers such as
lack of time).
Notwithstanding the very useful information
provided by these previous reviews concerning
the potential efficacy of different strategies to
increase blood donation, none of them systematically reviewed the literature and provided information on the effect sizes of such approaches. In
addition, in the review by Ferguson et al [5],
studies whose outcome was the intention or
willingness to give blood were also included in
their narrative analysis. Acknowledging the scientific evidence that intention is identified as one of
the most important determinants of blood donation
[2], it is, nonetheless, well documented that not all
people will act on their stated intention, that is there
is an “intention-behavior gap” [8]. Moreover, in a
review by Webb and Sheeran [9] on the effectiveness of interventions to promote health-related
behaviors, it was observed that a medium-to-large
effect size in increasing levels of intention leads to
a small-to-medium effect in changing behaviors.
Consequently, the efficacy of interventions on
blood donation and not only intention or willingness has to be determined.
To our knowledge, there is no systematic review
on the efficacy of interventions to promote blood
donation. Thus, the objective of the present
systematic review was to identify the most effective
types of interventions and their modes of delivery to
increase blood donation.
225
METHODS
Study Eligibility Criteria
The focus of the present systematic review was
on interventions aimed at increasing blood donation
among the general population (ie, recruitment
studies) and blood donors (ie, retention studies).
Studies aimed at increasing different or specific
types of donations such as bone marrow donation or
apheresis were not included in the present review.
To evaluate the efficacy of interventions, only
experimental and quasi-experimental studies were
included in the review; studies adopting a 1-group
pre-post design were excluded.
Studies concerned with different stages in the
lifetime experience of blood donors, that is, novice,
experienced, lapsed, or temporarily deferred donors,
were included. According to Masser [10], individuals who already gave blood once were considered
as novice donors (also referred to as first-time
donors), and individuals who already gave blood
more than once were considered as experienced
donors (also referred to as repeat donors) 1. Studies
interested in lapsed donors (ie, individuals who
already gave blood in the past but who did not
donate blood in the past 2 years without any known
cause for permanent or temporary deferral [11]) or
temporarily deferred donors (ie, individuals who
attended a blood center but were not eligible to
donate blood for a certain period because of various
medical reasons [12]) were also considered in the
present systematic review.
Studies reporting results on the attendance at a
blood drive or the number of successful blood
donations after the exposure to an intervention were
both included in the present systematic review.
Attendance was defined as the number of participants
who registered at a blood drive. This outcome was
chosen given that although some individuals can be
deferred for various medical reasons, attendance
takes into account the fact that participants had acted
toward their goal of giving blood after the exposure
to the intervention [13,14]. To avoid duplication of
the results, attendance was primarily used in the
statistical analyses in cases where attendance and the
number of actual blood donations were both provided
in a given study. Studies reporting the results of an
1
Other classifications can be found in the scientific literature
(eg, Ferguson and Bibby (2002): occasional, b5 donations;
regular, ≥5 donations).
226
GODIN ET AL
intervention on the level of intention or willingness as
well as studies reporting results in terms of
physiologic reaction, stress, and symptoms related
to blood donation were excluded from the present
systematic review.
Search Strategy
The following databases were investigated:
MEDLINE/PubMed (1950+), PsycINFO (1806+),
CINAHL (1982+), and EMBASE (1974+). Proquest Dissertations and Theses (1861+) was also
investigated for gray literature (ie, unpublished
trials). No restriction was placed on the year of
publication of the articles. The search was
performed between June 8 and July 31, 2010. In
all the databases, the search terms were always
related to 2 themes, that is, blood donation and
intervention (see Table 1). In MEDLINE/PubMed,
a combination of keywords and MeSH terms was
used. In PsycINFO, only keywords were used
because no psychological index terms corresponded to blood donation and intervention. In
CINAHL, a combination of keywords and descriptors was used. In EMBASE, a combination of
keywords and Emtree terms was used. Finally, in
Proquest Dissertations and Theses, only keywords
were used. The search was limited to studies
published in English. Additional studies were also
included by checking the references of the articles
included in the systematic review (ie, secondary
references) as well as by looking at our personal
collection of articles on blood donation.
Data Extraction
Data were independently extracted by GG and
LAVI using a standardized data extraction form (see
Table 2). Disagreements were resolved by discussion. Before extracting the data, several decisions
were made. First, in instances where there was more
than 1 experimental group in a study, only the group
that had the greatest impact on blood donation was
Table 1. Search Terms Used for Investigating All 5 Databases
Keywords
Theme 1: blood donation
Theme 2: intervention
Blood donation OR
Blood donor OR
Blood bank
Intervention OR
Program OR
Education
considered and compared with the control group.
This decision was taken given that the main objective
of the present review was to verify which types of
intervention were more successful in increasing
blood donation and also because this group corresponded to the hypothesis stated in the articles
included in the meta-analysis in 90% of the cases.
Second, when several tests of an intervention were
performed among independent samples but reported
in 1 article, these were all included in the analysis (ie,
studies 1 and 2). Consequently, the number of
independent comparisons in the meta-analysis might
be more important than the number of included
studies in some of the subgroup analyses. Finally,
authors were personally contacted by e-mail for
additional information when necessary.
Interventions were classified according to an
adapted version of the categorization proposed in a
recent review by Ferguson et al [5] (see Table 3).
These authors regrouped the social interventions
into 4 types: (1) altruism-egoism, (2) reminders and
commitments, (3) foot-in-the-door, and (4) intention based. Social interventions were aimed at
increasing blood donation and were mainly focused
on recruitment. They also identified behavioral
interventions that can be classified in 3 categories:
(5) distraction, (6) muscle tension, and (7) caffeine
and water loading. These behavioral interventions
are primarily aimed at reducing unpleasant symptoms associated with blood donation and mainly
focused on retention of donors.
The first category of social interventions
included interventions that use manipulations of
altruism and egoism to encourage blood donation.
In the present review, however, the interventions
based on altruism were imbedded in a larger
motivational category that included all the interventions aimed at increasing motivation toward
blood donation. These studies used techniques
such as providing information about positive or
negative consequences of blood donation, barriers
management, modeling, comparison with the
social norm, providing feedback and praise, and
other. However, given the importance of the
concept of altruism in the literature on blood
donation, the independent efficacy of interventions based on altruism was compared with the
efficacy of other motivational interventions in the
subgroup analyses. The second category included
interventions generally focusing on retention that
use a reminder, such as a telephone call, before
INTERVENTIONS PROMOTING BLOOD DONATION
227
Table 2. Information Contained in the Data Extraction Form
Type of information
Subcategory
Main objective of the study
Recruit new donors
Encourage donors to return
Both objectives
Unspecified
Potential donors/general population
Novice donors
Experienced donors
Unspecified
Randomized, controlled trial
Quasi-experimental design
Type of assignment
Strategy used for assignment
Technique used for assignment
Person who performed assignment
Type of control group
Motivational
Reminders
Measurement of cognitions
Incentives
Muscle tension
Face-to-face
By mail (eg, letter)
By telephone (eg, telephone call prompt)
By e-mail
By tape recorder
Mix
Yes
No
Registrations at a blood drive
Blood donations
Both
Type of behavioral measure
Validation of the measure
Size
Attrition at postintervention
Attrition at follow-up
Characteristics of participants at baseline
Equivalence of groups at baseline
Dropout analysis
Statistical test used
Intention to treat
Covariates
Technique to replace missing data
Percentage of participants who registered at a blood drive
Percentage of participants who gave blood
Other
Population of the study
Study design
Group allocation
Type of intervention
Mode of delivery
Use of a theory
Behavior measured
Behavioral measure
Sample
Statistical analyses
Results of the intervention
the blood drive takes place. The third category
was the foot-in-the-door [15] or door-in-the-face
[16] interventions that are well-known techniques
issued from the social psychology literature. The
last category, that is, the intention-based category,
initially included interventions that manipulated
intentions. Given that the intention-based category
was too specific to the measurement of intention,
a new category was created (ie, measurement of
cognitions or the question-behavior effect) to
include all studies that measured cognitions with
the specific objective of promoting blood donation. Finally, the use of incentives (financial or
not) was added as a separate category, and the
behavioral categories were classified using the
description provided by Ferguson et al [5]. All of
228
GODIN ET AL
Table 3. Classification of the Interventions Based on the Categorization of Ferguson et al [5]
Type of intervention
Motivational
Cognitions based
Foot-in-the-door/door-in-the-face
Altruism
Modeling
Reminders
Measurement of cognitions
Incentives
Muscle tension
Definition
Interventions aimed at increasing motivation toward blood donation
Interventions targeting psychosocial cognitions related to motivation,
such as social norms, attitudes, and barriers
Interventions using the foot-in-the-door, the door-in-the-face,
or a combination of both techniques to motivate people to give blood
Foot-in-the-door involves asking a small request that should be accepted
and then asking a critical large request. Door-in-the-face involves
asking a large request that should be refused
and then asking a critical small request.
Interventions using altruistic motives to motivate people to give blood
Interventions showing another person giving blood to motivate
people to give blood
Interventions using reminders about the next eligibility date and/or the next
appointment to give blood (eg, telephone call prompt)
Interventions using the completion of a questionnaire about the intention to give
blood to activate cognitions about blood donation (eg, question-behavior effect)
Interventions using incentives for donating blood such as a T-shirt, money,
prizes, tickets, and other
Interventions using the contraction of muscles during blood donation to avoid
dizziness and fainting
the interventions were independently classified
according to their type by LAVI and ABG,
and disagreements were resolved by a third
reviewer (GG).
Data Analyses
Descriptive statistics of the studies, such as
frequencies and means, was analyzed using SAS
version 9.1 (SAS Institute, Cary, NC). All the other
statistical analyses were done using Cochrane's
Review Manager 5.1 (The Nordic Cochrane Center,
Copenhagen, Denmark) [17] statistical software. To
compute pooled effect sizes, a random-effect model
using the Mantel-Haenszel method was chosen for
all the analyses because we assumed that the
magnitude of the effect sizes would vary across
studies given the difference in samples and interventions across studies [18]. When computing the
pooled effect size, each study was weighted
according to its sample size. Between-study heterogeneity was verified by using 2 common statistical
approaches: a χ 2 test (Cochran's Q) and the I 2
statistic [19], representing the percentage of total
variation in estimated effects that is due to
heterogeneity rather than chance. A significant Q
statistic (P b .05) indicates significant heterogeneity
between the studies, whereas an I 2 statistic of 25% is
considered low heterogeneity; 50%, moderate
heterogeneity; and 75%, high heterogeneity [20].
To test the robustness of the results, separate 80%
prediction intervals (PIs) were calculated for each
pooled effect size (see Supplementary file 1 online
for the formula used). The computation of PIs is
recommended when confronted with a randomeffect meta-analysis that has a low number of studies
and high heterogeneity [21]. Prediction intervals
show the distribution of effect sizes around the
pooled effect size. The effect sizes of intervention
studies were reported as odds ratios (ORs) given that
most of the outcomes were reported as dichotomous
data (ie, the proportion of attendance or blood
donation in the experimental group compared with
the control group). Unless otherwise stated, all effect
sizes are zero order (ie, no covariates are included in
the computation of the effect size). The ORs were
also converted to Cohen's d [22] to facilitate
interpretation and allow comparison with standard
effect sizes reported in other studies. A Cohen's d of
0.20 is considered a small effect size; 0.50, a
medium effect size; and 0.80, a large effect size [22].
The efficacy of the interventions according to type
of intervention and mode of delivery was also
examined by means of prespecified subgroup
analyses. However, pooled effect sizes were computed only when the category contained results from
a minimum of 3 independent samples. Finally,
publication bias was assessed by visually inspecting
the distribution of the funnel plot in RevMan 5.1
(The Nordic Cochrane Center) in instances where
there were about 10 studies per category [23].
INTERVENTIONS PROMOTING BLOOD DONATION
RESULTS
The results of the search strategy are presented
in Figure 1 [24]. A total of 29 studies (k = 29)
were included in the present review, detailing 36
different interventions tested among independent
samples (s = 36). In the rest of the text, the letter
k will be used to represent the number of studies,
and the letter s to represent the number of
independent samples.
Characteristics of the Interventions
A summary of the interventions is given in
Supplementary file 2. When no effect sizes could be
229
calculated because of missing information such as
the number of participants per condition, the original results are reported. Most interventions were
aimed at retaining current donors (s = 12) or both
recruiting new donors and retaining donors (s = 8).
One intervention had the objective to recruit new
donors. In 15 interventions, the objective did not
clearly specify if it was more about recruiting or
retaining blood donors or vice versa. Again, few
studies clearly specified the type of donors targeted
by the intervention; only 6 studies mentioned this
information (novice donors, s = 5; experienced
donors, s = 1). Given that more than 3 studies were
targeting novice donors, a pooled effect size was
Fig 1. Preferred reporting for systematic reviews and meta-analyses flowchart [24].
230
GODIN ET AL
calculated (s = 5; OR, 1.21; 95% confidence
interval [CI], 1.10-1.32; Q = 1.68; P = .80; I 2 =
0%); it represented a small effect size (d = 0.11). No
significant heterogeneity between the studies was
detected. Unfortunately, given the low number of
studies, no pooled effect sizes could be calculated
for other types of donors, such as experienced,
lapsed, and deferred donors.
Of the 36 interventions included in the review,
24 were randomized controlled trials, although
most of them (s = 16) did not provide information
on the randomization procedure. The other 12
interventions either adopted a quasi-experimental
design (s = 3) or did not provide information
about the design adopted (s = 9). Although it
would have been interesting to compare the
difference in effect sizes between both pure and
quasi-experimental studies, 2 problems arose: (1)
many studies did not specify their study design
(s = 9) and (2) older studies tended to have little
information about their randomization procedure
compared with more recent studies; this phenomenon could reflect different editorial requirements
through the years.
Twenty-one interventions were clearly based
on a theory. The theories most frequently used to
design interventions promoting blood donations
were the self-perception theory [25] (s = 4), the
social learning/social cognitive theory [26,27]
(s = 4), the theory of reasoned action [28] and
its extension, the theory of planned behavior [29]
(s = 4), and Schwartz's norm activation theory
[30] (s = 2).
Characteristics of the Participants
Only 12 studies provided information on the
age of their participants. The mean age of the
participants was 25.72 ± 8.61 years (range,
18-44.3 years). Eighteen interventions were
conducted among samples of college students,
and 3 others, among samples of high school
students. Seventeen interventions indicated the
percentage of male participants in their sample.
A little less than half (44.8%) of the samples
were composed of male respondents. Three
interventions mentioned having samples containing participants of both sexes without specifying
the percentage of male respondents. One intervention had a sample exclusively of female
participants. The typical sample was thus composed of college students with slightly more than
half being female students.
Efficacy of the Interventions
A summary of the pooled effect sizes computed,
their heterogeneity, and their robustness is presented in Table 4. Where exact information about
numbers of participants or effect sizes were
missing [31-35], we attempted to obtain the
relevant information directly from the authors. In
3 cases [33-35], we were unable to obtain this
information. Thus, of the 29 studies describing 36
Table 4. Pooled Effect Sizes, OR with 95% CIs
Variable
Type of intervention
Motivational
Cognitions based
Foot-in-the door/door-in-the-face
Altruism
Reminders
Measurement of cognitions
Mode of delivery
Face-to-face
Mail
Telephone
Mix of modes
s
n
OR (95% CI)
20
8
6
4
7
3
37 102
23 198
1303
7619
286 615
6812
2.09 (1.64-2.67)
2.47 (1.42-4.28)
1.86 (1.02-3.39)
3.89 (1.03-14.76)
1.91 (1.22-2.99)
1.23 (1.06-1.41)
5
4
10
10
239
5536
287 882
15 896
2.57
1.25
2.41
1.60
(1.29-5.11)
(1.12-1.40)
(1.38-4.20)
(1.27-2.00)
Q
I2
(80% PI)
117.21 ‡
42.18 ‡
12.72 ⁎
51.43 ‡
19.71 †
3.01
84%
83%
61%
94%
70%
34%
(1.23-3.53)
(0.88-6.97)
(0.74-4.69)
(0.30-50.96)
(0.91-4.01)
(0.85-1.79)
5.14
2.49
84.50 ‡
40.23 ‡
22%
0%
89%
78%
(1.11-5.95)
(0.89-1.76)
(0.77-7.50)
(1.07-2.39)
NOTE. I2 is an indicator of heterogeneity. Twenty-five percent is considered low heterogeneity; 50%, moderate heterogeneity; and 75%,
high heterogeneity.
⁎ P b .05, significant heterogeneity (Q statistic).
†
P b .01, significant heterogeneity (Q statistic).
‡
P b .001, significant heterogeneity (Q statistic).
INTERVENTIONS PROMOTING BLOOD DONATION
interventions, 26 studies describing 33 interventions among independent samples could be
included in the meta-analysis.
Motivational Interventions
Twenty-two interventions [32,33,36-50] were
classified in the motivational category. However,
2 interventions [33,48] did not report the number of
participants in their samples and were thus excluded
from the meta-analysis. The study of Evans [33]
reported nonsignificant results (P = .67), whereas
the study of Sarason et al [48] had significant results
(P b .01). The pooled effect size (s = 20) indicated
that motivational interventions were successful in
increasing blood drive attendance, and this represented a small-to-medium effect size (d = 0.41).
Significant heterogeneity between the studies was
detected, which justified subgroup analyses. The
80% PI indicated, however, that the significant
result was robust.
Cognitions-Based Interventions
Nine interventions [32,33,36,38,41,43,44] were
classified as cognitions based because they were
targeting psychosocial cognitions related to motivation, such as social norms, attitudes, and barriers.
However, 1 intervention [33] could not be entered
in the meta-analysis because it did provide the
number of participants per condition (ie, in the
experimental and control groups). This study
reported nonsignificant results (P = .67). The
pooled effect size (s = 8) indicated that cognitions-based interventions were successful in increasing attendance at blood drives, and this
represented a medium effect size (d = 0.50).
Significant heterogeneity was detected, and examination of the 80% PI indicated that the result
should be interpreted carefully. Moreover, the
examination of the funnel plot (see Supplementary
file 3 online) indicated that there is probably a
publication bias [23].
Foot-in-the-Door and Door-in-the-Face
Techniques
A total of 6 interventions [37,40,42,46] were using
the foot-in-the-door, the door-in-the-face, or a
combination of both techniques to encourage people
to give blood. The pooled effect size (s = 6) indicated
that these types of interventions were successful at
increasing attendance at blood drives. When the
pooled OR was converted to Cohen's d [22], it
231
represented a small-to-medium effect size (d = 0.34).
The Q statistic indicated that there was significant
heterogeneity between the studies, although the I 2
statistic was still below 75%. The 80% PI indicated
that the result is not very robust.
Altruism
Four interventions [39,45,50] aimed to increase
blood donation by stressing the altruistic reasons to
give blood. The pooled effect size (s = 4) revealed
that this kind of intervention was successful in
increasing blood drive attendance; this represented
a medium-to-large effect size (d = 0.75). Significant
heterogeneity between the studies was detected,
and the 80% PI was very large, which indicates that
the result is not robust at all.
Modeling
Three interventions [47-49] were using modeling, such as showing a model giving blood, to
increase blood donations. However, 1 study [48]
did not report the exact number of participants in
their experimental and control groups. This study
by Sarason et al [48] reported that their intervention had significant results (P b .01). Given that
only 2 studies could have been entered in a metaanalysis, no pooled effect size was computed for
this type of intervention. Examination of the
individual effect sizes of the studies indicated that
all the studies reported significant results (Rushton
and Campbell [47]: OR, 39.67; 95% CI, 1.281229.87; and Sarason et al [49]: OR, 1.25; 95% CI,
1.06-1.47). Additional studies will be needed
before any effect size can be computed about the
efficacy of interventions that use modeling to
increase blood donation.
Reminders
Seven interventions [51-55] were using reminders, such as telephone call prompts, as a
means to increase blood donations. The pooled
effect size (s = 7) for this type of intervention
indicated that they were successful in increasing
attendance at blood drives, and this represented a
small-to-medium effect size (d = 0.36). Significant
heterogeneity between the studies was detected,
and the 80% PI was close to significance, which
indicates that the result is somewhat robust. The
inspection of the funnel plot (see Supplementary
file 4 online) indicated the likely presence of a
publication bias [23].
232
GODIN ET AL
Measurement of Cognitions
Modes of Delivery
Three interventions [13,56,57] used answering
questions about cognitions related to blood donation to increase this behavior. One study [57]
reported a continuous outcome (ie, mean number of
blood donations), which allowed the calculation of
a standard mean difference (SMD). This SMD was
then converted to an OR to pool it with the rest of
the studies (see Supplementary file 5 online for the
formula used). The pooled effect size (s = 3)
indicated that interventions using the measurement
of cognitions are effective in increasing blood
donation and that no significant heterogeneity was
detected. Once converted to Cohen's d, the pooled
effect size represented a small effect size (d = 0.11).
The 80% PI, however, indicated that the result
should be interpreted carefully.
The efficacy of interventions according to their
mode of delivery was also evaluated. In 1
intervention [41], a brochure was used to contact
participants, but it was not specified if it was sent by
mail or if it was given in person, thereby making it
impossible to classify its mode of delivery.
Incentives
Two interventions [34,58] used incentives, such
as tickets to a Broadway play and a college football
game, gift certificates, prizes, free meals, and other,
to increase blood donation. One intervention [34]
did not report the exact number of study participants but stated that the intervention was successful
in increasing blood drive attendance (P b .01). The
other study by Ferrari et al [58] also had significant
results (OR, 3.86; 95% CI, 1.47-10.13). The low
number of studies using incentives and the fact that
only 1 study reported the information necessary to
compute an effect size precluded any meta-analysis.
Additional studies will be needed to verify the
accuracy of the present finding and before any
definite conclusion can be reached about the
efficacy of incentives to increase blood donation.
Muscle Tension
Two interventions [31,59] used muscle tension
to decrease negative symptoms associated to
blood donation such as dizziness and fainting to
increase blood donors return. Given the low
number of studies using this technique, no effect
size was computed. Examination of the individual
effect sizes indicated that both studies had
nonsignificant results (Ditto et al [59]: OR,
1.48; 95% CI, 0.95-2.30; and Ditto et al [31]:
OR, 1.16; 95% CI, 0.82-1.66). Additional studies
will be needed to verify the definite efficacy of
muscle tension on blood donation.
Face-to-Face
A total of 6 interventions [34,37,40,43,47,58]
were delivered face-to-face. However, 1 study [34]
did not report the exact number of participants in
their intervention but still presented significant
results (P b .01). The pooled effect size was thus
computed on 5 studies, and it indicated that
interventions delivered face-to-face are effective in
increasing attendance at blood drives. This
represented a medium effect size (d = 0.52). No
significant heterogeneity between the studies was
detected, and the 80% PI indicated that the result
is robust.
Mail
Four interventions [13,46,53] were delivered by
mail, such as the mailing of a brochure, a letter, or a
questionnaire. The pooled effect size (s = 4)
indicated that interventions using this mode of
delivery were effective in increasing blood drive
attendance, and this represented a small effect size
(d = 0.12). No heterogeneity between the studies
was detected, and the 80% PI was close to
significance, which indicates that the result is
somewhat robust.
Telephone
A total of 11 interventions [32,33,38,40,42,5052,54,55] were delivered by telephone, such as
telephone call prompts, making it the most common
mode of delivery. However, 1 study [33] did not
report the number of participants per condition and
was thus excluded from the meta-analysis. This
study by Evans [33] reported that its intervention
did not significantly increase blood donation in the
experimental condition (P = .67). The pooled effect
size (s = 10) for interventions delivered by
telephone indicated that this mode of delivery was
effective in increasing blood drive registration, and
this value was very close to a medium effect size
(d = 0.49). Significant heterogeneity was detected,
and the 80% PI indicated that the results should be
INTERVENTIONS PROMOTING BLOOD DONATION
interpreted carefully. In addition, the visual inspection of the funnel plot (see Supplementary file 6
online) indicated that there might be a publication
bias [23].
Tape Recorder
One study [39] that reported 2 interventions
carried out on 2 different samples mentioned using
a tape recorder to deliver the interventions. Given
the low number of studies, no pooled effect size was
computed. Examination of the individual effect
sizes indicated that this mode of delivery is
effective in increasing blood drive attendance
(Ferrari and Leippe [39], study 1: OR, 2.24; 95%
CI, 0.36-13.78; study 2: OR, 22.14; 95% CI, 2.58190.16). Additional studies are needed before a
definite conclusion about the efficacy of interventions delivered by tape recorder can be reached.
E-mail
Only 1 study [56] mentioned the use of an e-mail
to conduct its intervention. Examination of the
effect size indicated that this mode of delivery is
effective (OR, 2.44; 95% CI, 1.04-5.73) and
represents a small-to-medium effect size (d = 0.45).
Mix of Modes of Delivery
A total of 11 interventions [31,36,40,44,45,48,
49,57,59] were using a combination of different
modes of delivery, such as a letter delivered by
mail and a telephone call or instructions given
face-to-face and a video presentation. However, 1
study [48] did not provide the exact number of
participants in its experimental and control groups
and was thus excluded from the meta-analysis.
This study by Sarason et al [48] reported a
significant difference in the number of blood
drive registrations between the experimental and
the control groups (P b .01). The pooled effect
size (n = 10) indicated that interventions using a
mix of different modes of delivery were successful in increasing attendance at blood drives. This
represented a small-to-medium effect size (d =
0.26). Significant heterogeneity was detected, but
the 80% PI indicated that the result is robust.
DISCUSSION
According to our review, the first studies
reporting interventions to increase blood donation
were published in the 1970s. Therefore, it is rather
surprising that more than 4 decades later, the
233
number of papers available in the scientific literature
is low (n = 29). When we look at the flowchart (see
Fig 1), this phenomenon does not seem to originate
from the fact that many interventions did not report a
behavioral outcome (n = 25) or that studies adopted
a 1-group pre-post design (n = 3) but, rather, that the
literature in blood donation has mainly focused on
identifying the motives of blood donors by means of
surveys (n = 102).
Effective Interventions
Earlier reviews [5,60] also identified reminders,
such as telephone call prompts to inform donors of
the time, date, and location of blood drives, as an
effective means to increase blood donation. Moreover, the results of a recent study conducted among
first-time donors indicated that a telephone call
reminder about the upcoming opportunity to give
blood increases the number of donors who return
and that they also return more quickly compared
with those not telephoned [61]. Telephone calls act
as a “cue-to-action,” that is, they represent an
environmental cue that is used to remind people to
perform a given behavior [62].
In the present review, cognition-based interventions included all interventions targeting psychosocial constructs underlying motivation, such as
social norms, attitudes, and barriers (for a complete
taxonomy, see Michie et al [63]). The effectiveness
of this type of intervention at increasing blood
donations provides further support to previous
observations that intention, a construct expressing
ones' motivation toward a given behavior, is one of
the main determinants of blood donation [2,64].
This observation is also in agreement with the
theory of planned behavior [29], a model that
already has been successfully applied to predict
blood donation [64-66].
Interventions based on altruism seemed to be
effective in encouraging people to give blood.
However, this result should be interpreted carefully
given that quantification of the efficacy proved
difficult. Earlier reviews [4,60] already noted that
most donors cite altruistic motives as the main
reason to give blood. It was also suggested that
women and young people respond well to altruistic
appeals [60]. Thus, given that most interventions
were offered at samples of college students with
slightly more than half of these samples containing
female students, this might explain the efficacy of
this type of interventions. It remains to be
234
documented if appeals to altruism are effective
among older male samples.
In the present review, interventions using the
measurement of cognitions were also effective to
increase attendance at blood drives. The impact of
this type of intervention on behavior is generally
considered as the question-behavior effect or the
mere-measurement effect. According to the question-behavior effect, asking questions on relevant
cognitions can change behavior [67]. It was not
surprising that only 3 studies were identified in that
category given that the literature on the questionbehavior effect is still in its infancy; only recently
has such an effect been observed in blood donation
[13,57]. It will thus be important to obtain
additional information on the type of questions
that provoke taking action as well as the potential
moderators of this effect (eg, level of experience
with blood donation).
Interventions Whose Efficacy Could Not
Be Determined
Caution should be exercised before concluding
that offering incentives is an effective way to
encourage blood donation because only 2 studies
tested this approach, and no effect size could be
computed. It is also noteworthy that incentives were
sometime used in combination with other intervention components (eg, altruistic motives to give
blood and T-shirt). Other reviews [4,60] have
indicated that there was little evidence that the use
of incentives such as money, prizes, and similar
tokens were effective in increasing blood donation.
There is even evidence that they may actually
decrease donations by decreasing the altruistic
motives to give blood [5,60].
In the present review, the foot-in-the-door and the
door-in-the-face techniques emerged as effective
methods to increase blood donation, although the
result is not very robust. Piliavin [4] already
acknowledged that the use of these 2 techniques
was yielding to inconsistent results. Ferguson et al
[5] also noted that the foot-in-the-door technique
appeared a more effective approach than the doorin-the-face technique. In the present study, we did
not compute an effect size for each of the 2
techniques separately, given the small number of
studies using only 1 of the 2 techniques. Nonetheless, even if such techniques were found effective at
increasing blood donations, their use on a large scale
can be questioned for ethical or practical reasons.
GODIN ET AL
Effective Modes of Delivery
The present review also quantitatively evaluated
which modes of delivery are most effective at
increasing blood donations. In this regard, 3 modes
emerged as the most promising avenues to deliver
interventions on blood donation: face-to-face, by
telephone, and by mail. Previous reviews [4,60] had
highlighted that face-to-face solicitation was a
highly effective recruitment technique that could
be up to 4 times more effective than recruitment by
telephone. In the present meta-analysis, there was a
trend suggesting that the more a mode of delivery
was personal, the more effective it was in increasing
blood donation. However, given that face-to-face
solicitation is more time-consuming and more
expensive [60], interventions delivered by telephone might be the most cost-efficient method.
Finally, there is a lack of studies using new
technologies; only 1 study [56] used an e-mail to
carry out its intervention. Therefore, it will be
important in the future to document the potential of
using e-mail, SMS, and other new technologies as
novel approaches to promote blood donation,
especially among young blood donors.
Areas Where Additional Studies Are Needed
Unfortunately, very few interventions were
developed to target a specific type of donors (eg,
novice donors, regular donors, etc). This is rather
surprising given that previous studies have clearly
shown that the motives for giving blood differ
between types of donors. For example, it has been
shown that the determinants of return to give blood
again differ between novice and experienced donors
[64] and current and lapsed donors [11]. This would
suggest that different interventions should be
developed for first-time, novice, experienced, and
regular donors. Alas, the results of the present
review that indicated that interventions among
novice donors are effective could not be compared
with those of other types of donors. Moreover, most
interventions were carried out among samples of
college students, thereby making the results hard to
generalize to the general population.
The present review also revealed that some
authors omit important information when describing their intervention. Authors should be required to
refer to the Consolidated Standards for Reporting
Trials (CONSORT) statement [68]. The CONSORT statement comprises a 25-item checklist of
INTERVENTIONS PROMOTING BLOOD DONATION
235
information authors should report in their randomized controlled trials. The use of this checklist
should lead to more transparent reporting and
facilitate replication of the findings.
Another observation is that additional interventions based on modeling and muscle tension are
needed before a definite conclusion about their
efficacy can be reached. Modeling is a technique
grounded in the social learning theory of Bandura
[27]. It has already been successfully applied to
encourage children to eat more fruits and vegetables
[69] and to perform regular physical activity [70]. It
remains to be seen if this technique can be applied
in adult populations. In addition, behavioral interventions such as applied muscle tension are still a
new type of interventions to reduce discomfort (eg,
dizziness) associated with donating blood [59].
Although the results obtained were nonsignificant,
it could still be an interesting avenue for retaining
donors given that it can be implemented directly on
site during blood drives.
Limitations of the Systematic Review
The present review has some limitations that
are worth mentioning. First, the relatively small
number of studies prevented the computation of
some comparisons and affected the robustness of
the pooled effect sizes. Second, not all interventions could be included in the metaanalysis because they did not report the information needed to compute an effect size (ie, the
number of participants in the experimental and
the control groups).
CONCLUSIONS
Finally, to our knowledge, this is the first study to
systematically review blood donation interventions
and to quantify their effect. As such, the present
review contributes to improve current knowledge,
to identify gaps in knowledge, and to suggest new
directions for future interventions aimed at promoting blood donation.
REFERENCES
[1] American Red Cross. Give Blood. 2011 [cited 2011 January
12]; Available from: http://www.redcross.org/portal/site/en/
uitem.mend8aaecf214c576bf971e4cfe43181aa0/?vgnextoid=
d0061a53f1c37110VgnVCM1000003481a10aRCRD.
[2] Ferguson E. Predictors of future behaviour: a review of
the psychological literature on blood donation. Br J Health
Psychol 1996;1:287-308.
[3] Linden JV, Gregorio DI, Kalish RI. An estimate of blood
donor eligibility in the general population. Vox Sang
1988;54:96-100.
[4] Piliavin PA. Why do they give the gift of life? A review of
research on blood donors since 1977. Transfusion 1990;30:
444-59.
[5] Ferguson E, France CR, Abraham C, Ditto B, Sheeran P.
Improving blood donor recruitment and retention: integrating theoretical advances from social and behavioral science
research agendas. Transfusion 2007;47:1999-2010.
[6] Ringwald J, Zimmermann R, Eckstein R. Keys to open the
door for blood donors to return. Transfus Med Rev
2010;24:295-304.
[7] Bednall TC, Bove LL. Donating blood: a meta-analytic
review of self-reported motivators and deterrents. Transfus
Med Rev 2011;25:317-34.
[8] Sheeran P. Intention-behavior relations: a conceptual and
empirical review. Eur Rev Soc Psychol 2002;2:1-36.
[9] Webb TL, Sheeran P. Does changing behavioral intentions
engender behavior change? A meta-analysis of the
experimental evidence. Psychol Bull 2006;132:249-68.
[10] Masser BM, White KM, Hyde MK, Terry DJ. The
psychology of blood donation: current research and future
directions. Transfus Med Rev 2008;22:215-33.
[11] Germain M, Glynn SA, Schreiber GB, Gelinas S, King M,
Jones M, et al. Determinants of return behavior: a
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
comparison of current and lapsed donors. Transfusion
2007;47:1862-70.
Riley W, Schwei M, McCullough J. The United States'
potential blood donor pool: estimating the prevalence of
donor-exclusion factors on the pool of potential donors.
Transfusion 2007;47:1180-8.
Godin G, Sheeran P, Conner M, Germain M. Asking
questions changes behavior: mere measurement effects on
frequency of blood donation. Health Psychol 2008;27:
179-84.
James RC, Matthews DE. The donation cycle: a framework
for the measurement and analysis of blood donor return
behaviour. Vox Sang 1993;64:37-42.
Freedman JL, Fraser SC. Compliance without pressure: the
foot-in-the-door technique. J Pers Soc Psychol 1966;14:
195-202.
Cialdini RB, Vincent JE, Lewis SK, Catalan J, Wheeler D,
Darby BL. Reciprocal concessions procedure for inducing
compliance: the door-in-the-face technique. J Pers Soc
Psychol 1975;31:206-15.
Review Manager (RevMan) 5.1 ed. Copenhagen: The Nordic
Cochrane Center, The Cochrane Collaboration; 2008.
Borenstein M, Hedges LV, Higgins JPT, Rothstein HR.
Introduction to meta-analysis. Chippenham: John Wiley &
Sons; 2009.
Higgins JP, Thompson SG. Quantifying heterogeneity in a
meta-analysis. Stat Med 2002;21:1539-58.
Higgins JP, Thompson SG, Deeks JJ, Altman DG.
Measuring inconsistency in meta-analyses. BMJ 2003;
327:557-60.
Higgins JPT, Thompson SG, Spiegelhalter DJ. A
re-evaluation of random-effects meta-analysis. J R Statist
Soc 2009;172:137-59.
236
[22] Cohen J. A power primer. Psychol Bull 1992;112:155-9.
[23] Sutton AJ, Duval SJ, Tweedie RL, Abrams KR, Jones DR.
Empirical assessment of effect of publication bias on metaanalyses. BMJ 2000;320:1574-7.
[24] Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred
reporting items for systematic reviews and meta-analyses:
the PRISMA statement. Ann Intern Med 2009;151:264-9.
[25] Bem DJ. Self-perception theory. Adv Exp Soc Psychol
1972;6:1-62.
[26] Bandura A. Self-efficacy: toward a unifying theory of
behavioral change. Psychol Rev 1977;84:191-215.
[27] Bandura A. Social learning theory. Prentice-Hall: Englewood Cliff; 1977.
[28] Fishbein M, Ajzen I. Belief, attitude, intention and
behavior: an introduction to theory of research. Don
Mills: Addison-Wesley; 1975.
[29] Ajzen I. The theory of planned behavior. Organ Behav
Hum Decis Process 1991;50:179-211.
[30] Schwartz SH. Normative influences on altruism. Adv Exp
Soc Psychol 1977;10:221-79.
[31] Ditto B, France CR, Holly C. Applied tension may help
retain donors who are ambivalent about needles. Vox Sang
2010;98:e225-30.
[32] Sinclair KS, Campbell TS, Carey PM, Langevin E, Bowser
B, France CR. An adapted postdonation motivational
interview enhances blood donor retention. Transfusion
2010;50:1178-786.
[33] Evans DE. Development of intrinsic motivation for
voluntary blood donation among first-time donors [Ph.D.].
Wisconsin: The University of Wisconsin – Madison; 1981.
[34] Jason LA, Jackson K, Obradovic JL. Behavioral approaches in increasing blood donations. Eval Health Prof
1986;9:439-48.
[35] Sarason IG, Sarason BR, Pierce GR, Shearin EN. A social
learning approach to increasing blood donations. J Appl
Soc Psychol 1991;21:896-918.
[36] Chamla JH, Leland LS, Walsh K. Eliciting repeat blood
donations: tell early career donors why their blood type is
special and more will give again. Vox Sang 2006;90:
302-7.
[37] Cialdini RB, Ascani K. Test of a concession procedure
for inducing verbal, behavioral, and further compliance
with a request to give blood. J Appl Psychol 1976;61:
295-300.
[38] Clee MA, Henion KE. Blood donorship and psychological
reactance. Transfusion 1979;19:463-6.
[39] Ferrari JR, Leippe MR. Noncompliance with persuasive
appeals for a prosocial, altruistic act: blood donating. J Appl
Soc Psychol 1992;22:83-101.
[40] Foss RD, Dempsey CB. Blood donation and the foot-inthe-door technique: a limiting case. J Pers Soc Psychol
1979;37:580-90.
[41] Gimble JG, Kline L, Makris N, Muenz LR, Friedman LI.
Effects of new brochures on blood donor recruitment and
retention. Transfusion 1994;34:586-91.
[42] Hayes TJ, Dwyer FR, Greenwalt TJ, Coe NA. A comparison
of two behavioral influence techniques for improving blood
donor recruitment. Transfusion 1984;24:399-403.
[43] Jason LA, Rose T, Ferrari JR, Barone R. Personal versus
impersonal methods for recruiting blood donations. J Soc
Psychol 1984;123:139-40.
GODIN ET AL
[44] LaTour SA, Manrai AJ. Interactive impact of informational
and normative influence on donations. J Mark Res 1989;26:
327-35.
[45] Reich P, Roberts P, Laabs N, Chinn A, McEvoy P,
Hirschler N, et al. A randomized trial of blood donor
recruitment strategies. Transfusion 2006;46:1090-6.
[46] Royse D. Exploring ways to retain first-time volunteer
blood donors. Res Soc Work Pract 1999;9:76-85.
[47] Rushton J, Campbell A. Modelling, vicarious reinforcement
and extraversion and blood donating in adults: immediate and
long-term effects. Eur J Soc Psychol 1977;7:297-306.
[48] Sarason IG, Sarason BR, Pierce GR, Shearin EN, Sayers
MH. A social learning approach to increasing blood
donations. J Appl Soc Psychol 1991;21:896-918.
[49] Sarason IG, Sarason BR, Pierce GR, Sayers MH,
Rosenkranz SL. Promotion of high school blood donations:
testing the efficacy of a videotaped intervention. Transfusion 1992;32:818-23.
[50] Upton WE. Altruism, attribution, and intrinsic motivation
in the recruitment of blood donors [Ph.D.]. New York, NY:
Cornell University; 1973.
[51] Ferrari JR, Barone RC, Jason LA, Rose T. The effects of a
personal phone call prompt on blood donor commitment.
J Community Psychol 1985;13:295-8.
[52] Lipsitz A, Kallmeyer K, Ferguson M, Abas A. Counting on
blood donors: increasing the impact of reminder calls.
J Appl Soc Psychol 1989;19:1057-67.
[53] Pittman TS, Pallak MS, Riggs JM, Gotay CC. Increasing
blood donor pledge fulfillment. Pers Soc Psychol Bull
1981;7:195-200.
[54] Whitney JG, Hall RF. Using an integrated automated
system to optimize retention and increase frequency of
blood donations. Transfusion 2010;50:1618-24.
[55] Wiesenthal DL, Spindel L. The effect of telephone
messages/prompts on return rates of first-time blood
donors. J Community Psychol 1989;17:194-7.
[56] Cioffi D, Garner R. The effect of response options on
decisions and subsequent behavior: sometimes inaction is
better. Pers Soc Psychol Bull 1998;24:463-72.
[57] Godin G, Sheeran P, Conner M, Delage G, Germain M,
Bélanger-Gravel A, et al. Which survey questions change
behavior? Randomized controlled trial of mere measurement interventions. Health Psychol 2010;9:636-64.
[58] Ferrari JR, Barone RC, Jason LA, Rose T. The use of
incentives to increase blood donations. J Soc Psychol
1985;125:791-3.
[59] Ditto B, France CR, Albert M, Byrne N, Smyth-Laporte J.
Effects of applied muscle tension on the likelihood of blood
donor return. Transfusion 2009;49:858-62.
[60] Oswalt RM. A review of blood donor motivation and
recruitment. Transfusion 1977;17:123-35.
[61] Godin G, Amireault S, Vezina-Im LA, Germain M,
Delage G: The effects of a phone call prompt on subsequent blood donation among first-time donors. Transfusion, Epub aehad of print, June 9, 2011. DOI: 10.1111
(http://www.ncbi.nlm.nih.gov/pubmed/21658045).
[62] Abraham C, Michie S. A taxonomy of behavior change
techniques used in interventions. Health Psychol 2008;27:
379-87.
[63] Michie S, Johnston M, Abraham C, Lawton R, Parker D,
Walker A. Making psychological theory useful for
INTERVENTIONS PROMOTING BLOOD DONATION
[64]
[65]
[66]
[67]
implementing evidence based practice: a consensus
approach. Qual Saf Health Care 2005;14:26-33.
Godin G, Conner M, Sheeran P, Belanger-Gravel A,
Germain M. Determinants of repeated blood donation
among new and experienced blood donors. Transfusion
2007;47:1607-15.
Godin G, Sheeran P, Conner M, Germain M, Blondeau
D, Gagne C, et al. Factors explaining the intention to
give blood among the general population. Vox Sang
2005;89:140-9.
Armitage CJ, Conner M. Social cognitive determinants of blood donation. J Appl Soc Psychol 2001;31:
1431-57.
Dholakia UM. A critical review of question-behavior effect
research. Rev Mark Res 2010;7:145-97.
237
[68] Schulz KF, Altman DG, Moher D. CONSORT 2010
statement: updated guidelines for reporting parallel group
randomized trials. Ann Intern Med 2010;152:726-32.
[69] Lowe CF, Horne PJ, Tapper K, Bowdery M, Egerton C.
Effects of a peer modelling and rewards-based intervention
to increase fruit and vegetable consumption in children. Eur
J Clin Nutr 2004;58:510-22.
[70] Hardman CA, Horne PJ, Fergus Lowe C. Effects of
rewards, peer-modelling and pedometer targets on children's physical activity: a school-based intervention study.
Psychol Health 2011;26:3-21.
[71] Brehm JW. A theory of psychological reactance. New
York: Academic Press; 1966.
[72] Titmuss RM. The gift relationship: from human blood to
social policy. London: Allen & Unwin; 1970.
INTERVENTIONS PROMOTING BLOOD DONATION
237.e1
APPENDIX A
Supplementary file 1. Formula used to calculate 80% PIs and transformations applied.
The following formula and transformations were applied given that the effect sizes of the present review
were OR:
pffiffiffiffiffiffi
PI ¼ M * t T 2 þ Vm *
PI = e x (exponential of the result obtained)
Effect size [M*] = lnOR
t = t value
T 2 = τ 2 (an indicator of heterogeneity computed automatically by RevMan)
Variance [Vm⁎] = (lnCIupper − lnOR)/1.96 = SE 2
Source: Borenstein et al [18].
237.e2
Supplementary File 2. Summary of the Studies Included and Their Results
Reference
Chamla et al [36]
Theory
used
Objective of the study
Population
Study design
Sample
Type of intervention
Mode of delivery
Retention
of donors
RCT
N = 316 42.1% ♂
Cognitions based
Mail and
telephone
SCT
Unspecified
N = 189
College students
of both sexes
Foot-in-thedoor/door-in-theface
Face-to-face
RCM
OR (95% CI)
Unspecified
Cioffi and
Garner [56]
Unspecified
Unspecified
RCT †
N = 1121
College students
Measurement of
cognitions
E-mail
FPP
Clee and
Henion [38]
Retention
of donors
Novice donors
RCT †
N = 338
Cognitions based
Telephone
RT
Ditto et al [59]
Retention
of donors
RCT
Video and faceto-face
–
Retention of donors
Muscle tension
Video and faceto-face
–
Potential, novice, and
experienced donors
RCT
Cognitions based
Telephone
EAT
F2,440 = 0.392,
P = .67
Ferrari et al [51]
Recruitment
and retention
of donors
Unspecified
Unspecified
RCT †
N = 1209
Age, 21.9 ± 3.4 y
49.6% ♂
N = 726
Age Exp:
22.2 ± 6.8 y
Age Cont:
21.7 ± 5.8 y
41% ♂
N = 488
Age, 18 y
40.2% ♂
N = 68
Age, 18-20 y
Muscle tension
Ditto et al [31]
Novice and
experienced
donors
Experienced
donors
Reminders
Telephone
–
Ferrari et al [58]
Unspecified
Novice and
experienced
donors
RCT †
Incentives
Face-to-face
–
Exp: prompt condition
Cont: nonprompt
condition
13.91 (1.68-114.82)
Exp: E condition
Cont: C condition
3.86 (1.47-10.13)
Evans, 1981 [33]
(study 2)
RCT
N = 80 College
students
of both sexes
Exp: experimental
Cont: control
2.08 (1.14-3.80)
Exp: extremethen-critical
request condition
Cont: criticalrequest-only control
1.17 (0.36-3.78)
Exp: active-no
response option
Cont: no message sent
2.44 (1.04-5.73)
Exp: HN/LI
Cont: LN/HI
1.52 (0.64-3.62)
Exp: upper body
Cont: no treatment
1.48 (0.95-2.30) §
Exp: applied tension
Cont: no treatment
1.16 (0.82-1.66)) ⁎⁎
GODIN ET AL
Cialdini and
Ascani [37]
Novice and
experienced
donors
Unspecified
Potential
and donors †
RCT †
Unspecified
Potential
and donors †
RCT †
Foss and Dempsey
[40] (study 1)
Recruitment and
retention of donors
Potential
and donors †
Unspecified
Foss and Dempsey
[40] (study 2)
Recruitment and
retention of donors
Potential
and donors †
Unspecified
Foss and Dempsey
[40] (study 3)
Recruitment
and retention
of donors
Recruitment and
retention of donors
Potential and donors †
Unspecified
Novice, experienced,
and temporary
deferred donors
RCT
Godin et al [13]
Retention of donors
Novice and
experienced donors
RCT †
Godin et al [57]
Retention of donors
Novice donors
RCT
Hayes et al [42]
Recruitment and
retention of donors
Potential, novice, and
experienced donors
Jason et al [43]
(study 1)
Unspecified
Jason et al [43]
(study 2)
Jason et al [34]
Gimble et al [41]
N = 84
Age: 19.6 y
30.1% ♂
N = 112
Age, 18.8 y
19.6% ♂
N = 76
Dormitory residents
47.8% ♂
N = 135
Dormitory residents of
both sexes
Tape recorder
NAT
Exp: combined condition
Cont: no message
2.24 (0.36-13.78)
Exp: low RD/no access
Cont: high RD/no access
22.14 (2.58-190.16)
Exp: 3-d delay
Cont: control condition
1.75 (0.39-7.95)
Exp: poster request
Potential and donors ¶
Cont: control condition
1.07 (0.06-18.04)
Exp: poster request
Cont: control condition
7.73 (1.54-38.66)
Exp: group E
Cont: group C
1.11 (1.02-1.21)
Altruism
Tape recorder
NAT
Foot-in-thedoor/door-in-theface
Foot-in-thedoor/door-in-theface
Face-to-face
SPT
Telephone
SPT
N = 163
Dormitory residents
of both sexes
N = 65 874
Age, 38-40 y
Exp: 57%-63% ♂
Cont.: 57-59% ♂
N = 4672
Age Exp: 44.7 ± 11.8 y
Age Cont: 43.8 ± 12.1 y
Exp: 61.7% ♂
Cont: 61.3% ♂
N = 4391
Age, 30.4 ± 12.9 y
47% ♂
Foot-in-thedoor/door-in-theface
Cognitions based
Face-to-face
and telephone
SPT
Unspecified
CIPT
Measurement of
cognitions
Mail
TPB
Exp: experimental group
Cont: control group
1.24 (1.09-1.41) ††
Measurement of
cognitions
Mail and
telephone
TPB
RCT †
N = 914
Adults
Telephone
SPT
Unspecified
RCT †
N = 300
College students
Foot-in-thedoor/door-in-theface
Cognitions based
Mail and
telephone
–
Recruitment of
donors
Unspecified
QE
N = 64
College students
Cognitions based
Face-to-face
–
Unspecified
Unspecified
QE
Unspecified N
Workers and
university students
Incentives
Face-to-face
–
Exp: implementation
intention
Cont: control condition
1.16 (0.98-1.36)
Exp: FID
Cont.: control request
2.82 (1.69-4.69)
Exp: ML
Cont: M
4.13 (0.45-37.57)
Exp: friends
Cont: strangers
2.86 (0.80-10.24)
F1,25 = 12.30, P b .01
(continued on next page)
237.e3
Altruism
INTERVENTIONS PROMOTING BLOOD DONATION
Unspecified
Ferrari and
Leippe [39]
(study 1)
Ferrari and Leippe [39]
(study 2)
237.e4
Supplementary File 2. (continued)
Reference
Objective of the study
Population
Study design
Unspecified
Unspecified
LaTour and
Manrai [44] (study 2)
Unspecified
Unspecified
RCT ⁎
Lipsitz et al [52]
(study 1)
Unspecified
Unspecified
RCT ⁎
L i p s i t z e t a l [ 52 ]
(study 2)
Unspecified
Unspecified
RCT ⁎
Pittman et al [53]
(study 1)
Unspecified
Unspecified
RCT ⁎
Pittman et al [53]
(study 2)
Unspecified
Unspecified
RCT ⁎
Reich et al [45]
Retention of donors
Novice donors
RCT
Royse [46]
Retention of donors
Novice donors
RCT ⁎
Unspecified
Unspecified
Unspecified
Sarason et al [48]
Recruitment and
retention of donors
Potential and
novice donors
Sarason et al [49]
Recruitment and
retention of donors
Potential and
novice donors
Rushton and
Campbell [47]
Type of intervention
Theory
used
Mode of delivery
OR (95% CI)
N = 680
Residents of a
Midwestern community
Cognitions based
Mail and
telephone
TRA
N = 1200
Residents of a
Midwestern community
Cognitions based
Mail and
telephone
TRA
N = 156
College students
Reminders
Telephone
–
N = 156
University students
Reminders
Telephone
–
N = 284
College students,
faculty, and staff
N = 122
College students,
faculty, and staff
N = 6919
52.5% ♂
Reminders
Mail
–
Reminders
Mail
–
Altruism
Telephone and
e-mail
–
N = 1003
Median age, 22 y
46% ♂
N = 43
Age, 18-21 y
All ♀
Foot-in-thedoor/door-in-theface
Modeling
Mail
SPT
Face-to-face
SLT
Unspecified
N = 9378
High school students
Modeling
Face-to-face
and video
SLT
Unspecified
N = 4970
High school students
48% ♂
Modeling
Face-to-face
and video
SLT
Exp: informational and
normative influences
Cont: no informational
nor normative influences
13.70 (4.74-39.65)
Exp: strong informational
influence and
normative influence
Cont: no informational
nor normative influence
4.17 (2.01-8.66)
Exp: experimental
Cont: control
2.63 (1.27-5.42)
Exp: experimental condition
Cont: control condition
2.86 (1.01-8.10)
Exp: 3-name reminder
Cont: 1-name reminder
1.59 (0.96-2.64)
Exp: group reminder) O
Cont: 1-name reminder
2.27 (0.79-6.49)
Exp: script A
Cont: script B
1.23 (1.09-1.38) #
Exp: volunteer group
Cont: control group
1.12 (0.79-1.59)
Exp: vicarious
punishment condition
Cont: no-model condition
39.67 (1.28-1229.87)
Combined treatment
(55.4%) vs control
(44.6%) = P b .013 ‡‡
Exp: videotape
Cont: control presentation
1.25 (1.06-1.47)
GODIN ET AL
LaTour and
Manrai [44] (study 1)
RCT ⁎
Sample
Retention of donors
Novice and
experienced donors
RCT
N = 427
Age, 31.1 ± 13.5 y
Exp: 46.2% ♂
Cont: 35.8% ♂
Cognitions based
Telephone
–
Upton [50]
Retention of donors
Experienced and lapsed
donors
Unspecified
N = 1261
82% ♂
Altruism
Telephone
TOA
Whitney and
Hall [54]
Retention of donors
Novice and experienced
donors
QE
N = 285 879
54% ♂
Reminders
Telephone
–
Wiesenthal
and Spindel [55]
Retention of donors
Novice donors
Unspecified
N = 209
Reminders
Telephone
–
Exp: motivational
interview group
Cont: no-interview
control group
1.61 (0.93-2.78)
Exp: low reward/
high motivation
Cont: high motivation/
high reward
6.91 (4.22-11.34)
Exp: heard message
Cont: did not
hear message
1.09 (1.07-1.12)
Exp: experimental
condition 1
Cont: control group
1.84 (0.76-4.49)
INTERVENTIONS PROMOTING BLOOD DONATION
Sinclair et al [32] ‡
The symbol (♂) represents the percentage of the sample that is male.
Abbreviations: NAT indicates norm activation theory [30]; CIPT, consumer information processing theory; Cont, control group used to compute the effect size; EAT, endogenous attribution theory;
Exp, experimental group used to compute the effect size; FPP, feature-positive paradigm; QE, quasi-experimental design; RCM, reciprocal concession model [16]; RCT, randomized, controlled trial; RT:
reactance theory [71]; SCT, social cognitive theory [26]; SLT, social learning theory [27]; SPT, self-perception theory [25]; TOA, theory of altruism [72]; TPB, theory of planned behavior [29]; TRA, theory of
reasoned action [28]; HN/LI, high need/low implication; LN/HI, low implication/high need; low RD, low reponsibility denial; high RD, high responsibility denial; FID, foot-in-the-door; ML, media plus a
personal letter; M, media only.
⁎ No information on the randomization procedure.
†
The authors did not specify which type of donors was included in their study.
‡
Although the study had a 12-month follow-up, we only used the 9-month outcome to have intention-to-treat results.
§
The results of men and women were combined within the control and the experimental groups.
O
The group reminder condition was used as the experimental group to compute the effect size to differentiate it from the study 1 of Pittman et al [53].
¶
The poster request condition was used as the experimental group to compute the effect size, given that it was testing the effect of the foot-in-the-door technique.
#
The results reported are for the second donation (ie, percentage of first-time donors who returned for a second donation).
⁎⁎ Analyses controlled for age, sex, previous blood donations, experience, and body mass index.
††
Analyses controlled for age.
‡‡
Analyses controlled for school, blood center, donation history, and size.
237.e5
237.e6
Supplementary file 3. Funnel plot of cognitionbased interventions (s = 8).
GODIN ET AL
Supplementary file 5. Formula used to convert
SMD into OR.
pffiffiffi
logOR ¼ dðπ= 3Þ
d = SMD
Source: Borenstein et al [18].
Supplementary file 6. Funnel plot of interventions delivered by telephone (s = 10).
Supplementary file 4. Funnel plot of interventions using reminders (s = 7).
Efficacy of Interventions Promoting Blood Donation:
A Systematic Review
Gaston Godin, Lydi-Anne Vézina-Im, Ariane Bélanger-Gravel, and Steve Amireault
Findings about the efficacy of interventions promoting
blood donation are scattered and sometime inconsistent. The aim of the present systematic review
was to identify the most effective types of interventions and modes of delivery to increase blood
donation. The following databases were investigated:
MEDLINE/PubMed, PsycINFO, CINAHL, EMBASE, and
Proquest Dissertations and Theses. Additional studies
were also included by checking the references of the
articles included in the review and by looking at our
personal collection. The outcomes of interest were
either blood drive attendance or blood donations. A
total of 29 randomized controlled trials or quasiexperimental studies were included in the review,
detailing 36 interventions tested among independent
samples. Interventions targeting psychosocial cognitions (s = 8, s to represent the number of
independent samples; odds ratio [OR], 2.47; 95%
confidence interval [CI], 1.42-4.28), those stressing
the altruistic motives to give blood (s = 4; OR, 3.89;
95% CI, 1.03-14.76), and reminders (s = 7; OR,
1.91; 95% CI, 1.22-2.99) were the most successful
in increasing blood donation. The results suggest that
motivational interventions and reminders are the most
effective in increasing blood donation, but additional
studies are needed to evaluate the efficacy of other
types of interventions.
© 2012 Elsevier Inc. All rights reserved.
LOOD AGENCIES ARE investing much
effort to recruit new blood donors and to
ensure that those who have donated return for
additional donations. This preoccupation is wellfounded given that, according to the American Red
Cross, only 3% of the population gives blood [1].
To further complicate matters, blood donation is a
cyclic process in which once an individual has
donated blood, he/she must wait a certain time (eg,
56 days) before being eligible to give blood again
[2]. Moreover, many medical and nonmedical
reasons such as foreign traveling can prevent
someone from successfully giving blood [2]. This
is done to ensure donor and recipient safety [3].
Effective interventions to retain blood donors and
recruit new donors are thus imperative to maintain a
safe and sufficient level of blood supply.
In a previous review on the motives and the
characteristics of blood donors, Piliavin [4] made
an inventory of potential successful approaches to
increase blood donation, particularly among first-
time donors. For instance, this author suggested that
an increase in the perception for the need of blood
in the community and in the perception of the
adequate support for blood donation provided in
these communities as well as the use of modeling
were promising approaches to increase blood
donation. The author also observed that face-toface solicitations, the use of personal reminders,
and the foot-in-the-door technique (consisting of a
small request followed by a bigger one) appeared to
be effective techniques to motivate individuals
to give blood, given that individuals are sensitive to
social pressure. These observations were also
supported in a more recent review by Ferguson
et al [5], in which the use of reminders and the footin-the-door technique were mentioned as effective
strategies to favor blood donation.
Mixed evidences were observed concerning the
efficacy of interventions aimed at increasing
altruistic motives vs the use of different forms of
incentives [4]. However, as pointed out earlier by
Piliavin [4], the use of incentives might undermine
the development of an altruistic motivation and
provoke a backfire effect on future donations
among active blood donors. Moreover, in a recent
review of the literature among active blood donors,
Ringwald et al [6] mentioned that the use of
incentives might not be the best strategy to keep
long-term and committed blood donors given that
as a donor advances in his/her career, his/her
motives for giving blood become more internalized.
The role of altruism also needs some consideration.
According to a deeper analysis by Ferguson et al [5]
B
From the Canada Research Chair on Behaviour and Health,
Laval University, Québec, Canada; Faculty of Nursing, Laval
University, Québec, Canada; and Department of Social and
Preventive Medicine, Laval University, Québec, Canada.
The authors declare that they have no conflict of interest.
Address reprint requests to Gaston Godin, PhD, Canada
Research Chair on Behaviour and Health, Laval University,
Québec, Canada G1V 0A6.
E-mail: gaston.godin@fsi.ulaval.ca
0887-7963/$ - see front matter
© 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.tmrv.2011.10.001
224
Transfusion Medicine Reviews, Vol 26, No 3 (July), 2012: pp 224-237.e6
INTERVENTIONS PROMOTING BLOOD DONATION
of previous studies aimed at promoting blood
donation with the use of the altruism concept,
results seem to provide more support for the use of
the moral norm concept or the benevolence
response rather than pure altruism. Consequently,
the role of altruism in the promotion of blood
donation has to be more clearly defined to further
increase the efficacy of such interventions.
Bednall and Bove [7] also did a review of selfreported motivators and deterrents of blood donation. In particular, they verified that the motives
behind blood donation differed among different
types of donors (first-time, repeat, lapsed, apheresis, and eligible nondonors). They concluded that,
among first-time and repeat donors, the most
common motivators were convenience (eg, blood
drive nearby), prosocial motivation (eg, altruism),
and personal values (eg, moral norm). They also
identified several deterrents among which the most
frequently mentioned barrier was low self-efficacy
to donate (eg, ability to overcome barriers such as
lack of time).
Notwithstanding the very useful information
provided by these previous reviews concerning
the potential efficacy of different strategies to
increase blood donation, none of them systematically reviewed the literature and provided information on the effect sizes of such approaches. In
addition, in the review by Ferguson et al [5],
studies whose outcome was the intention or
willingness to give blood were also included in
their narrative analysis. Acknowledging the scientific evidence that intention is identified as one of
the most important determinants of blood donation
[2], it is, nonetheless, well documented that not all
people will act on their stated intention, that is there
is an “intention-behavior gap” [8]. Moreover, in a
review by Webb and Sheeran [9] on the effectiveness of interventions to promote health-related
behaviors, it was observed that a medium-to-large
effect size in increasing levels of intention leads to
a small-to-medium effect in changing behaviors.
Consequently, the efficacy of interventions on
blood donation and not only intention or willingness has to be determined.
To our knowledge, there is no systematic review
on the efficacy of interventions to promote blood
donation. Thus, the objective of the present
systematic review was to identify the most effective
types of interventions and their modes of delivery to
increase blood donation.
225
METHODS
Study Eligibility Criteria
The focus of the present systematic review was
on interventions aimed at increasing blood donation
among the general population (ie, recruitment
studies) and blood donors (ie, retention studies).
Studies aimed at increasing different or specific
types of donations such as bone marrow donation or
apheresis were not included in the present review.
To evaluate the efficacy of interventions, only
experimental and quasi-experimental studies were
included in the review; studies adopting a 1-group
pre-post design were excluded.
Studies concerned with different stages in the
lifetime experience of blood donors, that is, novice,
experienced, lapsed, or temporarily deferred donors,
were included. According to Masser [10], individuals who already gave blood once were considered
as novice donors (also referred to as first-time
donors), and individuals who already gave blood
more than once were considered as experienced
donors (also referred to as repeat donors) 1. Studies
interested in lapsed donors (ie, individuals who
already gave blood in the past but who did not
donate blood in the past 2 years without any known
cause for permanent or temporary deferral [11]) or
temporarily deferred donors (ie, individuals who
attended a blood center but were not eligible to
donate blood for a certain period because of various
medical reasons [12]) were also considered in the
present systematic review.
Studies reporting results on the attendance at a
blood drive or the number of successful blood
donations after the exposure to an intervention were
both included in the present systematic review.
Attendance was defined as the number of participants
who registered at a blood drive. This outcome was
chosen given that although some individuals can be
deferred for various medical reasons, attendance
takes into account the fact that participants had acted
toward their goal of giving blood after the exposure
to the intervention [13,14]. To avoid duplication of
the results, attendance was primarily used in the
statistical analyses in cases where attendance and the
number of actual blood donations were both provided
in a given study. Studies reporting the results of an
1
Other classifications can be found in the scientific literature
(eg, Ferguson and Bibby (2002): occasional, b5 donations;
regular, ≥5 donations).
226
GODIN ET AL
intervention on the level of intention or willingness as
well as studies reporting results in terms of
physiologic reaction, stress, and symptoms related
to blood donation were excluded from the present
systematic review.
Search Strategy
The following databases were investigated:
MEDLINE/PubMed (1950+), PsycINFO (1806+),
CINAHL (1982+), and EMBASE (1974+). Proquest Dissertations and Theses (1861+) was also
investigated for gray literature (ie, unpublished
trials). No restriction was placed on the year of
publication of the articles. The search was
performed between June 8 and July 31, 2010. In
all the databases, the search terms were always
related to 2 themes, that is, blood donation and
intervention (see Table 1). In MEDLINE/PubMed,
a combination of keywords and MeSH terms was
used. In PsycINFO, only keywords were used
because no psychological index terms corresponded to blood donation and intervention. In
CINAHL, a combination of keywords and descriptors was used. In EMBASE, a combination of
keywords and Emtree terms was used. Finally, in
Proquest Dissertations and Theses, only keywords
were used. The search was limited to studies
published in English. Additional studies were also
included by checking the references of the articles
included in the systematic review (ie, secondary
references) as well as by looking at our personal
collection of articles on blood donation.
Data Extraction
Data were independently extracted by GG and
LAVI using a standardized data extraction form (see
Table 2). Disagreements were resolved by discussion. Before extracting the data, several decisions
were made. First, in instances where there was more
than 1 experimental group in a study, only the group
that had the greatest impact on blood donation was
Table 1. Search Terms Used for Investigating All 5 Databases
Keywords
Theme 1: blood donation
Theme 2: intervention
Blood donation OR
Blood donor OR
Blood bank
Intervention OR
Program OR
Education
considered and compared with the control group.
This decision was taken given that the main objective
of the present review was to verify which types of
intervention were more successful in increasing
blood donation and also because this group corresponded to the hypothesis stated in the articles
included in the meta-analysis in 90% of the cases.
Second, when several tests of an intervention were
performed among independent samples but reported
in 1 article, these were all included in the analysis (ie,
studies 1 and 2). Consequently, the number of
independent comparisons in the meta-analysis might
be more important than the number of included
studies in some of the subgroup analyses. Finally,
authors were personally contacted by e-mail for
additional information when necessary.
Interventions were classified according to an
adapted version of the categorization proposed in a
recent review by Ferguson et al [5] (see Table 3).
These authors regrouped the social interventions
into 4 types: (1) altruism-egoism, (2) reminders and
commitments, (3) foot-in-the-door, and (4) intention based. Social interventions were aimed at
increasing blood donation and were mainly focused
on recruitment. They also identified behavioral
interventions that can be classified in 3 categories:
(5) distraction, (6) muscle tension, and (7) caffeine
and water loading. These behavioral interventions
are primarily aimed at reducing unpleasant symptoms associated with blood donation and mainly
focused on retention of donors.
The first category of social interventions
included interventions that use manipulations of
altruism and egoism to encourage blood donation.
In the present review, however, the interventions
based on altruism were imbedded in a larger
motivational category that included all the interventions aimed at increasing motivation toward
blood donation. These studies used techniques
such as providing information about positive or
negative consequences of blood donation, barriers
management, modeling, comparison with the
social norm, providing feedback and praise, and
other. However, given the importance of the
concept of altruism in the literature on blood
donation, the independent efficacy of interventions based on altruism was compared with the
efficacy of other motivational interventions in the
subgroup analyses. The second category included
interventions generally focusing on retention that
use a reminder, such as a telephone call, before
INTERVENTIONS PROMOTING BLOOD DONATION
227
Table 2. Information Contained in the Data Extraction Form
Type of information
Subcategory
Main objective of the study
Recruit new donors
Encourage donors to return
Both objectives
Unspecified
Potential donors/general population
Novice donors
Experienced donors
Unspecified
Randomized, controlled trial
Quasi-experimental design
Type of assignment
Strategy used for assignment
Technique used for assignment
Person who performed assignment
Type of control group
Motivational
Reminders
Measurement of cognitions
Incentives
Muscle tension
Face-to-face
By mail (eg, letter)
By telephone (eg, telephone call prompt)
By e-mail
By tape recorder
Mix
Yes
No
Registrations at a blood drive
Blood donations
Both
Type of behavioral measure
Validation of the measure
Size
Attrition at postintervention
Attrition at follow-up
Characteristics of participants at baseline
Equivalence of groups at baseline
Dropout analysis
Statistical test used
Intention to treat
Covariates
Technique to replace missing data
Percentage of participants who registered at a blood drive
Percentage of participants who gave blood
Other
Population of the study
Study design
Group allocation
Type of intervention
Mode of delivery
Use of a theory
Behavior measured
Behavioral measure
Sample
Statistical analyses
Results of the intervention
the blood drive takes place. The third category
was the foot-in-the-door [15] or door-in-the-face
[16] interventions that are well-known techniques
issued from the social psychology literature. The
last category, that is, the intention-based category,
initially included interventions that manipulated
intentions. Given that the intention-based category
was too specific to the measurement of intention,
a new category was created (ie, measurement of
cognitions or the question-behavior effect) to
include all studies that measured cognitions with
the specific objective of promoting blood donation. Finally, the use of incentives (financial or
not) was added as a separate category, and the
behavioral categories were classified using the
description provided by Ferguson et al [5]. All of
228
GODIN ET AL
Table 3. Classification of the Interventions Based on the Categorization of Ferguson et al [5]
Type of intervention
Motivational
Cognitions based
Foot-in-the-door/door-in-the-face
Altruism
Modeling
Reminders
Measurement of cognitions
Incentives
Muscle tension
Definition
Interventions aimed at increasing motivation toward blood donation
Interventions targeting psychosocial cognitions related to motivation,
such as social norms, attitudes, and barriers
Interventions using the foot-in-the-door, the door-in-the-face,
or a combination of both techniques to motivate people to give blood
Foot-in-the-door involves asking a small request that should be accepted
and then asking a critical large request. Door-in-the-face involves
asking a large request that should be refused
and then asking a critical small request.
Interventions using altruistic motives to motivate people to give blood
Interventions showing another person giving blood to motivate
people to give blood
Interventions using reminders about the next eligibility date and/or the next
appointment to give blood (eg, telephone call prompt)
Interventions using the completion of a questionnaire about the intention to give
blood to activate cognitions about blood donation (eg, question-behavior effect)
Interventions using incentives for donating blood such as a T-shirt, money,
prizes, tickets, and other
Interventions using the contraction of muscles during blood donation to avoid
dizziness and fainting
the interventions were independently classified
according to their type by LAVI and ABG,
and disagreements were resolved by a third
reviewer (GG).
Data Analyses
Descriptive statistics of the studies, such as
frequencies and means, was analyzed using SAS
version 9.1 (SAS Institute, Cary, NC). All the other
statistical analyses were done using Cochrane's
Review Manager 5.1 (The Nordic Cochrane Center,
Copenhagen, Denmark) [17] statistical software. To
compute pooled effect sizes, a random-effect model
using the Mantel-Haenszel method was chosen for
all the analyses because we assumed that the
magnitude of the effect sizes would vary across
studies given the difference in samples and interventions across studies [18]. When computing the
pooled effect size, each study was weighted
according to its sample size. Between-study heterogeneity was verified by using 2 common statistical
approaches: a χ 2 test (Cochran's Q) and the I 2
statistic [19], representing the percentage of total
variation in estimated effects that is due to
heterogeneity rather than chance. A significant Q
statistic (P b .05) indicates significant heterogeneity
between the studies, whereas an I 2 statistic of 25% is
considered low heterogeneity; 50%, moderate
heterogeneity; and 75%, high heterogeneity [20].
To test the robustness of the results, separate 80%
prediction intervals (PIs) were calculated for each
pooled effect size (see Supplementary file 1 online
for the formula used). The computation of PIs is
recommended when confronted with a randomeffect meta-analysis that has a low number of studies
and high heterogeneity [21]. Prediction intervals
show the distribution of effect sizes around the
pooled effect size. The effect sizes of intervention
studies were reported as odds ratios (ORs) given that
most of the outcomes were reported as dichotomous
data (ie, the proportion of attendance or blood
donation in the experimental group compared with
the control group). Unless otherwise stated, all effect
sizes are zero order (ie, no covariates are included in
the computation of the effect size). The ORs were
also converted to Cohen's d [22] to facilitate
interpretation and allow comparison with standard
effect sizes reported in other studies. A Cohen's d of
0.20 is considered a small effect size; 0.50, a
medium effect size; and 0.80, a large effect size [22].
The efficacy of the interventions according to type
of intervention and mode of delivery was also
examined by means of prespecified subgroup
analyses. However, pooled effect sizes were computed only when the category contained results from
a minimum of 3 independent samples. Finally,
publication bias was assessed by visually inspecting
the distribution of the funnel plot in RevMan 5.1
(The Nordic Cochrane Center) in instances where
there were about 10 studies per category [23].
INTERVENTIONS PROMOTING BLOOD DONATION
RESULTS
The results of the search strategy are presented
in Figure 1 [24]. A total of 29 studies (k = 29)
were included in the present review, detailing 36
different interventions tested among independent
samples (s = 36). In the rest of the text, the letter
k will be used to represent the number of studies,
and the letter s to represent the number of
independent samples.
Characteristics of the Interventions
A summary of the interventions is given in
Supplementary file 2. When no effect sizes could be
229
calculated because of missing information such as
the number of participants per condition, the original results are reported. Most interventions were
aimed at retaining current donors (s = 12) or both
recruiting new donors and retaining donors (s = 8).
One intervention had the objective to recruit new
donors. In 15 interventions, the objective did not
clearly specify if it was more about recruiting or
retaining blood donors or vice versa. Again, few
studies clearly specified the type of donors targeted
by the intervention; only 6 studies mentioned this
information (novice donors, s = 5; experienced
donors, s = 1). Given that more than 3 studies were
targeting novice donors, a pooled effect size was
Fig 1. Preferred reporting for systematic reviews and meta-analyses flowchart [24].
230
GODIN ET AL
calculated (s = 5; OR, 1.21; 95% confidence
interval [CI], 1.10-1.32; Q = 1.68; P = .80; I 2 =
0%); it represented a small effect size (d = 0.11). No
significant heterogeneity between the studies was
detected. Unfortunately, given the low number of
studies, no pooled effect sizes could be calculated
for other types of donors, such as experienced,
lapsed, and deferred donors.
Of the 36 interventions included in the review,
24 were randomized controlled trials, although
most of them (s = 16) did not provide information
on the randomization procedure. The other 12
interventions either adopted a quasi-experimental
design (s = 3) or did not provide information
about the design adopted (s = 9). Although it
would have been interesting to compare the
difference in effect sizes between both pure and
quasi-experimental studies, 2 problems arose: (1)
many studies did not specify their study design
(s = 9) and (2) older studies tended to have little
information about their randomization procedure
compared with more recent studies; this phenomenon could reflect different editorial requirements
through the years.
Twenty-one interventions were clearly based
on a theory. The theories most frequently used to
design interventions promoting blood donations
were the self-perception theory [25] (s = 4), the
social learning/social cognitive theory [26,27]
(s = 4), the theory of reasoned action [28] and
its extension, the theory of planned behavior [29]
(s = 4), and Schwartz's norm activation theory
[30] (s = 2).
Characteristics of the Participants
Only 12 studies provided information on the
age of their participants. The mean age of the
participants was 25.72 ± 8.61 years (range,
18-44.3 years). Eighteen interventions were
conducted among samples of college students,
and 3 others, among samples of high school
students. Seventeen interventions indicated the
percentage of male participants in their sample.
A little less than half (44.8%) of the samples
were composed of male respondents. Three
interventions mentioned having samples containing participants of both sexes without specifying
the percentage of male respondents. One intervention had a sample exclusively of female
participants. The typical sample was thus composed of college students with slightly more than
half being female students.
Efficacy of the Interventions
A summary of the pooled effect sizes computed,
their heterogeneity, and their robustness is presented in Table 4. Where exact information about
numbers of participants or effect sizes were
missing [31-35], we attempted to obtain the
relevant information directly from the authors. In
3 cases [33-35], we were unable to obtain this
information. Thus, of the 29 studies describing 36
Table 4. Pooled Effect Sizes, OR with 95% CIs
Variable
Type of intervention
Motivational
Cognitions based
Foot-in-the door/door-in-the-face
Altruism
Reminders
Measurement of cognitions
Mode of delivery
Face-to-face
Mail
Telephone
Mix of modes
s
n
OR (95% CI)
20
8
6
4
7
3
37 102
23 198
1303
7619
286 615
6812
2.09 (1.64-2.67)
2.47 (1.42-4.28)
1.86 (1.02-3.39)
3.89 (1.03-14.76)
1.91 (1.22-2.99)
1.23 (1.06-1.41)
5
4
10
10
239
5536
287 882
15 896
2.57
1.25
2.41
1.60
(1.29-5.11)
(1.12-1.40)
(1.38-4.20)
(1.27-2.00)
Q
I2
(80% PI)
117.21 ‡
42.18 ‡
12.72 ⁎
51.43 ‡
19.71 †
3.01
84%
83%
61%
94%
70%
34%
(1.23-3.53)
(0.88-6.97)
(0.74-4.69)
(0.30-50.96)
(0.91-4.01)
(0.85-1.79)
5.14
2.49
84.50 ‡
40.23 ‡
22%
0%
89%
78%
(1.11-5.95)
(0.89-1.76)
(0.77-7.50)
(1.07-2.39)
NOTE. I2 is an indicator of heterogeneity. Twenty-five percent is considered low heterogeneity; 50%, moderate heterogeneity; and 75%,
high heterogeneity.
⁎ P b .05, significant heterogeneity (Q statistic).
†
P b .01, significant heterogeneity (Q statistic).
‡
P b .001, significant heterogeneity (Q statistic).
INTERVENTIONS PROMOTING BLOOD DONATION
interventions, 26 studies describing 33 interventions among independent samples could be
included in the meta-analysis.
Motivational Interventions
Twenty-two interventions [32,33,36-50] were
classified in the motivational category. However,
2 interventions [33,48] did not report the number of
participants in their samples and were thus excluded
from the meta-analysis. The study of Evans [33]
reported nonsignificant results (P = .67), whereas
the study of Sarason et al [48] had significant results
(P b .01). The pooled effect size (s = 20) indicated
that motivational interventions were successful in
increasing blood drive attendance, and this represented a small-to-medium effect size (d = 0.41).
Significant heterogeneity between the studies was
detected, which justified subgroup analyses. The
80% PI indicated, however, that the significant
result was robust.
Cognitions-Based Interventions
Nine interventions [32,33,36,38,41,43,44] were
classified as cognitions based because they were
targeting psychosocial cognitions related to motivation, such as social norms, attitudes, and barriers.
However, 1 intervention [33] could not be entered
in the meta-analysis because it did provide the
number of participants per condition (ie, in the
experimental and control groups). This study
reported nonsignificant results (P = .67). The
pooled effect size (s = 8) indicated that cognitions-based interventions were successful in increasing attendance at blood drives, and this
represented a medium effect size (d = 0.50).
Significant heterogeneity was detected, and examination of the 80% PI indicated that the result
should be interpreted carefully. Moreover, the
examination of the funnel plot (see Supplementary
file 3 online) indicated that there is probably a
publication bias [23].
Foot-in-the-Door and Door-in-the-Face
Techniques
A total of 6 interventions [37,40,42,46] were using
the foot-in-the-door, the door-in-the-face, or a
combination of both techniques to encourage people
to give blood. The pooled effect size (s = 6) indicated
that these types of interventions were successful at
increasing attendance at blood drives. When the
pooled OR was converted to Cohen's d [22], it
231
represented a small-to-medium effect size (d = 0.34).
The Q statistic indicated that there was significant
heterogeneity between the studies, although the I 2
statistic was still below 75%. The 80% PI indicated
that the result is not very robust.
Altruism
Four interventions [39,45,50] aimed to increase
blood donation by stressing the altruistic reasons to
give blood. The pooled effect size (s = 4) revealed
that this kind of intervention was successful in
increasing blood drive attendance; this represented
a medium-to-large effect size (d = 0.75). Significant
heterogeneity between the studies was detected,
and the 80% PI was very large, which indicates that
the result is not robust at all.
Modeling
Three interventions [47-49] were using modeling, such as showing a model giving blood, to
increase blood donations. However, 1 study [48]
did not report the exact number of participants in
their experimental and control groups. This study
by Sarason et al [48] reported that their intervention had significant results (P b .01). Given that
only 2 studies could have been entered in a metaanalysis, no pooled effect size was computed for
this type of intervention. Examination of the
individual effect sizes of the studies indicated that
all the studies reported significant results (Rushton
and Campbell [47]: OR, 39.67; 95% CI, 1.281229.87; and Sarason et al [49]: OR, 1.25; 95% CI,
1.06-1.47). Additional studies will be needed
before any effect size can be computed about the
efficacy of interventions that use modeling to
increase blood donation.
Reminders
Seven interventions [51-55] were using reminders, such as telephone call prompts, as a
means to increase blood donations. The pooled
effect size (s = 7) for this type of intervention
indicated that they were successful in increasing
attendance at blood drives, and this represented a
small-to-medium effect size (d = 0.36). Significant
heterogeneity between the studies was detected,
and the 80% PI was close to significance, which
indicates that the result is somewhat robust. The
inspection of the funnel plot (see Supplementary
file 4 online) indicated the likely presence of a
publication bias [23].
232
GODIN ET AL
Measurement of Cognitions
Modes of Delivery
Three interventions [13,56,57] used answering
questions about cognitions related to blood donation to increase this behavior. One study [57]
reported a continuous outcome (ie, mean number of
blood donations), which allowed the calculation of
a standard mean difference (SMD). This SMD was
then converted to an OR to pool it with the rest of
the studies (see Supplementary file 5 online for the
formula used). The pooled effect size (s = 3)
indicated that interventions using the measurement
of cognitions are effective in increasing blood
donation and that no significant heterogeneity was
detected. Once converted to Cohen's d, the pooled
effect size represented a small effect size (d = 0.11).
The 80% PI, however, indicated that the result
should be interpreted carefully.
The efficacy of interventions according to their
mode of delivery was also evaluated. In 1
intervention [41], a brochure was used to contact
participants, but it was not specified if it was sent by
mail or if it was given in person, thereby making it
impossible to classify its mode of delivery.
Incentives
Two interventions [34,58] used incentives, such
as tickets to a Broadway play and a college football
game, gift certificates, prizes, free meals, and other,
to increase blood donation. One intervention [34]
did not report the exact number of study participants but stated that the intervention was successful
in increasing blood drive attendance (P b .01). The
other study by Ferrari et al [58] also had significant
results (OR, 3.86; 95% CI, 1.47-10.13). The low
number of studies using incentives and the fact that
only 1 study reported the information necessary to
compute an effect size precluded any meta-analysis.
Additional studies will be needed to verify the
accuracy of the present finding and before any
definite conclusion can be reached about the
efficacy of incentives to increase blood donation.
Muscle Tension
Two interventions [31,59] used muscle tension
to decrease negative symptoms associated to
blood donation such as dizziness and fainting to
increase blood donors return. Given the low
number of studies using this technique, no effect
size was computed. Examination of the individual
effect sizes indicated that both studies had
nonsignificant results (Ditto et al [59]: OR,
1.48; 95% CI, 0.95-2.3...
Purchase answer to see full
attachment