DIGITAL
Original Research HEALTH
Digital Health
Design and testing of communication materials Volume 6: 1–12! The Author(s) 2020
Article reuse guidelines:
for a breastfeeding social media marketing sagepub.com/journals-
permissions
campaign: Breastfeed4Ghana DOI: 10.1177/2055207620909291
journals.sagepub.com/home/dhj
R Aryeetey1 , O Lasisi1, A Hromi-Fiedler2, G Carroll2, R P
erez-Escamilla2
and K Harding2
Abstract
Breastfeed4Ghana was a social media-based campaign implemented to address identified gaps in the protection, promo-
tion, and support of breastfeeding in Ghana. This paper describes the process of campaign materials development and
testing to ensure their cultural and content appropriateness. The 60 campaign materials, each consisting of an image and
text message, underwent a process of creation, testing, revision, and finalization. Existing research evidence and infant and
young child feeding communication tools that were culturally relevant for Ghana were used to develop the materials. All
materials were tested and finalized through an iterative process that incorporated input from six focus group discussions
(FGDs) with mothers, and content and technical experts. The materials were revised to ensure scientific accuracy, under-
standability, and cultural appropriateness of the messages, as well as alignment of the messages with the images. Finalized
materials were reviewed and approved by the Ghana Food and Drugs Authority. Analysis for this paper involved summa-
rizing and categorizing the types and sources of input as well as the research team’s responses to the input received. The 60
campaign materials received a total of 132 inputs. Most inputs came from FGDs (78.4%); and most inputs were on the
campaign material images. The evidence-informed process of materials creation, use of multiple input sources, and a
broad-based iterative process allowed the creation of 60 evidence-based and culturally appropriate materials for a
breastfeeding social media campaign in Ghana. This paper could serve as a guide for other social media campaign efforts
looking to develop culturally appropriate materials.
Keywords
Social media, breastfeeding, promotion, Ghana, messages
Submission date: 4 September 2019; Acceptance date: 31 January 2020
Introduction Ghana, use of social media has increased among
Social media online communities that enable users to adults 18 years and older by an impressive 60%
7,8
create and share electronic content1 are becoming ubiq- between 2015 and 2017, (20–32% of all adults).
uitous among new mothers, especially those seeking Use is higher (43%) among young adults (18- to
information and support.2 Connecting on social 36-year-olds), but specific data on Ghanaian mothers
media can provide new mothers with external valida-
tion of their motherhood.3 Indeed, breastfeeding moth-
ers have reported using social media to make 1University of Ghana School of Public Health, Legon Accra, Ghana
2
connections, find information, and be entertained.4,5 Yale School of Public Health New Haven, Connecticut, USA
While the majority of research regarding social Corresponding author:
Richmond Aryeetey, University of Ghana School of Public Health, Akilagpa
media and parenting has been conducted in high-
Sawyer Rd, Legon, Accra, Ghana.
income countries, the rapid uptake of social media Email: raryeetey@ug.edu.gh
has been a global phenomenon.6 Specifically in Twitter: @rnokai
Creative Commons NonCommercial-NoDerivs CC BY-NC-ND: This article is distributed under the terms of the Creative Commons Attribution-
NonCommercial-NoDerivs 4.0 License (https://creativecommons.org/licenses/by-nc-nd/4.0/) which permits non-commercial use, reproduction
and distribution of the work as published without adaptation or alteration, without further permission provided the original work is attributed as specified on
the SAGE and Open Access pages (https://us.sagepub.com/en-us/nam/open-access-at-sage).
2 DIGITAL HEALTH
is not currently available. If the adoption of social been lauded as complementary to such interventions
media in Ghana continues at the rate suggested by with a theoretical underpinning, and has also been pro-
the Pew Research Center data, we would anticipate posed specifically for interventions that target user
market saturation by 2023.7,8 Therefore, it is important engagement.28–30 Despite existing social marketing
to identify how best to harness social media platforms and social media frameworks, there is a lack of evi-
for public health campaigns in contexts akin to Ghana. dence on the development of health-focused social
Social media-based health promotion campaigns media campaigns such as breastfeeding. Indeed, no
can have a direct and positive effect on behavior.9,10 examples of the development of such social media cam-
Over the last 15 years, social media has increasingly paigns in sub-Saharan Africa have been reported in
been used as a platform for health promotion, partic- literature.
ularly in high-income countries.10,11 However, little Therefore, to fill this gap in the literature in
research has been conducted on the effect of social sub-Saharan Africa, we present the development
media-based campaigns for protecting, promoting, and testing process of campaign materials for
and supporting breastfeeding.12 As a result, the 2016 Breastfeed4Ghana, a Facebook- and Twitter-based cam-
Lancet series on breastfeeding called for further paign designed to improve breastfeeding knowledge and
research on the role of social media in breastfeeding perceptions among Ghanaian adults.31 We detail the
promotion, noting the limited guidance on how to iterative, user-centered development of evidence-based
develop such campaigns.13,14 and culturally appropriate materials for this campaign
The benefits of optimal breastfeeding practices are that was grounded in the social ecological model,26 and
widely documented.13,15 Specifically, the World Health utilized inputs from multiple sources (Box 1). We also
Organization (WHO) recommendation of exclusive present descriptive results to outline this process.
breastfeeding (EBF) through the first 6 months,16,17 is
linked with child survival, optimal physical growth,
and cognitive development.15 Despite tremendous Materials and methods
progress in achieving optimal breastfeeding practices The study was implemented in an iterative stepwise
across Ghana, the rate of EBF has decreased without process. First, campaign materials composed of a text
clear explanation.18,19 In response, a committee of message accompanied by a complementary photograph
breastfeeding stakeholders in Ghana implemented the were designed (Box 1). Next, the campaign materials
Becoming Breastfeeding Friendly (BBF) initiative in were tested in focus group discussions (FGDs) and
2016, which resulted in recommendations and action their feedback was used to revise and improve the sim-
plans to scale up breastfeeding policies and programs plicity, cultural appropriateness, and acceptability of
nationally.20–22 Key recommendations that emerged the materials. In addition to the FGDs, inputs for revi-
included 1) fostering strong advocacy for breastfeed- sion were obtained from a diverse group of content and
ing, 2) harnessing support for maternity protection leg- technical experts including experts in maternal–child
islation, and 3) effectively disseminating accurate and nutrition and lactation (Breastfeed4Ghana advisory
actionable breastfeeding information to the public.20 group), the BBF committee, and the Food and Drugs
The committee of stakeholders who implemented Authority (FDA) (Box 1). The experts reviewed the
BBF proposed that a breastfeeding social media cam- materials for validity (e.g., confirming that the mes-
paign be designed as a low-cost and novel health com- sages were consistent with national and global recom-
munications intervention to help implement these mendations for breastfeeding), alignment (e.g.,
recommendations. confirming that messages corresponded with the
In response to these key evidence-informed recom- images used in terms of infant age), and clinical appro-
mendations, and the lack of evidence on how to devel- priateness (e.g., confirming that the images depicted
op a breastfeeding campaign via social media in this appropriate breastfeeding technique, including posi-
context, methods designing such a campaign were tioning and latch). All inputs were reviewed and incor-
needed. Such a design needed to integrate strong exist- porated, iteratively, to enhance both the acceptability
ing evidence from multiple disciplines. For instance,
and validity of the materials prior to finalization.
well-established frameworks for social marketing
Details of this study approach are provided here.
exist, such as the social marketing wheel.23,24
Traditional social marketing practices have been
Formative development of campaign material
applied for breastfeeding promotion,25 and the social
ecological model26 has repeatedly been applied in the The Breastfeed4Ghana social media-based communi-
context of breastfeeding protection, promotion, and cation campaign was grounded in the social ecological
support.27 A user-centered, person-based approach to model, a theory-based framework for understanding
designing interactive digital health interventions has complex and multifaced determinants of behavior
Aryeetey et al. 3
Box 1. Description of terminologies and acronyms
Terminology Description
Message Message is defined here as the recommendation or advocacy statement that
was presented in a text format. Sixty messages were developed for this
campaign. All messages are listed in Additional File 1.
Image Photos of persons, situations and, locations were included to complement the
messages. Each message was accompanied by a complementary image.
Campaign material Campaign material denotes a combination of a message and a complementary
image. A total of 60 campaign materials were designed for the campaign.
Input Feedback on the core campaign materials obtained from various sources
including end users, and content and technical experts. The process of
receiving input from different sources is summarized in Figure 1.
Content and technical experts These included nutrition and lactation experts (advisory group), BBF com-
mittee members, and the FDA.
BBF committee members BBF is acronym for Becoming Breastfeeding Friendly. The BBF committee is a
cross-sectoral committee of experts whose work resulted in recommenda-
tions for breastfeeding policy in Ghana as part of the BBF project.20 The BBF
committee was consulted for content and technical input.
Breastfeed4Ghana advisory group These were experts in breastfeeding and lactation who were consulted for
content and technical input.
FDA FDA is acronym for the Food and Drugs Authority in Ghana. The FDA is
mandated by law to approve breastfeeding messages intended for dis-
semination to the public. The FDA provided technical and content input as
part of its approval process.
across four levels that reflect personal and environmen- were designed based on existing Infant and Young Child
tal factors: intrapersonal, interpersonal, community, Feeding (IYCF) recommendations of the Ghana Health
and society levels.26 Thus, the campaign provided mes- Service, WHO, United Nations Children’s Fund, and
sages targeting not only breastfeeding mothers, but the Ghana BBF committee (Box 1).20,35,36
also their close family and associates, as well as The first step was the development of an initial set of
decision-makers who influence the broader breastfeed- 11 campaign messages (Figure 1). Complementary
ing environment. The overall campaign was designed to images were selected and used to create this first set
provide positive, simple, educational messages for of campaign materials (Box 1). These initial materials
breastfeeding protection, promotion, and support. were presented to mothers of young children (n¼ 19)
Positively oriented messages emphasize the benefits of who provided feedback in two community group dia-
breastfeeding, and have been considered more persua- logues prior to the formal launch of the research phase
sive than negatively oriented messages.32,33 Thus, all of the study. Based on this feedback, the materials were
messages were designed to be brief, easily understood, revised and submitted to the FDA for official approval.
culturally appropriate, and relevant for the target pop- Input from the mothers who participated in the com-
ulations of breastfeeding mothers, their close family munity dialogues and from the content and technical
and associates, and decision-makers who influence the experts guided the development and subsequent finali-
broader breastfeeding environment, so that users can zation of 49 additional materials, yielding a total of 60
easily apply them as a basis for behavior change.34 campaign materials.
Each material had a brief message that included The digital design of the campaign materials was led
between 7 and 31 words, paired with a single photo, by a digital creative agency with input from the study
to reinforce and complement the message. The messages investigators as well as the content and technical experts.
4 DIGITAL HEALTH
Inputs for core campaign materials Evolution of the core campaign materials
Initial inputs: 
BBF findings & recommendations
GHS IYCF communication tools
Community dialogs with mothers
FDA approval 
11 Initial materials 
BBF recommendations 
1
IYCF communication tools 49 Additional messages  
Testing and finalization inputs 
Breastfeed4Ghana advisory group 60 Core campaign materials2
FGDs (n = 6)
BBF committee 
60 Final core campaign materials 
FDA approval
Figure 1. Summary of the evolution of the Breastfeed4Ghana core campaign materials from multiple sources of input.
BBF¼Becoming Breastfeeding Friendly; BF4GH¼ Breastfeed4Ghana; FDA¼ Food and Drugs Authority; FGD¼ focus group discussion;
GHS¼Ghana Health Services; IYCF¼ Infant and Young Child Feeding.
1Lessons learned from the first 11 materials supported the development of the additional 49 materials.
2The 60 core campaign materials comprise the 11 initial plus the additional 49 materials.
The creative agency was responsible for creating appro- Women who participated in the FGDs were
priate, high-quality photographs to complement the recruited at selected health facilities or community-
campaign messages. based child welfare clinics (CWC) in Accra, Ghana
by the campaign coordinator, with assistance from
Testing for acceptability and cultural public health nurses at the various CWCs. The
appropriateness women were included if they had delivered and
breastfed a healthy infant in the preceding 2 years.
Subsequently, all 60 campaign materials underwent A total of 48 women participated in the 6 FGDs con-
a similar review and finalization process (Figure 1). ducted (Table 1).
Because the initial 11 materials had already received In each FGD, an average of 15 materials (range:
input from mothers in the target population, they 9–20) were displayed and discussed, one at a time,
were subsequently only reviewed again by the expert either electronically (using an LCD projector), or as a
advisory group. In addition, the images of the initial color-printed large poster (on A3-sized paper). The
materials were revised for cultural appropriateness, FGDs were facilitated by the campaign coordinator.
using original photos. The FGDs used for testing the Each FGD was audio recorded, and a notetaker was
materials aimed to assess understanding and cultural present both to record notes and to support the facili-
appropriateness of the materials in terms of 1) the mes- tator. The audio recordings from each FGD were tran-
sage, 2) the image, and 3) message and image align- scribed verbatim. Through an ongoing process of
ment, from the primary end-user perspective (i.e., material review, input received from FGDs and experts
social media users from the target audience). An inter- were used to revise the materials (Figure 1).
view guide was developed and pretested for use in con- Finalization of the materials was carried out by the
ducting the FGDs (see Additional File 2). study investigators working closely with the digital
Formative Testing &
Development
development finalizing
Aryeetey et al. 5
Table 1. Socio-demographic characteristics of focus group dis- responses. Thereafter, two additional investigators
cussion participants. independently reviewed and provided input. Review
and final decisions on inputs were achieved via discus-
All participants
(n¼ 48)1 sion and consensus between at least two investigators.
The results present a descriptive assessment of the
Characteristics results.
Age (years) 33 6
Results
Number of children 2 1 The initial stage of the analysis included recording the
 evolution of the materials during testing and finaliza-Age of youngest child (months) 17 20
tion. Table 2 provides examples of campaign materials
Education level completed and illustrates the process involved in organizing, sum-
marizing, and finalizing the materials (Additional File 3
None/primary school 6 (13) provides the full output).
Junior high/vocational 10 (21) Source of inputs
Secondary or higher 32 (67) The 60 core campaign materials were tested and final-
ized based on 132 inputs from four sources: FGDs,
Ever used social media 37 (77) Breastfeed4Ghana advisory group, BBF committee,
and the FDA (Figure 2). The majority (68%) of these
Social media platforms used
inputs came from FGDs, followed by inputs from the
WhatsApp 34 (71) Breastfeed4Ghana advisory group (28%). The median
number of different sources of input for a given mate-
Facebook 28 (58) rial was two (Table 3). Although two of the materials
that were previously tested and approved by the FDA
Note: 1Data are presented as mean SD or n (%). during formative development were retested in FGDs,
a total of nine materials that had been previously tested
creative agency, and harnessing all input from the and approved were not retested in FGDs. All 60 mate-
FGDs and the content and technical experts, in accor- rials were reviewed by the three content and technical
dance with the campaign design theory. Thereafter, expert groups. However, 22 of the materials did not
color-printed versions of the 60 materials were submit- receive any input from these sources; one of these
ted to the Ghana FDA for final approval, as legally had been previously tested and approved and was
required (Figure 1). therefore not retested in FGDs.
Data analysis Types of inputs
The quantitative and qualitative analytical approaches The majority of the input from FGDs (78.4%) and
described here, utilized inputs obtained during the test- content and technical expert groups (76.3%), was to
ing and finalization stages of the campaign material request changes to the images in the material.
evolution (Figure 1). This analysis involved a two- However, the nature of these inputs differed according
stage process. First, campaign materials were organized to source. In FGDs, image-related inputs were often to
into a database that allowed tracking of type, frequen- support the alignment of the material, ensure variety in
cy, and source of inputs (i.e., from FGD, IYCF images used, and to address the likability and aesthetics
experts, BBF committee, or FDA), as well as the deci- of an image, whereas experts’ inputs on the images
sions made based on each input by the investigators. were primarily regarding cultural appropriateness and
Types of input were classified on the three core ele- breastfeeding technique (e.g., positioning and latch)
ments of the materials: 1) image, 2) message, and 3) (Additional File 3).
alignment between image and message. Revisions Inputs were also provided on the messages, although
based on input were classified as 1) incorporating the at a lower rate (31.4% of material reviewed by FGDs
input, or 2) not incorporating input. The latter only and 36.8% reviewed by content and technical experts).
occurred with input from FGDs; the reasons for not Most of the message-related inputs consisted of word
incorporating inputs are reported in Additional File 3. choice, grammatical changes, or spelling. Three of the
One investigator conducted an initial classification and four (75%) FDA inputs were regarding the message,
assessment of sources and forms of input and and the only BBF committee input was on the content
6 DIGITAL HEALTH
Table 2. Examples of how inputs were used to modify campaign materials.
Input from BF4GH advisory
Material group, FDA, and BBF Changes made in response
number Original material Intermediary material Final material Input from FGD committee to inputs
A1.1_1 N/A N/A BF4GH: Replace image with BF4GH: Image was replaced to
culturally appropriate ver- be culturally appropriate
sion (Ghanaian images)
A1.2_2 FGD1: Image is positive and indi- BF4GH: Improve on image and FGD1 & BF4GH: Another image
cates joy and connection FGD comments by repre- was developed to better
between mother and baby senting “nation” and represent and align with
Disagreement with the message “family” in the image to the message
FGD6: Image is positive with better correspond with text FGD6: Background of new
happy and healthy family, image was made brighter
though father could be more
excited and smiling.
Background is dark and could be
Tested in FGD1 Tested in FGD6
brighter
Reviewed by BF4GH
A1.3_3 FGD1: Emphasis was put on the BF4GH: Revise image for var- FGD1: Image was revised to a
newborn in the image, and the iation in models smiling mother and young
mother is making eye contact boy based on FG comments
Disagreement on the positioning FGD6: No changes needed
for breastfeeding BF4GH: Image was revised for
Participant thinks the message is variation in models across
for mother and child subsequent material
Message is clear and agreeable numbers
Tested in FGD6 FGD6: Image and message were
Tested in FGD1
Reviewed by BF4GH understood and acceptable
A1.4_4 FGD1: Disagreement with the BF4GH: Message should be FGD1: Different image was
message that is unlike most revised to a maternal selected to address com-
women’s experience benefit that has strong ments on posture; howev-
Message and image are not evidence, i.e., protection er, the specific image
aligned against breast cancer suggestion strayed from
Man in image shows “support to a Image of just a woman and the core design of the
breastfeeding mother,” which her baby breastfeeding materials (i.e., 1 image
is a good thing with proper positioning that corresponds to the
Tested in FGD6 A better image could be a before and smiling message)
Tested in FGD1
Reviewed by BF4GH and after image showing effect BF4GH & FGD6: Message was
Reviewed by BF4GH of breastfeeding on weight revised to focus on breast
status cancer rather than weight
FGD6: Disagreement with the loss
message; it is contradictory to
most women’s experience
Image was acceptable
Generated some questions and
discussion about positioning
Note: BBF¼Becoming Breastfeeding Friendly; BF4GH¼Breastfeed4Ghana; FDA¼ Food and Drugs Authority; FGD¼ focus group discussion.
Aryeetey et al. 7
Examples of input and responses for selected
Sources of input (n = 132 inputs) materials
1% 3% As shown in Figure 3a, the message of the material was
to motivate caregivers to “put your baby to the breast
anywhere and anytime that she needs to feed.”
28% Although the FGDs considered the message accept-
able, they felt the image (on the left-hand side of the
panel) did not convey the concept of “anywhere,”
which was considered a critical aspect of the message.
Therefore, the image of the breastfeeding baby was
68%
encircled by icons of different locations to represent
this concept of breastfeeding “anywhere” (Figure 3a).
Figure 3b demonstrates a case in which FGD inputs
FGD BF4GH BBF FDA were not incorporated due to cultural sensitivities. In
this situation, the image was modified in response to
FGD input, which indicated it appeared “staged” and
Figure 2. Proportion of sources across the 132 inputs received for also demonstrated the wrong breast milk expression
the 60 core campaign materials during testing and finalization. technique. Thus, a revised image (on the right-hand
BBF¼Becoming Breastfeeding Friendly; BF4GH¼ Breastfeed4
¼ ¼ side of the panel) was used. However, although theGhana; FDA Food and Drugs Authority; FGD focus group
FGD suggested using an image of a breast pump, this
discussions
revision was not carried out because in Ghana, hand
expression is common practice.
of a message. One of the 60 messages was changed The one message that was changed completely
completely during the testing and finalization period during testing and finalization was that of material
(material no. A0104_04), and one message was simpli- no. A0104_04, which originally stated “Breastfeeding
fied and given a more positive framing: from is good for mothers. It helps them lose pregnancy
“Exclusively breastfed babies have lower risk of chron- weight.” This generated discussion and debate in
ic disease when they are adults” to “Exclusively FGDs about the women’s own experience with preg-
breastfed babies will become healthier adults” (material nancy weight loss and breastfeeding, and among the
no. A0401_16) (Additional File 3). Breastfeed4Ghana advisory group who agreed that
Few inputs were received from the Breastfeed4Ghana the evidence for this was not strong enough to promote
advisory group, BBF committee, or FDA on material as a benefit of breastfeeding. Therefore, the message
alignment (2.6%); this type of input was received for 15 was changed to: “Breastfeeding is good for mothers.
of the 51 materials reviewed by FGDs. Most of the It helps protect against breast cancer” (see Additional
FGD inputs on alignment were accompanied by sugges- File 3).
tions to change the image rather than the message.
Discussion
This paper presents the iterative, user-centered
Response to inputs
approach taken to develop and test materials for a
The study investigators made the final decision on breastfeeding social media campaign with the aim of
inputs from the various sources by working in collab- insuring that the materials would be culturally appro-
oration with the creative agency to incorporate the priate. Campaign materials were informed not only by
required changes into the materials. All inputs from the end user, but also with evidence and input from
the content and technical expert groups were incorpo- various experts and stakeholders. We found that the
rated directly into the materials, as were 73.3% of nature of the inputs from the target population in
inputs from FGDs. Of the 24 FGD inputs that were the FGDs versus the content from technical experts
not incorporated into the materials, seven were incon- (i.e., Breastfeed4Ghana advisory group, BBF commit-
sistent with the material format for the campaign, and tee, and the FDA) differed, providing justification for
four were not considered feasible due to resource or consulting both content and context experts alongside
time constraints. The rest were not considered signifi- the target population. The expert input was valuable
cant to require revisions, in line with the design theory for ensuring message validity, while end-user input was
of the campaign. focused on message clarity and acceptability.
8 DIGITAL HEALTH
Table 3. Summary of sources and types of input, and the responses to input on the 60 core campaign materials.
Source of input for 60 materials n %1
Different sources of input [median (IQR)] 2 (2,3)
Only FGD input received 21 35.0
Only BF4GH, BBF, or FDA input received 8 13.3
No FGD, BF4GH, BBF, or FDA input received 1 1.7
Type of input from FGD (% out of 51 materials)
Change image 40 78.4
Change message 16 31.4
Change image–message alignment 15 29.4
Type of Input from BF4GH advisory group (% out of 37 materials)
Change image 28 75.7
Change message 4 10.8
Change image–message alignment 5 13.5
Type of input from BBF committee (% for 1 of the materials)
Change image 0 0
Change message 1 100.0
Change image–message alignment 0 0
Type of input from FDA (% out of 4 materials)
Change image 1 25.0
Change message 4 100%
Change image–message alignment 0 0
Response to the 132 inputs n %1
BF4GH, BBF, or FDA input was addressed (% out of 42 inputs) 42 100.0
FGD input was addressed (% out of 90 inputs) 66 73.3
Reasons for not incorporating FGD inputs (% out of 24 inputs)
Disagreed with suggestion 10 11.1
Deviated from campaign or material design 7 7.8
Not feasible within time/resources 4 4.4
Other2 3 3.3
Note: BBF¼Becoming Breastfeeding Friendly; BF4GH¼Breastfeed4Ghana; FDA¼ Food and Drugs Authority; FGD¼ focus group discussion;
IQR¼ interquartile range.
1Percentage is out of 60 materials unless otherwise specified.
2“Other” reasons included alignment/consistency with broader campaign theory, and interest in retesting the materials for additional input.
Aryeetey et al. 9
(a)
FGD input: Image was well accepted
message and image are not aligned Response: Image was revised to include
a better image would represent location icons to depict "anywhere" 
locations that depict "anywhere" 
(b)
Response: Image was revised to correct
FGD input: Image shows wrong manual expression technique. despite the suggestion
expression technique, and appears ‘staged' for including a breastmilk pump, we chose to
a better image should be of a breast pump keep hand expression as the image because this
is more common in Ghana. 
Figure 3. Image demonstrating range of responses to input from a focus group discussion (FGD) toward finalizing the Breastfeed4Ghana
campaign materials. Panel A demonstrates an example in which the FGD input was to address poor message–image alignment. Panel B
demonstrates an example in which some of the FGD input, specifically to include an image of a breast pump, was not utilized out of
respect for local context practice.
It has previously been demonstrated that the mes- misalignment can create a vague understanding of
sages from communication interventions have greater what is being communicated to end users, thereby lead-
cultural acceptability if appropriate needs assessment ing to misperceptions of the message.40
have been designed to understand the situation and Following its design, the campaign was run on
needs of the target audience.37 Such acceptability test- Facebook and Twitter for a 6-month period. The
ing is a key element of designing health interventions.38 results are reported in a forthcoming publication
A user-centered approach aligns with an emphasis on (Maternal and Child Nutrition Journal), and indicate
acceptability, and such an approach is particularly acceptability of the campaign and engagement with
effective for interventions targeting user engagement.28 the material content among both men and women in
Given that, in the context of social media, user engage- Ghana.31
ment is a key performance metric for campaigns,39 the
choice of user-centered design was justified for
Lessons learned
Breastfeed4Ghana as well as other health promotion
campaigns. In the process of designing the materials, the research
The visual elements (i.e., images) of the materials team identified a number of practical lessons that can
designed in the current study constituted an important serve as a useful guide to those who will be developing
area of feedback from the users. Image–message materials for campaigns similar to Breastfeed4Ghana.
10 DIGITAL HEALTH
First, sufficient time must be allocated to material and engagement with breastfeeding stakeholders in
development and testing. Approximately 5 months Ghana. The development stage of such a campaign
were allocated for developing the 60 materials for this should emphasize cultural acceptability and relevance,
campaign, which proved sufficient. However, adequate and aim to achieve acceptability among the target pop-
time between each FGD must be allowed to review the ulation of the campaign. This can be achieved by seek-
proposed changes and revise the materials, especially if ing input from a variety of sources with unique
FGD participants request new images that will need to perspectives on the content, culture, dissemination
be sourced and approved by the research team. The platforms, including end users.
duration required for campaign material development
and testing will, however, depend on the number, con- Acknowledgements: The authors appreciate the contribution of
text, and nature of the materials being developed. all focus group discussion participants for sharing their opinions on
Another important lesson from this study is the materials. We are also grateful to the Ghana BBF committee and
acknowledging that the images are an essential part other experts (including Gabriela Buccini) who provided input for
of material development. Stock photos were initially developing and testing the materials. We are also appreciative of
Elfortify David Sarfo and John Mawuli Nyadegbe at IZSCOPE
used before recognizing they were not culturally accept-
Ltd for the digital designing of the materials. The following research
able. Thus, only original images were ultimately used to and administrative assistants were critical for the design and testing
ensure cultural acceptability and campaign branding. of the Breastfeed4Ghana SM communication materials and are duly
Furthermore, it was originally planned to utilize the acknowledged: Abena Engmann, Faustina Aryeetey, Helena Bentil,
same images across multiple messages. However, Vivian Kapio Abem, Cara Safon, Augustina Boadu, and Katie
FGD feedback revealed that the recycling of images Doucet.
was poorly received. Across all inputs from all source
types, we found the image to be the element that Contributorship: The study was conceived and designed by
required the greatest attention and time for revisions. RA, KH, RPE, and AHF. Data gathering was led by RA
with contributions from KH, RPE, OL, and GC. RA and
KH led the data analysis and interpretation, with contribu-
Limitations
tions and final approval from all other authors.
Despite the value that this study adds to the literature,
there are some important limitations to these findings. Conflict of interest: The authors declared no potential con-
The end-user input we sought in this process was lim- flicts of interest with respect to the research, authorship, and/
ited to mothers in Ghana, despite the campaign’s broad or publication of this article.
approach to provide messages for family, friends,
employers, and co-workers to support and protect Ethical approval: Ethical approval of this study was granted
breastfeeding. Although input was sought from a by the Institutional Review Boards at University of Ghana
diverse group of experts, input from civil society that and Yale University. All FGD participants provided written
could have provided additional value was not received. informed consent to participate. In all cases, appropriate con-
Finally, we have not presented the results on campaign sent was obtained by the creative agency from the persons
engagement in this paper, but rather focused on the whose photographs were taken and/or used in preparation of
acceptability of the content. However, such results on the campaign materials. Permission was obtained from insti-
campaign engagement and content acceptability during tutions, including hospitals in Accra, where the photographs
the active campaign period have since been published, were taken by the creative agency.
which confirm that the campaign material was accept-
able and the target population (i.e., breastfeeding Funding: The authors disclosed receipt of the following finan-
mothers, family members and associates, and cial support for the research, authorship, and/or publication
decision-makers who influence the broader breastfeed- of this article: The Breastfeed4Ghana study was funded by
ing environment) did engage with it.31 the International Society for Research in Human Milk and
Lactation through Yale University (PI. Kassandra Harding),
subaward number GR101839 (CON-80001098).
Conclusions
This study demonstrated an iterative, user-centered, Guarantor: RA is guarantor.
evidenced-based approach to designing a culturally
appropriate and relevant breastfeeding social media Peer review: This manuscript was reviewed by reviewers who
campaign in Ghana that was grounded in the social have chosen to remain anonymous.
ecological model. Starting with a foundation of the
context-specific needs is essential. In the current ORCID iD: R Aryeetey https://orcid.org/0000-0003-4667-
study, this evidence came from the BBF assessment, 592X
Aryeetey et al. 11
Supplemental material: Supplemental material for this arti- 15. Victora CG, Bahl R, Barros AJD, et al. Breastfeeding in
cle is available online. the 21st century: epidemiology, mechanisms, and lifelong
effect. Lancet 2016; 387: 475–490.
16. Butte N, Lopez-Alarcon MG and Garza C. Nutrient ade-
References
quacy of exclusive breastfeeding for the term infant during
1. Tufts University. Social media overview, https://commu the first six months of life. Geneva: World Health
nications.tufts.edu/marketing-and-branding/social- Organization, 2002.
media-overview/ (2019, accessed 3 May 2019). 17. World health Organization. Global strategy on infant and
2. Regan S and Brown A. Experiences of online breastfeed- young child feeding. Geneva: World Health Organization,
ing support: support and reassurance versus judgement 2003.
and misinformation. Matern Child Nutr 2019; 15: e12874. 18. Ghana Statistical Service (GSS), Ghana Health Service
3. Schoppe-Sullivan SJ, Yavorsky JE, Bartholomew MK, (GHS), ICF International. Demographic and health
et al. Doing gender online: new mothers’ psychological survey 2014. Rockville, MA: GSS, GHS, ICF
characteristics, Facebook use, and depressive symptoms. International, 2015.
Sex Roles 2017; 76: 276–289. 19. Ghana Statistical Service (GSS). Multiple indicator clus-
4. Tomfohrde OJ and Reinke JS. Breastfeeding mothers’ ter survey (MICS 2017/2018): survey findings report.
use of technology while breastfeeding. Comput Human Accra, Ghana: GSS, 2018.
Behav 2016; 64: 556–561. 20. Aryeetey R, Hromi-Fiedler A, Adu-Afarwuah S, et al.
5. Asiodu IV, Waters CM, Dailey DE, et al. Breastfeeding Pilot testing of the Becoming Breastfeeding Friendly tool-
and use of social media among first-time African box in Ghana. Int Breastfeed J 2018; 13: 30.
American mothers. J Obstet Gynecol Neonatal Nurs 21. Buccini G, Harding KL and Hromi-Fiedler A. How does
2015; 44: 268–278. “Becoming Breastfeeding Friendly” work? A program
6. StatCounter. StatsCounter GlobalStats, https://gs.stat impact pathways analysis. Matern Child Nutr 2018; 15:
counter.com/social-media-stats (2019, accessed 23 e12766.
November 2019). 22. Hromi-Fiedler AJ, Dos Santos Buccini G, Gubert MB,
7. Poushter J, Bishop C and Chwe H. Social media use con- et al. Development and pretesting of “Becoming
tinues to rise in developing countries, but plateaus across Breastfeeding Friendly”: empowering governments for
developed ones: digital divides remain, both within and global scaling up of breastfeeding programmes. Matern
across countries. Washington, D.C.: Pew Research Center, Child Nutr 2018; 15: e12659.
https://www.pewglobal.org/2018/06/19/social-media-use- 23. Evans E. How social marketing works in health care. Br
continues-to-rise-in-developing-countries-but-plateaus- Med J 2006; 332: 1207–1210.
across-developed-ones/ (2018, accessed 29 March 2019). 24. National Cancer Institute. Making health communication
8. Pousther J. Smartphone ownership and internet usage programs work: a planner’s guide. Bethesda, MD: NCI,
continues to climb in emerging economies: but advanced 2002.
economies still have higher rates of technology use. 25. P
erez-Escamilla R. Breastfeeding social marketing: les-
Washington, D.C.: Pew Research Center, http://www. sons learned from USDA’s “Loving Support” campaign.
pewglobal.org/2016/02/22/smartphone-ownership-and- Breastfeeding Med 2012; 7: 358–363.
internet-usage-continues-to-climb-in-emerging-econo 26. Bronfenbrenner U. Toward an experimental ecology of
mies/ (2016, accessed 29 March 2019). human development. Am Psychol 1977; 32: 513–531.
9. Wakefield MA, Loken B and Hornik RC. Use of mass 27. Chopel A, Soto D, Joiner B, et al. Multilevel factors
media campaigns to change health behaviour. Lancet influencing young mothers’ breastfeeding: a qualitative
2010; 376: 1261–1271. CBPR study. J Hum Lact 2019; 35: 301–317.
10. Giustini D, Ali SM, Fraser M, et al. Effective uses of 28. Yardley L, Morrison L, Bradbury K, et al. The person-
social media in public health and medicine: a systematic based approach to intervention development: application
review of systematic reviews. Online J Public Health to digital health-related behavior change interventions.
Inform 2018; 10: e215–e. J Med Internet Res 2015; 17: e30.
11. Laranjo L, Arguel A, Neves AL, et al. The influence of 29. Band R, Bradbury K, Morton K, et al. Intervention plan-
social networking sites on health behavior change: a sys- ning for a digital intervention for self-management of
tematic review and meta-analysis. J Am Med Inform hypertension: a theory-, evidence- and person-based
Assoc 2015; 22: 243–256. approach. Implement Sci 2017; 12: 25.
12. Kahn JG, Yang JS and Kahn JS. Mobile health needs 30. Blandford A, Gibbs J, Newhouse N, et al. Seven
and opportunities in developing countries. Health Affairs lessons for interdisciplinary research on interactive digital
2010; 29: 252–258. health interventions. Digit Health 2018; 4: 2055207618
13. Rollins NC, Bhandari N, Hajeebhoy N, et al. Why invest, 770325.
and what it will take to improve breastfeeding practices? 31. Harding K, P
erez-Escamilla R, Carroll G, et al.
Lancet 2016; 387: 491–504. Four dissemination pathways for a social media-based
14. Abrahams SW. Milk and social media: online communi- breastfeeding campaign: evaluation of the impact on
ties and the International Code of Marketing of Breast- key performance indicators. J Med Internet Res Nurs
milk Substitutes. J Hum Lact 2012; 28: 400–406. 2019; 2: e14589.
12 DIGITAL HEALTH
32. Bartels RD, Kelly KM and Rothman AJ. Moving 37. Petty RE and Cacioppo JT. Communication and persua-
beyond the function of the health behaviour: the effect sion: central and peripheral routes to attitude change.
of message frame on behavioural decision-making. New York: Springer-Verlag, 1986.
Psychol Health 2010; 25: 821–838. 38. Sekhon M, Cartwright M and Francis JJ. Acceptability
33. Gallagher KM, Updegraff JA. Health message framing of health care interventions: a theoretical framework and
effects on attitudes, intentions, and behavior: a meta- proposed research agenda. Br J Health Psychol 2018; 23:
analytic review. Ann Behav Med 2012; 43: 101–116. 519–531.
34. Zarcadoolas C. The simplicity complex: exploring simpli- 39. Neiger BL, Thackeray R, Van Wagenen SA, et al. Use of
fied health messages in a complex world. Health Promot social media in health promotion: purposes, key perfor-
Int 2011; 26: 338–350. mance indicators, and evaluation metrics. Health Promot
35. Ghana Health Service. National Infant and Young Child Pract 2012; 13: 159–164.
Feeding for Ghana: strategy document. Accra: GHS, 2007. 40. Hewitt D. Conceptual complexity, environmental com-
36. Ghana Health Service (GHS), United Nations Children’s plexity, communication salience and attitude change
Fund (UNICEF). Key messages booklet: The community Eur J Soc Psychol 1972;2:285–306.
Infant and Young Child Feeding counselling package.
Accra, Ghana: GHS, 2012.