Co-Designing M-Healer: Supporting Lay
Practitioner Mental Health Workers
in Ghana
Liam Albright1(B), Hoa Le1(B), Suzanne Meller2(B), Angela Ofori Atta3(B),
Dzifa A. Attah3(B), Seth M. Asafo3(B), Pamela Y. Collins2(B),
Dror Ben Zeev2(B), and Jaime Snyder1(B)
1 Information School, University of Washington, Seattle, WA, USA
{liamaw,lnqhoa,jas1208}@uw.edu
2 School of Medicine, Department of Psychiatry,
University of Washington, Seattle, WA, USA
{smeller,pyc1,dbenzeev}@uw.edu
3 University of Ghana, Accra, Ghana
{daattah,smasafo}@ug.edu.gh
Abstract. Mental health is a vast problem around the globe and is one
of the key population health issues in the world today. At any given time,
up to 6.8% of the world’s population suffers from a serious mental illness
(SMI) such as schizophrenia or bipolar disorder. The impacts of SMI on
a population are especially challenging in low and middle-income coun-
tries (LMIC). Mobile healthcare application research is a growing area
of research aiming to ameliorate these challenging impacts. In Ghana,
a LMIC in West Africa, mental healthcare systems are severely under-
resourced and people with SMI often receive care from lay practitioners
such as traditional and faith healers rather than trained mental health
clinicians. These challenges exist alongside developed wireless infrastruc-
ture. In these contexts, mobile applications can substantially increase
access to health information. This is the basis for our work develop-
ing a mobile health (mHealth) application to support mental health lay
practitioners in Ghana. We describe the ways that our principled design
research practice is intersecting with local faith-based practices, vernac-
ular expertise and values, and the practicalities of technology adoption
in Ghana.
Keywords: Mobile health · Mental health · Ghana
1 Introduction
By leveraging wireless infrastructure and available technologies, mobile applica-
tions can substantially increase access to health information, especially in under-
resourced communities [7,19]. In particular, in many low- and middle-income
countries (LMIC), challenges related to health care equity exist alongside strong
©c Springer Nature Switzerland AG 2021
C. Stephanidis et al. (Eds.): HCII 2021, CCIS 1499, pp. 115–124, 2021.
https://doi.org/10.1007/978-3-030-90179-0_16
116 L. Albright et al.
wireless infrastructure. HCI work in the area of information and communica-
tion in developing contexts (ICT4D) advocates for (1) the critical importance
of establishing strong collaborations with local stakeholders and (2) avoiding a
“one size fits all” approach to technology design [26]. In describing our ongoing
work designing and implementing a mobile app to support non-clinically trained
spiritual healers in Ghana, we highlight ways that our principled design research
practice intersected with local faith-based practices, vernacular expertise and
values, and the practicalities of technology adoption in West Africa.
Mental illness is a significant global health problem. At any given time, up to
6.8% of the world’s population suffers from a serious mental illness (SMI) such
as schizophrenia or bipolar disorder [14]. These psychiatric conditions result in
particularly devastating long-term impacts in LMIC [29]. In West Africa, hard-
ships associated with SMI are compounded by pervasive societal stigma, scarce
treatment options, and systematic neglect and abuse [20,21,23]. West African
mental healthcare systems are severely under-resourced leading to substantial
unmet mental health needs [10,12,13,25]. People with SMI often receive care
from traditional and faith healers rather than trained mental health clinicians
[11,18,22]. Previous research in Ghana found that stakeholders from all sec-
tors (patients, providers, government officials, and faith healers) were open to
exploring mobile health approaches to improve the clinical outcomes of those in
need [3].
To address this gap, we are developing a smartphone intervention, M-Healer,
to provide local healers with educational content and practical tutorials for tech-
niques such as de-escalation, deep breathing, and re-framing anxious thoughts.
Our goals are two-fold: (1) to reduce human rights violations commonly associ-
ated with the care of people with SMI in LMIC [8,28] by providing information
about more humane alternatives and (2) to evaluate the impacts of a mobile app
on treatment outcomes for individuals with SMI in West Africa.
2 Related Work
Health information and interventions delivered via mobile systems (mHealth)
have the ability to improve mental health-care outcomes [5,9,15]. A systemic
review and meta-analysis of randomized controlled trials that used smartphone
applications to treat symptoms of anxiety found that the most successful pro-
grams combined mobile content with the delivery of face-to-face interventions
[9]. A team of researchers responding to a large treatment gap in mental health
in Zimbabwe (approximately 10 psychiatrists for a population of 13 million)
investigated if lay health workers delivering psychological interventions using a
form of cognitive-behavioral therapy (CBT) could improve symptoms of depres-
sion and anxiety. Participants were taught a structured approach to identifying
problems and finding workable solutions through 4 individual sessions, first with
lay health workers then in peer-led groups. The study concluded that primary
care administered by lay health worker combined with education and support
resulted in improved symptoms at 6 months [6].
Co-Designing M-Healer: Supporting Lay Practitioner Mental Health Workers 117
Our approach involves local stakeholders in the app design process. Related
work uses similar co-design methodology. Two case studies in Mozambique and in
Romania [27] document procedures, successes, and hurdles of community-based
participatory design. In Romania, hierarchy in social interaction and decision-
making led to some participants not feeling empowered to engage fully in the
design process [27]. In contrast to the intended impact, participatory and co-
design methods highlighted lack of agency experienced by people the process was
meant to empower. However, a co-design study in rural South Africa focused on
a billing system for voice services provided by a community network shows that
the co-design process can result in an intervention that is usable and integrated
with local culture [24]. Another group of researchers used co-design methods
to enhance communication between mothers and Neonatal Intensive Care Unit
(NICU) staff in South Africa [16]. While researchers noted that co-design sessions
seemed to make participants feel more comfortable about contributing substan-
tively, they also observed that junior nurses did not feel free to critique ideas
of more senior members of the healthcare system [16]. In response, researchers
introduced a card-sorting method to encourage group cohesion.
3 Methodology
During fall 2019, members of our team traveled to Ghana to (1) learn more about
the context in which our intervention would be implemented and (2) conduct
co-design sessions with local stakeholders. As a result, we were able to compile
a set of system, interface, and content requirements for a fully functional pro-
totype. User feedback sessions with target users were conducted during summer
2020. User feedback sessions were conducted on-site by collaborators from the
University of Ghana, in accordance with WHO recommended COVID-19 pre-
ventative practices. This work was approved by the Institutional Review Boards
of the University of Washington and the University of Ghana.
3.1 Design Research and Requirements Gathering
Prior to field work, We conducted survey of digital infrastructure requirements,
focusing on the likelihood of certain technologies being readily used, data and
storage constraints, and other factors that would constrain the future application
development. Our team also completed a series of design research activities aimed
at better understanding Ghanaian culture, care and treatment of people with
mental health issues in West Africa, common practices of faith healers that
provide care to individuals with SMI, and the nature of relationships among
clinicians, religious leaders, and policymakers. Interview protocols and elicitation
materials drew on the team’s previous work in Ghana [3], mHealth literature,
and user-centered technology design best practices.
In Ghana, interviews and verbal prototyping sessions with 18 faith healers
took place at three prayer camps located just outside the capital of Accra. Field
118 L. Albright et al.
work culminated with a half day co-design focus group with 12 healers repre-
senting each of the three camps. During these activities, the research team met
daily to discuss findings and to reflect on qualitative and design methods as
they unfolded. These discussions were documented in field notes and serve as
the primary source of data for the emergent themes we highlight here.
3.2 System Design
In response to design research insights, emergent co-design concepts, and require-
ments gathered on site, we developed the M-Healer prototype to be a stable and
self-contained Android platform. The android platform was chosen because it is
estimated to have 84% of the mobile operating system market share in Ghana
[17]. Due to low broadband support and costs associated with data plans in West
Africa, the application was built to be self-contained with content delivered via
optimized and light-weight animated videos downloaded with install files. This
means that M-Healer has no backend functionality, API integration, or database
connections. The user interface was also designed dark mode first as dark mode
has shown to be more efficient in battery consumption [1]. See Fig. 1.
Fig. 1. Functional prototype
The application was written in Kotlin, the official android platform develop-
ment language. A benefit of Kotlin, as opposed to Flutter or React Native, is
that it is optimized for performance across all android phones. It scores better
on performance metrics such as CPU usage, memory use, battery efficiency, all
vital for an app operating in environments with limited access to power and data
connectivity. Additionally, Kotlin typically leads to a codebase that is more com-
pact and concise than traditional java android applications [2]. Kotlin has also
shown the ability to reduce the frequency of Null Pointer Exceptions [2] which
is often a hurdle in Java Android development. The application’s user interface
design was built using a modified version of Google’s Material Design language.
The app uses Material Design system icons and square buttons in a grid-like
layout with flat lateral navigation flows. The main navigation buttons are also
Co-Designing M-Healer: Supporting Lay Practitioner Mental Health Workers 119
relatively larger than the traditional Material Design button size with the six of
them each evenly taking up 16.67% of the screen for ease of clicking.
The M-Healer application itself consists of a home screen where users have
the option to click on six buttons that each correspond to a content bucket. Con-
tent areas, established in collaboration with stakeholders, include de-escalation
and relaxation techniques, tutorials on alternatives to problematic practices such
as physical restraints and forced fasting, and general education regarding human
rights in the care of people with SMI. Original content was aligned with, but did
not explicitly refer to, religious messaging and language discussed during inter-
views and co-design sessions. From content landing pages, users can choose from
a list of videos, accessed through a video player integrated in the application.
Users also have the option to switch video language between Twi (a common
spoken language in the southern region of Ghana) and English.
Documentation, including (1) key path information architecture, (2) style
guide for implementing preliminary M-Healer branding, (3) examples of ani-
mated video and text content, and (4) specifications for loading the free-standing
app onto smart phones, was written for non-technical audiences and tested with
clinical research colleagues at the University of Ghana.
(a) Community/Medical (b) GhanaianCulture
(c) Nature (d) ExplicitlyReligious
Fig. 2. Four visual motifs
4 Preliminary User Feedback
An initial set of user feedback findings emerged from our fieldwork. While still
in Ghana, we created a series of five visual design directions for the look and
120 L. Albright et al.
feel of the app, ranging from highly medicalized to explicitly religious, as shown
in Figs. [2a, 2b, 2c, 2d]. The explicitly Christian imagery (e.g., a bible, a dove,
and a wooden cross) was very popular with focus group participants. However,
several of the faith healers argued that the design concept that drew on images
of nature (e.g., water droplets on grass blades, sun beams bursting from behind
clouds, glowing candle flames) would be the most accessible option for prayer
camp leaders of different faiths, such as Islam and traditional religions, who are
also intended users of the app.
Once we had a functional prototype of the app, our colleagues in Ghana
took it to our primary stakeholders to gather feedback. The protocol focused
on assessing the interface and content for usability issues. Several users men-
tioned being excited about the application and could envision themselves using
the application on a daily basis. Users were mostly able to navigate through
the application, access the video content, and summarize the video content [4].
However, the testing sessions took place on phones not owned by the partici-
pants which caused minor hurdles for some. The video content was found to be
broadly acceptable and useful for the participants. Participant feedback on how
to improve the design of the application was focused on the meaning of icons and
the dark background. The previously cited paper also found some of the partici-
pants had trouble understand the connection between specific icons and content
areas. Additionally, one participant mentioned that the main application icon
was blurry. Some participants also mentioned the colors in the application being
a bit dark, noting that a brighter interface could be easier to see for users with
vision problems. Finally, when participants were asked about future features in
the application, they mentioned wanting more spoken voice cues describing dif-
ferent modules of the application. Participants described future iterations of the
application that might include these features as intriguing and exciting.
4.1 Influence of Technical Constraints on Design Process
One of the major technical constraints on our application development process
was the need to balance infrastructure requirements with design of an engag-
ing user experience. For complex interventions that require having access to
dynamic data or more extensive software architecture designs, balancing these
types of design requirements can be challenging. This project forced our mobile
app development team to rethink assumptions, both implicit and explicit, about
the ubiquity of reliable data connections and power sources. Thinking about
workflows and processes to easily incorporate people and environments that do
not have access to traditional development tools is valuable for future endeavors.
In the case of M-Healer, the environment in which our application operates
lacks consistent access to a mobile grid and power. Therefore, content was inte-
grated into the application itself and loaded at startup. Because we identified
this limitation early in our development process, we were proactive about con-
sidering implications in all aspects of our design. Application file size became a
primary design requirement throughout the development process, playing a role
in decisions from color to length of scripted animations.
Co-Designing M-Healer: Supporting Lay Practitioner Mental Health Workers 121
Building the application with these design constraints also meant having
to go without some of the usual tools associated with Android development.
With no backend to the project, the application had to be manually loaded
onto phones. Without typical manual processes used with Google Developer
Console tools, the application was shared in APK form and installed directly on
devices by colleagues in Ghana. Tracking usage data such as clicks without using
Google Analytics or another user data tracking API was also a challenge. For
applications that operate in resource-constrained environments like ours, devis-
ing extensible tracking methods that do not require connectivity is particularly
challenging. This continues to be a major focus of our work as we prepare to roll
out a large-scale field trial of the M-Healer intervention.
4.2 Code-Sharing and Workflow
In future stages of the app development, we will prioritize introducing concrete
code-sharing plans and workflows that are both accessible to non-technical team
members and in line with traditional software development tools (i.e., Git Hub).
Two factors made it necessary for the code base to be more available to the full
team than perhaps is typical for app development projects like this: (1) close
coordination across design activities (i.e., platform, interface, brand/identity,
content) resulted in an iterative process involving multiple cycles of code updates
and evaluation by the team and (2) user feedback sessions run by local collabora-
tors who are trained as clinicians rather than technologists made it necessary to
package a preliminary release of the app in a format that was easy to extract and
install. From an application development standpoint, a key hurdle was sharing
code. With only one member of our team experienced in the Git environment,
a lot of code was shared over email, making version control very challenging.
Gathering and responding to software bug fixes also presented challenges.
The team members tasked with installing the preliminary version of the appli-
cation on phones in Ghana mainly used WhatsApp for communication among
themselves and with other international team members. This meant issues, fixes,
and eventually updated software were frequently shared over WhatsApp. During
this process, we noticed that as a chat and messaging platform WhatsApp was
not quite as optimized for developer communication as is a platform like Slack.
For instance, Slack has a feature called Share Code Snippets that allows users to
share code snippets and have the code appear very similar if not the same as it
does in the development environment it was copied from. An interesting solution
for this issue would be for WhatsApp to add more features to support software
developers and create an identity for itself as the communicating platform of
choice for international software development.
4.3 Innovation and Broadening Impacts of User Centered Design
Initial user feedback was elicited and gathered by Ghanaian members of our
team who are clinical psychologists familiar with the contexts of the prayer
camps but with limited exposure to user centered design practice. Members of
122 L. Albright et al.
our team trained in usability studies attempted to craft a protocol that was clear,
concise, and easily understood by both the clinicians eliciting feedback and the
participants providing it. While these questions did provide some insights into
what was working and what was not in the preliminary version of the app, we
recognize that mastering techniques for follow-up questions and clarifications
can be challenging, even for very experienced designers. Additionally, questions
were written in English and translated to Twi in the moment. For the most part,
responses were given in Twi, and were then translated by researchers, sometimes
at that time and sometimes later, when notes were compiled on standardized
data sheets to send to colleagues in the U.S. As we have reported, participants
in this initial round of user feedback interviews found the application usable and
feasible. We are confident that this feedback is reliable, while at the same time
looking forward to larger scale field tests where we will have the opportunity
to refine our approach so that participants can give us more pointed feedback
regarding alternatives to navigation, graphic design, and visual identity.
For example, we used a modified fusion of Microsoft’s Metro and Google
Material Design. While a current design trend in mobile development is bottom
navigation using a carousel of selectable icons, our application differs quite a bit
from this trend and uses large evenly space squares for navigation through the
main content. Future testing could compare both navigation methods in terms
the accessibility issues most commonly experienced in this content. Further,
future rounds of user feedback will include training in user centered design to
local colleagues, and ideally direct observations of the app being used in situ.
From a software engineering standpoint, this project can be used as an example
of thinking about innovation and accessibility in different ways. To make this
intervention feasible it required thinking about innovation from the standpoint of
resource-constrained environments, users with low literacy and experiences with
technology distinct from Western user groups . In this context, our innovation
focused on permanence, longevity, and cultural competence rather than speed,
accuracy, or monetization. This app development project has been a platform
for our design team to explore the ways that constraints can enable us to shift
the focus of innovation from “the next new feature” to broadening the impacts
of user centered design insights.
5 Conclusion
Our methods were selected for their capacity to (1) avoid a “one size fits all”
approach to mHealth technology design and (2) support strong collaborations
with local stakeholders. However, we also recognize that our work exists and
relies on the global mobile telecommunications industry, including its endemic
inequities. Preliminary user feedback and emergent themes resulting from our
design research approach signal potential for this work to contribute a method-
ological resource for HCI researchers working in the areas of ICT4D, mHealth for
mental health, and supporting personal informatics for vulnerable populations.
We offer them here as a signal of the value we see in case studies of DEI efforts.
Co-Designing M-Healer: Supporting Lay Practitioner Mental Health Workers 123
References
1. URL. https://www.cnet.com/news/using-androids-dark-mode-improves-battery-
life-google-confirms-p/
2. Ardito, L., et al.: Effectiveness of Kotlin vs. Java in android app development
Tasks.. Inf. Softw. Technol. 127, 106374 (2020). https://doi.org/10.1016/j.infsof.
2020.106374.
3. Ben-Zeev, D.: Mobile health for mental health in west Africa: the case for
Ghana. Psychiatr. Serv. 69(7), 741–743 (2018). https://doi.org/10.1176/appi.ps.
201700555.
4. Ben-Zeev, D., et al.: Development of M-Healer: a digital toolkit to improve care
and reduce human rights abuses against people with mental illness in West Africa.
JMIR Mental Health 8(7), e28526 (2021)
5. Ben-Zeev, D., et al.: Feasibility, acceptability, and preliminary efficacy of a smart-
phone intervention for schizophrenia. Schizophrenia Bull. 40(6), 1244–1253 (2014)
6. Chibanda, D., Weiss, H.A., Verhey, R.: Effect of a primary care–based psycho-
logical intervention on symptoms of common mental disorders in Zimbabwe: a
randomized clinical trial. JAMA 316(24), 2618–2626 (2016). https://doi.org/10.
1001/jama.2016.19102
7. Donker, T., et al.: Smartphones for smarter delivery of mental health programs:
a systematic review. J. Med. Internet Res. 15(11), 247 (2013). https://doi.org/10.
2196/jmir.2791
8. Esan, O., et al.: A survey of traditional and faith healers providing mental health
care in three sub-Saharan African countries. Soc. Psychiatry Psychiatr. Epidemiol.
54, 395–403 (2019). https://doi.org/10.1007/s00127-018-1630-y, PMID: 30456425
9. Firth, J., et al.: Can smartphone mental health interventions reduce symptoms of
anxiety? A meta-analysis of randomized controlled trials. J. Affective Disord. 218,
15–22 (2017). https://doi.org/10.1016/j.jad.2017.04.046
10. Gureje, O., Lasebikan, V.: Use of mental health services in a developing country:
results from the Nigerian survey of mental health and well-being. Soc. Psychiatr.
Psychiatr. Epidemiol. 41(1), 44–49 (2006)
11. Gureje, O., et al.: The role of global traditional and complementary systems of
medicine in the treatment of mental health disorders. Lancet Psychiatr. 2(2), 168–
77 (2015)
12. Samba, D.: Ibadan, Nigeria, University College Hospital, Department of Psychia-
try, Mental Health Leadership and Advocacy Programme (mhLAP). http://www.
mhlap.org/country-of-activities/gambia (2012)
13. Johnson, K., et al.: Association of combatant status and sexual violence with health
and mental health outcomes in postconflict Liberia. JAMA 300(6), 676–690 (2008)
14. Kessler, R.C., et al.: The global burden of mental disorders: an update from the
WHO World Mental Health (WMH) surveys. Epidemiol. Psychiatr. Sci. 18(1),
23–33 (2009)
15. Kola, L., Abiona, D., Adefolarin, A., Ben-Zeev, D.: Mobile phone use and accept-
ability for the delivery of mental health information among perinatal adolescents
in Nigeria: survey study JMIR Ment. Health 8(1) (2021). https://doi.org/10.2196/
20314
16. Mburu, C.W., et al.: Co-designing with mothers and neonatal unit staff: use of
technology to support mothers of preterm infants. In: Proceedings of the Sec-
ond African Conference for Human Computer Interaction: Thriving Communities
(AfriCHI 2018). Association for Computing Machinery, vol. Article 12. New York,
NY, USA, pp. 1–10 (2018). https://doi.org/10.1145/3283458.3283487
124 L. Albright et al.
17. Mobile Operating System Market Share Ghana. Accessed 25 Nov 2020. https://
gs.statcounter.com/os-market-share/mobile/ghana
18. Ae-Ngibise, K., et al.: Whether you like it or not people with mental problems are
going to go to them: a qualitative exploration into the widespread use of tradi-
tional and faith healers in the provision of mental health care in Ghana. Int. Rev.
Psychiatr. 22(6), 558–567 (2010)
19. Norris, L., Swartz, L., Tomlinson, M.: Mobile phone technology for improved men-
tal health care in South Africa: possibilities and challenges. South African J. Psy-
chol. 43(3), 379–388 (2013). https://doi.org/10.1177/0081246313493376
20. Esan, O., et al.: Mental health care in Anglophone West Africa. Psychiatr. Serv.
65, 1084–1087 (2014)
21. Ofori-Atta, A., et al.: A situation analysis of mental health services and legislation
in Ghana: challenges for transformation. African J. Psychiatr. 13(2), 99–108 (2010)
22. Read, U.: I want the one that will heal me completely so it won’t come back again:
the limits of antipsychotic medication in rural Ghana. Transcultural Psychiatr.
49(3–4), 438–460 (2012)
23. Read, U.M., Adiibokah, E., Nyame, S.: Local suffering and the global discourse
of mental health and human rights: an ethnographic study of responses to mental
illness in rural Ghana. Globalization Health 5(1), 13 (2009)
24. Rey-Moreno, C., et al.: Co-designing a billing system for voice services in Rural-
South Africa: lessons Learned. In: Proceedings of the Fifth ACM Symposium on
Computing for Development (ACM DEV-5 2014). New York, NY, USA: Associa-
tion for Computing Machinery, pp. 83–92 (2014). https://doi.org/10.1145/2674377.
2674389
25. Roberts, H.: A way forward for mental health care in Ghana? Lancet357, 1859
(2001)
26. Rusatira, J., et al.: Enabling access to medical and health education in Rwanda
using mobile technology. Needs Assessment Dev. Mob. Med. Educator Apps
JMIR Med. Educ. 2(1), e7 (2016). https://doi.org/10.2196/mededu.5336.. https://
mededu.jmir.org/2016/1/e7
27. Sabiescu, A.G., et al.: Emerging spaces in community-based participatory design:
reflections from two case studies. In: Proceedings of the 13th Participatory Design
Conference: Research Papers 1. PDC 2014, pp. 1–10 (2014). https://doi.org/10.
1145/2661435.2661446
28. Human Rights Watch. Living in chains: Shackling of people with psychosocial
disabilities worldwide. ISBN: 978-1-62313-8653, 6 October 2020
29. Whiteford, H.A., et al.: Global burden of disease attributable to mental and sub-
stance use disorders: findings from the Global Burden of Disease Study 2010.
Lancet 382(9904), 1575–1586 (2013)