Content uploaded by Miriam Cunningham
Author content
All content in this area was uploaded by Miriam Cunningham on Jan 17, 2018
Content may be subject to copyright.
mHealth4Afrika Alpha Validation in Rural and Deep Rural
Clinics in Ethiopia, Kenya, Malawi and South Africa
Miriam Cunningham1, Paul M Cunningham1, Darelle van Greunen2,
Alida Veldsman2, Chipo Kanjo3, Emmanuel Kweyu4, Binyam Tilahun5
1 IIMC / mHealth4Afrika / IST-Africa, 13 Docklands Innovation Park, 128 East Wall Road, Dublin 3, Ireland.
Corresponding emails: miriam@iimg.com, paul@iimg.com
2 School of ICT, Nelson Mandela University, P O Box 77000, Port Elizabeth, 6035, South Africa
3 Chancellor College, University of Malawi, Zomba, Malawi
4 @iLabAfrica, Strathmore University, Nairobi, Kenya
5 Institute of Public Health, University of Gondar, Ethiopia
5University of Gondar, Ethiopia
Abstract - mHealth4Afrika is a collaborative research and
innovation project, co-funded under Horizon 2020. It is focused on
supporting Sustainable Development Goal 3 and Horizon 2020
Societal challenges by developing, researching and evaluating the
potential impact of co-designing an open source, multilingual enabled
mHealth platform to support quality community-based primary
maternal healthcare delivery at semi-urban, rural and deep rural
clinics, based on end-user requirements in Southern Africa (Malawi,
South Africa), East Africa (Kenya) & Horn of Africa (Ethiopia). This
paper aims to share the co-design process applied to develop and
validate the alpha prototype, and the implications this had on the
design of the beta platform. The validation of the alpha prototype was
undertaken with 49 participants from 14 healthcare clinics across
Northern Ethiopia, Western Kenya, Southern Malawi and Eastern
Cape, South Africa during November - December 2016, using a mix
of observation and semi-structured interviews. These findings have
informed the co-design of the mHealth4Afrika beta platform, which
will be installed in participating clinics on a phased basis during Q3
2017. The expected outcome is a multi-region proof of concept that
can make a significant contribution in accelerating exploitation of
mHealth across Africa.
Keywords - Africa, Ethiopia, Kenya, Malawi, South Africa,
Electronic Healthcare Records, Sensors, mHealth, Telemedicine
I. INTRODUCTION
A. Background
Maternal and child mortality are amongst the most
pressing current challenges in a developing country context,
and a key regional health challenge across Africa. WHO
reports that "99% of all maternal deaths occur in developing
countries" and "maternal mortality is higher in women living in
rural areas and among poorer communities" [1]. It highlights
that "the maternal mortality ratio in developing countries in
2015 is 239 per 100,000 live births versus 12 per 100,000 live
births in developed countries". The average maternal mortality
ratio in Sub-Saharan Africa (initial focus of mHealth4Afrika)
was significantly higher at 546 per 100,000 live births in 2015,
with a wide variation at national level in partner countries
(Ethiopia – 353 per 100,000 live births; Kenya – 510 per
100,000 live births; Malawi – 634 per 100,000 live births and
South Africa – 138 per 100,000 live births).
One of the key targets of Sustainable Development Goal 3
(Ensure healthy lives and promote well-being for all at all ages)
[2] is to reduce global maternal mortality ratio to less then 70
per 100,000 live births. Methods towards achieving this goal
include increasing access to primary antenatal and postnatal
care, particularly in rural and deep rural areas and to support
healthcare workers in resource-constrained environments
where there are no doctors. As healthcare professionals with
higher levels of professional training and experience are more
likely to be found in major population centres, the rationale of
leveraging technology to increase access to quality primary
healthcare in deprived urban, semi-urban as well as rural and
deep rural communities is clear [3].
The WHO Global Observatory for eHealth defines
mHealth or mobile health as "medical and public health
practice supported by mobile devices, such as mobile phones,
patient monitoring devices, personal digital assistants (PDAs),
and other wireless devices” [4]. The European Commission [5]
states that through the use of "sensors and mobile apps,
mHealth allows the collection of considerable medical,
physiological, lifestyle, daily activity and environmental data”
and "could also support the delivery of high-quality healthcare,
and enable more accurate diagnosis and treatment”. It is
interesting that [5] concluded that "mHealth in the high-income
countries is driven by the imperative to cut healthcare costs,
while in developing countries it is mainly boosted by the need
for access to primary healthcare”. Access to primary
healthcare is a critical challenge faced by rural and deep rural
communities in developing countries [3].
Despite this, WHO [4] confirmed that "Countries in the
African Region reported the fewest [mHealth] initiatives”
based on their 2011 survey. However, in recent years as
highlighted by [3], there has been a significant growth in
mHealth related activities across Africa. These have focused on
a range of critical areas including maternal, newborn, infant
and child health (e.g. [6], [7], [8]), HIV/AIDS (e.g. [7]; [9],
[10]), and palliative care services (e.g. [11], [12]).
There is a strong policy commitment to support healthcare
delivery in each of the mHealth4Afrika intervention countries
[3], [13], [14], [15], [16], [17]. Key focus areas include:
increasing access to healthcare, leveraging technology more
efficiently and addressing a severe shortfall in the availability
of experienced healthcare professionals, particularly in rural
and deep rural areas, where most Ethiopians, Kenyans,
Malawians and South Africans still live. The Governments of
Ethiopia, Kenya, Malawi and South Africa are providing free
primary antenatal and postnatal care services at clinic level to
address the high demand for antenatal care service. However to
date the frequency of contact for antenatal care is still
significantly lower than expected. This is due to a combination
of reasons, ranging from practical challenges related to the
time, distance and cost required to travel to clinics; the quality
of service provided at the clinic; to the need to raise awareness
of the benefits of attending antenatal and postnatal care at a
clinic.
B. Insight into Intervention Clinics
The intervention clinics for mHealth4Afrika incorporate a
mix of semi-urban, rural and deep rural clinics across Northern
Ethiopia, Western Kenya, Southern Malawi and Eastern Cape,
South Africa. There are no doctors in any semi-urban, rural or
deep rural clinics in all participating countries.
Currently there is either limited or no use of technology to
support maternal healthcare delivery in these clinics. None of
the participating clinics have access to a complete electronic
patient record system [3]. The default data capturing method in
all four countries is still based on large-format paper-based
registries. This can result not just in human error, but also in
significant delays in finding information, hence compromising
the quality of service provided. It also wastes scarce human
resources due to significant levels of duplication of data across
the range of registries in use. These include general, antenatal
care, Out-Patient Department, family planning and other
registries related to specific conditions that must be monitored
such as malaria, HIV, tuberculosis and sexually transmitted
infections. A serious constraint with paper-based registries is
finding records quickly in busy clinics as well as the time
required each month to manually tally the aggregated health
indicators required by the Ministry of Health for statistical
reporting. With the exception of South Africa, generally
clinical staff have very limited or no previous digital literacy
training. There are considerable differences in communication
infrastructure in clinics within and between each country.
Where computers are installed, this is usually for a specific
programme or only used by the clinic manager or data manager
for reporting and record keeping [3].
None of the participating clinics currently have either a
local area network, WiFi network or Internet access. It was
clear during the requirements assessment and baseline study [3]
and reconfirmed during the alpha validation that is it necessary
to both provide digital literacy training and install basic
electronic infrastructure for clinics to participate in the
validation of the mHealth4Afrika platform. An IEEE
Humanitarian Activities Committee (HAC) Projects Grant was
secured to procure necessary equipment for the initial
intervention clinics in Ethiopia, Malawi and Kenya. This
equipment included a touch screen laptop to host the electronic
patient records, 10 inch tablets to be used by nurses at the point
of care for data entry and retrieval of client medical history,
high performance internal and external routers to set up a fast
local area network to support data transmission both within the
clinics and in the surrounding areas, as well as back up hard
drives. This grant also supported the procurement of solar
systems for clinics in Malawi and one clinic in Ethiopia, which
are currently (and for the foreseeable future) off the public grid.
Demographics of the intervention clinics in the
participating countries are outlined below.
The North Gondar Zone, Amhara Region is in Northwest
part of Ethiopia, with a total area of c. 45,990 km2 and an
estimated population of 2.9 million (c.3% population of
Ethiopia). It is primarily an agricultural area. The intervention
health centres are in the Wogera Wordea. Amba Giorgis Health
Centre (semi-urban) serves 6 kebeles with a catchment area
population of 42,750. It is 45 kms to the nearest hospital. It has
a mix of 4 Health Officers and 19 Nurses with some
Laboratory Technicians and Pharmacy Technicians and a
Health Information Technician. Gedebye Health Centre (rural)
serves 8 kebeles with a catchment area population of 51,460
and is 50kms to the nearest hospital. It has one Health Officer,
14 Nurses and a Laboratory Technician. Amba Giorgis Health
Centre and Gedebye Health Centre have electricity and a
mobile signal. Chichiki Health Centre (deep rural) serves 3
kebeles with a catchment area population of 13,470 and is 77
kms to the nearest hospital (3 hours on a gravel road around a
mountain). It has 8 Nurses and 1 Pharmacy Technician. It has
no electricity and very recently acquired a mobile signal.
In Kenya the functional administration of public health
facilities has been devolved to county level governments, of
which there are 47 around the country. Bungoma County is in
Western Kenya and shares a border with Uganda. It has a total
area of c. 2,207 km2 and an estimated population of c. 1.361
million (c.3.1% population of Kenya). The intervention health
clinics are within a radius of 20kms from the county referral
hospital, which is located in Bungoma Town. All the clinics
have access to electricity (although the supply is intermittent
during the rainy season) and a mobile signal. Webuye Health
Centre (semi-urban) serves the township and four villages with
a catchment area population of 24,515. There are 2 Clinical
Officers, 9 Nurses, Laboratory Technician and 1 Public Health
Officer. Kabula Dispensary serves 50+ villages with a
catchment area population of 25,792 and is 10 kms to the
nearest hospital. It has 1 Clinical Officer, 7 Nurses, Public
Health Officer and a Laboratory Technician. Kimaeti Health
Centre (deep rural) serves 42 villages with a catchment area
population of 32,434 and is 20 kms to the nearest hospital. It
has 3 Clinical Officers, 8 Nurses, Public Health Officer,
Laboratory Technician and a Pharmacy Technician. Nalando
Health Centre (deep rural clinic) serves 10 villages with a
catchment area population of 14,770 and is 15 kms to the
nearest hospital. It has 8 Nurses and a Pharmacy Technician.
Zomba and Machinga Districts are located in Southern
Malawi. Zomba has a population of 667,900, whilst Machinga
has a population of 40,600 (NSO, 2015), c.4.1% and c. 2% of
the population of Malawi. Both are primarily agricultural areas.
The intervention health centers are: Naisi (semi-urban) and
Ngwelero (deep rural) in Zomba and Gawanani (rural) in
Machinga. None of the clinics have Internet access. Naisi
Health Clinic serves 49 villages with a catchment area
population of 17,144 and is 11 kms to the nearest hospital. It
has 1 Medical Assistant and 6 Nurses. It has electricity and a
3G mobile signal. Gawanani Health Clinic (rural clinic) serves
17 villages with a catchment area population of 10,668 and is
16 kms to the nearest hospital. It has 1 Clinical Officer, 2
Nurses, Laboratory Technician and Pharmacy Technician. It
has a 2G mobile signal and limited electricity. Ngwelero Health
Clinic (deep rural clinic) serves 74 villages with a catchment
area population of 33,206 and is 49 kms to the nearest hospital.
It has 1 Medical Assistant and 4 Nurses. It has no electricity
and a weak 2G mobile signal.
The Eastern Cape province is both the largest and poorest
province in South Africa. The maternity unit of Dora Nginza
Hospital, Port Elizabeth is the urban intervention clinic. The
rural intervention clinics in the Eastern Cape are located within
the Intsika Yethu Local Municipality, a three to four hour drive
from Port Elizabeth. This municipality has a total area of 2,711
km2 and a population of 145,372 (c.0.003% of population of
South Africa). The clinics have electricity and a mobile signal.
Clinical staff have some previous experience using tablets.
Banzi Health Clinic serves 8 villages with a catchment area
population of 5,738 and is 15 kms to the nearest hospital. It has
a Clinic Manager, Professional Nurse, Data capturer and
Nursing Assistant. St Marks Health Clinic serves 6 villages
with a catchment area population of 7,273 and is 25 kms to the
nearest hospital. It has a Clinic Manager, 2 Professional
Nurses, Data capturer and Nursing Assistant. Sabalelo Health
Clinic serves 9 villages with a catchment area population of
3,516 and is 55 kms to the nearest hospital. It has a Clinic
Manager, Data capturer and Nursing Assistant. Qamata Health
Clinic serves 10 villages with a catchment area population of
5,532 and is 12 kms to the nearest hospital. It has a Clinic
Manager, Professional Nurse, Data capturer and Nursing
Assistant. Ntshingeni Health Clinic serves 4 villages with a
catchment area population of 3,958 and is 40 kms to the nearest
hospital. It has a Clinic Manager, Professional Nurse and Data
capturer.
C. mHealth4Afrika Research Focus
mHealth4Afrika (2016 – 2018) is a collaborative research
and innovation project, co-funded under Horizon 2020,
European Commission’s Research and Innovation Framework
Programme. It is focused on supporting Sustainable
Development Goal 3 and Horizon 2020 Societal challenges by
researching and evaluating the potential impact of co-designing
an open source, multilingual enabled mHealth platform to
support quality community-based healthcare delivery at clinic
level (semi-urban, rural and deep rural), based on end-user
requirements in Southern Africa (Malawi, South Africa), East
Africa (Kenya) & Horn of Africa (Ethiopia). It is researching
the impact on clinic level primary healthcare delivery of the
adoption and adaptation of a framework to support quality
maternal and newborn healthcare delivery in resource-
constrained clinics across Africa. The platform integrates:
electronic health records to store patient history, data collected
at clinic visits as well as associated tests and test results;
medical sensors to capture the results of a range of readings
(blood pressure, pulse and oxygen in blood, glucose,
temperature, haemoglobin); and analytical and visualisation
tools to facilitate interpretation and monitoring of the patient
results.
The overall objectives include to:
research end-user community requirements for rural and
deep rural communities in developing country contexts;
research and evaluate the challenges and potential benefits
associated with co-designing a common multilingual
patient record framework and integrate readings and
clinical data from tablets and medical sensors used at the
point of care;
train healthcare workers in urban, rural and deep rural
clinics on the coordinated, integrated use of medical
sensors, patient records and mobile user interfaces to
support more efficient, high quality healthcare delivery in
resource constrained environments and
pilot the integrated solution in semi-urban, rural and deep
rural health clinics in Southern Africa (Malawi and South
Africa), East Africa (Kenya) and Horn of Africa (Ethiopia)
to assess usability and user acceptance and modifications
required to facilitate wider adoption at national and
regional level.
As part of the co-design methodology, an extensive
consultation was undertaken with key stakeholders at national
level in Ethiopia, Malawi, Kenya and South Africa between
October 2015 and January 2016 to inform the needs
requirements and carry out a base line study. Leveraging a mix
of semi-structured interviews, focus groups and analysis of
national protocols, an analysis of the patient record system,
user interface, sensor, linguistic, work progress, privacy and
infrastructure requirements was undertaken to inform the
development of the alpha version of the mHealthAfrika
platform. The needs assessment analysed end user profiles,
usability and user experience requirements, health data
elements to be captured, the overall clinical workflow and
reporting requirements related to maternal health care delivery.
The objective was to minimise any potential disruptions in the
participating clinics environment during the project duration.
The baseline study [3] examined the overall clinic
environment to determine how best to integrate the platform
into day-to-day operations and what aspects of the environment
needed to be modified to optimise use. It also captured
information about healthcare workers’ work practices and
previous exposure to and use of technology to determine the
potential impact of mHealth4Afrika on the working life of
healthcare workers and clinics in different settings. It provided
valuable insights into human resource capacity, practical and
technical challenges, and equipment and infrastructure related
deficits. It also identified constraints and training requirements
of healthcare workers that must be considered during the co-
design of the mHealth4Afrika platform.
mHealth4Afrika aims to provide both direct and indirect
contributions to primary healthcare delivery at clinic level
through supporting improvements in: (a) the quality of primary
healthcare delivery through timely capture of information,
systematic storage of important data points in the patient
electronic record and improved follow up; (b) data quality and
(c) services and tests available at the healthcare clinics. It aims
to assist in motivating clients to attend antenatal visits through
the use of state-of-the-art technologies at the point of care.
Through increased client engagement with antenatal visits, this
should increase the number of deliveries undertaken in clinics
and thus help to reduce maternal mortality. Sensors can be used
to identify non-communicable diseases (hypertension, diabetes)
at the point of care and facilitate triage (not currently practiced
at clinic level). Through the use of state-of-the-art technologies,
mHealth4Afrika aims to assist in building the status and skills
of healthcare workers in the participating clinics as well as
improve the services provided through the use of sensors. It
also provides opportunities for both application and digital
literacy training to support non-medical professionals to
undertake routine procedures such as patient registration and
performing triage with the sensors to free up more patient
engagement time for clinical staff.
This paper is focused on sharing insight into the co-design
process followed to develop and validate the alpha prototype
and the implications this had on the co-design of the beta
platform. Section 2 outlines the methodology applied in the
development and validation of the alpha prototype. Section 3
provides insights into the functionality of the alpha prototype,
its validation across Ethiopia, Kenya, Malawi and South
Africa, and its implications for the co-design of the beta
mHealth4Afrika platform, limitations of the study and potential
areas for future research. Section 4 presents the conclusions.
II. METHODOLOGY
mHealth4Afrika is applying an experimental research
strategy, carrying out “an empirical investigation under
controlled conditions designed to examine the properties of,
and relationship between, specific factors" [18, p65].
Qualitative data collection (incorporating a mix of semi-
structured interviews and focus groups) was selected for the
needs requirements and baseline study [3] undertaken in the
four intervention countries between November 2015 and
January 2016. These findings were complemented by detailed
analysis of the standard paper-based registries used for
antenatal care (ANC) in Ethiopia, Kenya, Malawi and South
Africa. This provided the mandatory data sets required for the
alpha prototype. The mandatory data sets identified for each of
the four intervention countries were compared with the WHO
Essential Interventions for Maternal, newborn and child health.
This analysis identified significant additions required beyond
the WHO Essential Interventions for Maternal, newborn and
child health, which were taken into account in the design of the
alpha prototype. This significantly expanded the datasets to be
collected to create the necessary basic superset of indicators
and data points required for adoption of the mHealth4Afrika
patient record system for maternal healthcare delivery in
Ethiopia, Kenya, Malawi and South Africa. The expanded data
sets were agreed and workflow defined.
The outputs of the qualitative data collection, desktop
research and consultation with national stakeholders were
leveraged as necessary inputs to the agile design-based research
and implementation of the alpha prototype. The alpha
prototype focused on providing a working proof of concept to
support validation with the intervention clinics of the workflow
for the first antenatal care visit in participating countries, user
interface and data sets to be collected.
Qualitative data collection was undertaken to validate the
alpha prototype between November and December 2016 with
49 participants (29 women, 20 men) from 14 healthcare clinics
in Northern Ethiopia, Western Kenya, Southern Malawi and
Eastern Cape, South Africa. Semi-structured interviews and
focus groups were selected as the most appropriate data
collection method to capture insight from the validation
exercise in Southern and East Africa, in combination with
observation. Based on the use of purposive sampling
techniques, the most appropriate approach for this study was
intensity sampling [19], [20].
In terms of validating the user interface, functionality,
workflow (including ANC) and data sets to be collected,
healthcare workers were invited to look at each screen within
the four functional areas (Registration of healthcare facilities
registration of healthcare staff, registration of clients (patients)
and gathering data from clients), verbalise what actions they
could take and then take these actions. The primary objectives
were to determine: (a) the usability of the prototype; (b) the
appropriateness of the prototype functionality; (c) the
alignment of the workflow and data collected with common
practices at a healthcare facility level; (d) the alignment of the
data collected with the requirements for monthly reporting; and
(e) the level of training required prior to implementing the beta
pilot phase.
Validation sessions, which incorporated a semi-structured
individual or group interview, were undertaken in Amhara
Region, Northwest Ethiopia with the following semi-urban,
rural and deep rural intervention health centres: Amba Giorgis
Health Centre (1 participant); University Hospital, Gondor (1
participant); Gedebye Health Centre (3 participants) and
Chichiki Health Centre (4 participants). Validation sessions and
semi-structured interviews in Bungoma County, Western
Kenya were carried out at the following health clinics: Webuye
(2 participants); Nalendo (1 participant); Kimeati (2
participants) and Kabula Dispensary (2 participants).
Validation sessions and semi-structured interviews in the
Zomba and Machinga Districts, Southern Malawi were carried
out at the following health clinics: Naisi (7 participants);
Gawanani (13 participants) and Ngwelero (6 participants).
Validation sessions and interviews in the Eastern Cape, South
Africa were carried out with representatives of Nelson Mandela
University Clinic (1 participant), Dora Nginza Hospital Port
Elizabeth (1 participant), Sabalelo Clinic, Cofimvaba (1
participant) and Livingstone, Provincial and Humansdorp (4
participants). Where the validation was undertaken by a group
of staff members, each participant in the group had the
opportunity to interact directly with the application so they had
first-hand experience of going through different sections and
using a touch screen interface. Identical questions were asked
and identical protocols were followed during each validation
session with health clinics across the four countries during
November and December 2016.
mHealth4Afrika secured the necessary ethical approval
required in each country. There were no risks to participants
based on participation in this study, which was voluntary.
Participants were all adults, nursing school or university
graduates, generally fluent in English, and no vulnerable people
were targeted. The clinic management had already signed an
Informed Consent form during Q4 2015 agreeing that the data
collected throughout the duration of the project could be used
for the purpose of research, informing policy and associated
publications. To ensure anonymity, each transcript per health
facility was allocated a unique numerical code. With the
consent of the participants, interviews were audio recorded for
the purpose of the transcript to complement field notes taken
during the interview.
Following the validation sessions, full transcripts based on
the audio recordings were created to provide raw data for
analysis. Each participant or group of participants was
allocated a code to ensure that data was sufficiently
anonymised prior to data analysis, which leveraged Creswell's
Data Analysis Spiral [19].
III. MHEALTH4AFRIKA ALPHA PROTOTYPE
A. Alpha Functionality
The mHealth4Afrika alpha specification was based on
inputs provided from the needs requirements and baseline
study [3], analysis of the paper-based registries in the four
intervention countries and consultation with clinicians. It was
designed to provide a common workflow for the four countries
around use cases that follow the natural workflow in the clinic.
When a client comes to the clinic, they first visit the
reception desk or records office. If they are a new patient for a
service such as antenatal care or a new patient, they are
registered as a patient in the paper-based registry and provided
with a medical record number on a physical card. If they are an
existing client in that clinic, they will provide their medical
record number for entry in the paper-based registry. Under both
scenarios they then join the queue for consultations.
A nurse undertakes the consultation. During the first ANC
visit, a detailed medical history is taken, followed by obstetric
history, general examination, systemic examination, screening
and immunisations and supplements. A number of rapid tests
may be undertaken for syphilis and HIV. Details captured
during each consultation are recorded in the paper-based
registry. Depending on the trimester of the pregnancy at the
first ANC visit, the date for the next visit is agreed.
Use cases were developed around the different roles and
actions taken to support the ANC1 workflow. The data
elements, workflow and associated logic were set up to provide
a common back end data storage and reporting framework. An
iterative approach was undertaken for the user interface based
on end user feedback, which resulted in a cleaner and simpler
look and feel.
The alpha prototype supported new patient/client
registration capturing the necessary biographical data and next
of kin details (Figure 1).
Figure 1: mHealth4Afrika Alpha - Add New Patient/Client
All patient details are available to the healthcare workers
and registration/records staff from a patient list (Figure 2).
Figure 2: mHealth4Afrika Alpha Patient List
From the patient list, the nurse can view the patient profile
and continue to add medical history (Figure 3), edit the profile
or add a new patient (Figure 1).
Figure 3: mHealth4Afrika Alpha - View Patient Details
Having selected the client's record, the nurse can then go
through the medical history, starting with pre-existing medical
conditions and family and genetic disorders (Figures 4 and 5).
Figure 4: mHealth4Afrika Alpha - Add Pre-existing medical conditions
Figure 5: mHealth4Afrika Alpha - Add Family & Genetic Disorders
The nurse will then capture information in relation to the
obstetric history and current pregnancy (Figures 6 and 7).
Once the background history is collected the nurse will
then work through details specific to the first ANC visit:
General examination, Systemic Examination, Screening
(Malaria, Rhesus Factor, Haemoglobin, Syphilis, Diabetes,
Hypertension, Asthma, Tuberculosis, HIV, Urine) and
Immunisation and Supplements.
The alpha prototype provided a working proof of concept
to demonstrate the initial common workflow.
Figure 6: mHealth4Afrika Alpha - Previous Pregnancies
Figure 7: mHealth4Afrika Alpha - Gestational Age
B. Alpha Validation
During the validation of the alpha platform user interface,
functionality workflow and data collection, nurses used a touch
screen laptop to undertake the following actions:
Health Facilities related functionality - set up, edit and
delete.
Healthcare Practitioner related functionality - set up, edit
and delete.
Patient related functionality – register, edit & delete patient
data; add & edit next of kin.
Pre-Existing Medical Conditions, Previous Pregnancy -
add and edit details
ANC 1 Workflow - add and edit details related to
Gestational Age, General Examination, Systemic
Examination, Screening, Immunisation and Supplements
Healthcare workers were observed interacting with
different functional areas of the mHealth4Afrika alpha
prototype and were given minimal guidance in using the
platform (Figure 8). As they completed each functional area,
they were requested to provide detailed feedback on each
aspect of the user interface, functionality and workflow. Each
validation session lasted a minimum of two to three hours.
Figure 8: Healthcare workers interacting with mHealth4Afrika
platform
Overall the feedback received on the page layouts was that
it was user-friendly and intuitive/easy to use. While a
significant number of the healthcare workers had little or no
previous experience or training using a laptop or keyboard,
they were able to work out how to complete the various
sections related to patient registration, setting up healthcare
facilities and collecting clinical data for a patient based on the
antenatal care workflow, without any training.
Participants gave suggestions in relation to how some of
the action buttons could be made more obvious (e.g. colour,
location, size) and how specific data sets could be re-positioned
(for example, next of kin and allergies) to make it more
obvious this data must be collected. The aggregated feedback
provided will be implemented in the beta platform.
Some of the clinic staff interviewed indicated that it would
be very helpful if they could be trained on using a keyboard so
data entry can be done more quickly. There were initially
delays finding where specific keys were positioned on the
keyboard.
While going through all the functional areas of the alpha
platform, healthcare workers were actively encouraged to
validate the data sets included (based on analysis of the paper-
based registries used by clinics at national level and initial
feedback provided at the user requirements stage) and identify
missing data sets that must be included in the beta platform.
In relation to patient registration, participants suggested re-
ordering the presentation of first and last name, and
recommended the addition of middle name, medical record
number and country specific address details. They confirmed it
would be very beneficial to add a Patient Overview page,
summarising main data sets and adding visualisation of sensor
readings.
In terms of the ANC workflow, participants proposed
re-ordering the navigation items and moving allergies to
Medical History. They recommended re-ordering of conditions
within specific areas, suggested additional option sets to be
added into drop down lists to speed up data entry and proposed
summary tables to provide easy to use overview of all data
inputs. They validated the conditional rendering of option sets
depending on whether the condition is relevant to that client
and provided further insights to be implemented. They
proposed additional data sets to be added to Obstetric History,
Systemic Examination, General Examination and Screening.
Participants also provided insights in relation to the types of
reports that would be useful.
The findings were first collected and integrated by country
and then compared across the four participating countries to
identify global additions and modifications to be made in the
beta platform. It was striking that there were significant
commonalities in findings across participating clinics and
countries. The comparative feedback provided was
incorporated into detailed specification for beta prototype –
additional data elements to be added, modifications to clinic,
patient and ANC workflows and user interface.
C. Implications for the Co-design of mHealth4Afrika Platform
mHealth4Afrika is taking a user-centered design,
collaborative open innovation based approach. As highlighted
in [3], the importance of interventions taking account of
information needs at different stages in the continuum of care is
well documented in literature (e.g., [21], [22], [23], [24]).
Undertaking the alpha validation across the four countries
during November and December 2016 facilitated the research
teams to engage directly with the healthcare workers who are
both the beneficiaries and validators of the platform as it
evolves.
The alpha validation was a very productive exercise,
providing valuable insights into how the healthcare workers
will use the platform, usability and appropriateness of the alpha
platform functionality and the data flow that they expect in the
beta platform. It was very interesting that while clinics
confirmed additional data sets identified by counterparts in
other clinics and countries, each clinic also identified new data
sets to be included. This is a very important part of the co-
design of mHealth4Afrika to ensure that the application
addresses the requirements of healthcare workers who provide
maternal health care delivery in real-life environments.
Healthcare workers were very pleased at how their initial
inputs had been operationalised and incorporated into the alpha
prototype. They found it unusual but very beneficial that they
were consulted regularly from the very beginning of the co-
design process. Healthcare workers committed considerable
time to methodically go through the platform section by section
validating data sets and suggesting additions. Based on the
alpha validation and inputs provided both the healthcare
workers and clinic managers believe mHealth4Afrika will
allow them to capture a much more comprehensive patient
history. The feedback reinforced the importance to include
visualisation and summary reports in the beta platform. It was
also interesting how open clinicians in different countries were
to considering adopting good practices from other countries,
particularly where some countries captured additional data.
The frequent movement of healthcare workers between
rural and deep rural clinics is well documented in literature
[25], [26], [27], [28]; [29]. As a result it is critical the user
interface is intuitive and easy to use to reduce the need for
intensive application training [3].
Limited prior use of technology and low digital literacy
skills require a proactive sensitization and training approach
[3]. A mixture of observation and discussion during the alpha
validation confirmed there is a requirement for digital literacy
training, including learning how to use a keyboard. It was
agreed there should be a mix of face-to-face training, an
electronic manual as reference material with lots of screenshots
and multimedia materials installed on the laptop and tablets.
Training materials will be installed on the equipment in the
clinics by the research teams while they are providing ongoing
regular training support to avoid the clinics incurring data
charges downloading materials remotely. It was agreed that it
would be beneficial to integrate tool tips into the beta platform
user interface as an online training facility to support accurate
data entry.
As a result of the alpha validation the updated data
elements, user attributes, program structure and complete
workflow for Medical History (Past/pre-existing medical
conditions, allergies and family and genetic disorders),
Obstetric History, Clinic appointments and Antenatal care
visits 1 - 4 (General examination, systemic examination,
screening and immunisation and supplements) were configured
for the beta platform. The user interface and navigation was
refined to be more intuitive. Additional option sets were added
to reduce the amount of end user input required. Tool tips were
added as an online training aid. Roles based end user
functionality and initial reporting was added in the beta
platform.
D. Limitations of the Findings
There are a number of key limitations of this study. A
deliberate limitation was to only engage with professional
healthcare participants in rural, deep rural and semi-urban
clinics, to gather intelligence from clinical staff responsible for
local healthcare delivery. The 49 participants (29 women, 20
men) were selected from a mix of semi-urban, rural and deep
rural healthcare clinics (4 in Northern Ethiopia; 4 in Western
Kenya; 3 in Southern Malawi; 4 in Eastern Cape, South Africa)
in four African member states. While this provides geographic
representation from Southern and East Africa, the study
findings may not be representative of the local situation across
these countries or other Southern and East African Member
States, let alone all African Member States.
E. Future Research
The main purpose of the alpha validation was to inform
the co-design of the beta version of mHealth4Afrika, which is
currently being implemented. The input received during the
alpha validation and pre-beta validation is being incorporated
into the beta platform, which will be installed in the clinics on
a gradual basis during Q3 2017. The feedback provided during
the beta validation will inform the final platform specification.
mHealth4Afrika is also researching the potential impact
of integrating non-intrusive, easy to use medical sensors to
capture clinical data at the point of care, save these readings
within the patient electronic medical record for comparative
IV. CONCLUSION
This paper provides insight into the co-design process
followed to develop and validate the alpha prototype and the
implications this had for the design of the beta platform.
The alpha prototype was developed based on initial inputs
from the needs requirements and baseline study undertaken
with the clinics, complemented by detailed analysis of the
standard paper-based registries used for ANC in Ethiopia,
Kenya, Malawi and South Africa. The alpha prototype
provided a working proof of concept to demonstrate the initial
common workflow to set up patients and healthcare workers,
manage patients and capture and retrieve medical data
associated with medical history, obstetric history and ANC1
stages (general examination, systemic examination, screening
and immunisation and supplements).
The alpha validation was undertaken with 49 healthcare
workers from 14 clinics in Northern Ethiopia, Western Kenya,
Southern Malawi and Eastern Cape, South Africa, leveraging
observation of actions taken interacting with the application
and semi-structured interviews. The alpha validation provided
valuable insights into how the healthcare workers will use the
platform, usability and appropriateness of the alpha prototype
functionality and the data flow that they expect in the beta
platform. These insights have informed the co-design of the
mHealth4Afrika beta platform, which will be installed in the
clinics gradually over Q3 2017.
ACKNOWLEDGEMENTS
This research was co-funded by the European Commission
under the Horizon 2020 Research and Innovation Framework
Programme (mHealth4Afrika, Grant Agreement No. 688015).
The interpretation of the results is the sole responsibility of the
primary researchers, based on the contributions of the
participants. The primary researchers would like to thank the
representatives of clinics in Ethiopia, Kenya, Malawi and South
Africa who participated in the alpha validation for their
invaluable contributions and insight.
REFERENCES
[1] World Health Organisation (2015): Trends in Maternal Mortality 1990
to 2015
http://apps.who.int/iris/bitstream/10665/194254/1/9789241565141_eng.
pdf.
[2] UN Sustainable Development Goal 3(Ensure healthy lives and promote
well-being for all at all ages)
http://www.un.org/sustainabledevelopment/health/
[3] Cunningham, P., Cunningham, M., van Greunen, D., Veldsman, A.,
Kanjo, C., Kweyu, E. and Gebeyehu, A. (2016) "Implications of
Baseline Study Findings from Rural and Deep Rural Clinics in Ethiopia,
Kenya, Malawi and South Africa for the co-design of mHealth4Afrika",
Proceedings of IEEE Global Humanitarian Technology Conference
(GHTC) 2016, IEEE Xplore, ISBN: 978-1-5090-2432-2, DOI:
10.1109/GHTC.2016.7857350
[4] World Health Organization (2011) mHealth: New Horizons for Health
through Mobile Technologies: Based on the Findings of the Second
Global Survey on eHealth (Global Observatory for eHealth Series,
Volume 3) 2011.
[5] European Commission (2014) Green Paper on mobile Health
("mHealth") COM(2014) 219 final.
http://ec.europa.eu/information_society/newsroom/cf/dae/document.cfm
?doc_id=5147
[6] Mechael, P. (2009). MoTECH: mhealth Ethnography Report (pp. 98):
Grameen Foundation.
www.kiwanja.net/database/document/document_motech_mhealth_ethno
graphy_report.pdf
[7] Braun, R., Catalani, C., Wimbush, J.; Israelski, D. (2013) Community
health workers and mobile technology: a systematic review of the
literature. PloS one, 8(6), p.e65772.
http://journals.plos.org/plosone/article/asset?id=10.1371%2Fjournal.pon
e.0065772.pdf
[8] Wolff-Piggott, B., Coleman, J., Rivett, U. (2017) The clinic-level
perspective on mHealth implementation: a South African case study,
Information Technology for Development, Taylor & Francis, DOI:
10.1080/02681102.2016.1233858
[9] Cell-Life (2013) M.A.M.A. "Supporting pregnant women and new
mothers in South Africa." http://www.emtct-iatt.org/wp-
content/uploads/2014/02/WHO-Supporting-pregnant-women-and-new-
mothers-in-South-Africa-Cell-Lifes-MAMA-SMS.pdf
[10] Forrest, J.I., Wiens, M., Kanters, S., Nsanzimana, S., Lester, R.T.; Mills,
E.J. (2015) Mobile health applications for HIV prevention and care in
Africa. Current Opinion in HIV and AIDS, 10(6), pp.464-471.
https://www.researchgate.net/profile/Jamie_Forrest2/publication/281628
330_Mobile_health_applications_for_HIV_prevention_and_care_in_Afr
ica/links/55f0b29608aedecb68ffbf1b.pdf
[11] Grant, L., Downing, J., Namukwaya, E., Leng, M., Murray, S.A. (2011)
Palliative care in Africa since 2005: good progress, but much further to
go. BMJ Supportive & Palliative Care. 2011;1:2 pp118-122 Published
Online First: 6 August 2011 doi:10.1136/bmjspcare-2011-000057
[12] Allsop, M.J., Powell, R.A.; Namisango, E. (2016) The state of mHealth
development and use by palliative care services in sub-Saharan Africa: a
systematic review of the literature. BMJ Supportive & Palliative Care,
pp.bmjspcare-2015.
http://eprints.whiterose.ac.uk/100135/18/Manuscript_mHealthSSA_revis
ion.pdf
[13] African Union Commission (2014) On the Wings of Innovation:
Science, Technology and Innovation for Africa, 2024 Strategy (STISA-
2024), AUC, Addis Ababa
[14] Federal Democratic Republic of Ethiopia National Planning
Commission (2015b) The Second Growth and Transformation Plan
(GTP II) (2015/16 – 2019/20)
[15] NESC (2007). Kenya Vision 2030: A globally competitive and
prosperous Kenya. National Economic and Social Council of Kenya.
[16] National Economic Council (1998) Malawi Vision 2020. National Long-
Term Development Perspective for Malawi.
[17] National Planning Commission, South Africa (2013) National
Development Plan 2030, Our future - make it work, Government of
South Africa
[18] Denscombe, M. (2010) The Good Research Guide. Open University
Press
[19] Creswell, J.W. (2007) Qualitative Inquiry & Research Design: Choosing
Among Five Approaches. (2nd ed.) Thousand Oaks, CA, USA
[20] Collins, K.M.T, Onwuegbuzie, A.J. and Jiao, Q.G. (2007) A Mixed
Methods Investigation of Mixed Methods Sampling Designs in Social
and Health Science Research. Journal of Mixed Methods Research 1, 3,
267-294.
[21] Obasola, O.I., Mabawonku, I. and Lagunju, I., (2015) A review of E-
health interventions for maternal and child health in sub-Sahara Africa.
Maternal and Child Health Journal, 19(8), pp.1813-1824.
[22] Kerber, K.J., de Graft-Johnson, J.E., Bhutta, Z.A., Okong, P., Starrs, A.
and Lawn, J.E., 2007. Continuum of care for maternal, newborn, and
child health: from slogan to service delivery. The Lancet, 370(9595),
pp.1358-1369.
http://www.who.int/pmnch/media/mnchnews/2007/lancet-COC.pdf
[23] Bhutta, Z.A., Chopra, M., Axelson, H., Berman, P., Boerma, T., Bryce,
J., Bustreo, F., Cavagnero, E., Cometto, G., Daelmans, B. and de
Francisco, A. (2010). Countdown to 2015 decade report (2000–10):
taking stock of maternal, newborn, and child survival. The Lancet,
375(9730), pp.2032-2044.
http://www.who.int/workforcealliance/media/events/2010/lancet_countd
own.pdf
[24] Cooper, E. (2013). Evaluating social impact of our work- CCPF Malawi.
http://villagereach.org/2013/04/19/evaluating-the-social-impact-of-our-
work-ccpf-malawi
[25] Kalach, J. (2011) Saving mothers and newborn lives in the community
using RapidSMS technology - experience from Musanze District of
Rwanda, International Conference on Community Health, Kigali,
Rwanda, 25 - 28 January 2011
[26] Ngabo, F., Nguimfack, J., Nwaigwe, F., Mugeni1, C., Muhoza, D.,
Wilson, D., Binagwaho, A. (2012). Designing and implementing an
innovative SMS-based alert system (RapidSMSMCH) to monitor
pregnancy and reduce maternal and child deaths in Rwanda. Pan African
Medical Journal, 13(31)
[27] Cole-Ceesay, R., Cherian, M., Sonko, A., Shivute, N., Cham, M., Davis,
M., Southall, D. (2010). Strengthening the emergency healthcare system
for mothers and children in the Gambia. Reproductive Health, 7(21).
doi: 10.1186/1742-4755-7-21.
[28] Fajembola, T. (2011). Abiye: Safemotherhood: A case of leadership in
turning the tide of maternal mortality in Nigeria. Nigerian health journal.
http://nigerianhealthjournal.com/?p=1526
[29] Oyeyemi, S. O. (2012). The use of cell phone for Maternal Health: The
Abiye project. Master Thesis, The Artic University of Norway (UIT)
http://munin.uit.no/handle/10037/4793