Journal of Public Health | Vol. 40, Supplement 2, pp. ii74–ii86 | doi:10.

1093/pubmed/fdy200

Can mHealth improve timeliness and quality of health data

collected and used by health extension workers in rural
Southern Ethiopia?
W. Mengesha1, R. Steege2, A.Z. Kea1, S. Theobald2, D.G. Datiko1
1
REACH Ethiopia, Hawassa, Ethiopia
2
Department of International Public Health, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, UK
Address correspondence to Rosalind Steege, E-mail: [email protected]

ABSTRACT

Background Health extension workers (HEWs) are the key cadre within the Ethiopian Health Extension Programme extending health care to
rural communities. National policy guidance supports the use of mHealth to improve data quality and use. We report on a mobile Health
Management Information system (HMIS) with HEWs and assess its impact on data use, community health service provision and HEWs’
experiences.

Methodology We used a mixed methods approach, including an iterative process of intervention development for 2 out of 16 essential
packages of health services, quantitative analysis of new registrations, and qualitative research with HEWs and their supervisors.

Results The iterative approach supported ownership of the intervention by health staff, and 8833 clients were registered onto the mobile
HMIS by 62 trained HEWs. HEWs were positive about using mHealth and its impact on data quality, health service delivery, patient follow-up
and skill acquisition. Challenges included tensions over who received a phone; worries about phone loss; poor connectivity and power failures
in rural areas; and workload.

Discussion Mobile HMIS developed through collaborative and locally embedded processes can support quality data collection, flow and better
patient follow-up. Scale-up across other community health service packages and zones is encouraged together with appropriate training,
support and distribution of phones to address health needs and avoid exacerbating existing inequalities.

Keywords CHWs, equity, ethics, Ethiopia, Health Management Information system, HEP, maternal health, mHealth, TB

RESUMEN

Antecedentes Los agentes de extensión sanitaria (AES) son el equipo clave dentro del Programa de Extensión Sanitaria de Etiopía, llevando la
atención para la salud a las comunidades rurales. Las directrices de las políticas nacionales apoyan el uso de mSalud para mejorar la calidad y el
uso de la información. Informamos sobre un sistema de Información para la Administración de la Salud móvil con los AES y evaluamos su
impacto en el uso de la información, la provisión de servicios de salud comunitaria y las experiencias de los AES.

Metodología Usamos un enfoque de métodos mixtos, incluyendo un proceso de desarrollo de la intervención reiterado para dos de los 16
paquetes de servicios de salud esenciales, análisis cuantitativo de las nuevas inscripciones e investigación cualitativa con los AES y sus
supervisores.

Resultados El enfoque reiterado respaldó la responsabilidad del personal de la salud con respecto a la intervención y se registraron 8.833
clientes en el sistema móvil de Información para la Administración de la Salud por medio de 62 AES capacitados. Los AES mostraron una actitud

W. Mengesha, eHealth Project Coordinator

R. Steege, PhD candidate
A.Z. Kea, Field Director for REACHOUT Project
S. Theobald, Chair in Social Science and International Public Health
D.G. Datiko, Director

© The Author(s) 2018. Published by Oxford University Press on behalf of Faculty of Public Health.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
ii74 which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
 CAN MHEALTH IMPROVE TIMELINESS AND QUALITY OF HEALTH DATA COLLECTED ii75

positiva en cuanto al uso de mSalud y su impacto en la calidad de los datos, la prestación de los servicios de salud, el seguimiento con el
paciente y la adquisición de capacidades. Los desafíos incluyeron conflictos con respecto a quién recibió un teléfono, preocupación por perder
dicho aparato, los apagones, conectividad y el volumen de trabajo.

Debate El sistema móvil de Información para la Administración de la Salud integrado, desarrollado mediante procesos de colaboración e
inserción en el marco local, puede respaldar la recolección y el flujo de datos fidedignos y un mejor seguimiento con el paciente. Se alienta la
ampliación transversal con otros paquetes de servicios para la salud comunitaria y zonas, junto con la capacitación adecuada, el respaldo y la
distribución de teléfonos para abordar las necesidades de salud y evitar incrementar las desigualdades existentes.

Palabras clave mSalud, género, agentes de extensión sanitaria, ética, Etiopía, sistema de Información para la Administración de la Salud, salud
materna, tubérculo, y comunidades.

RÉSUMÉ

Contexte Les agents de vulgarisation sanitaire (AVS) sont le cadre clé du programme éthiopien de vulgarisation sanitaire qui étend les soins de
santé aux collectivités rurales. L’orientation stratégique nationale appuie l’utilisation de santé mobile pour améliorer la qualité et l’utilisation des
données. Nous présentons des évidences sur un système mobile de gestion de l’information sur la santé avec les AVS et évaluons son incidence
sur l’utilisation des données, la prestation de services de santé communautaires et l’expérience des AVS.

Méthodologie Nous avons utilisé une approche à méthodes mixtes, y compris un processus itératif d’élaboration d’interventions pour 2 des
16 forfaits de services de santé essentiels, une analyse quantitative des nouvelles homologations, et une recherche qualitative avec les AVS et
leurs superviseurs.

Résultats L’approche itérative appuyait la prise en charge de l’intervention par le personnel de la santé, et 8 833 clients étaient inscrits au
système mobile de gestion de l’information sur la santé par 62 AVS formés. Le recours à la santé mobile ainsi que son incidence sur la qualité
des données, la prestation des services de santé, le suivi des patients et l’acquisition de compétences était favorisé par les AVS. Les difficultés
comprenaient des tensions par rapport à qui recevait le téléphone, des inquiétudes au sujet de la perte du téléphone, les pannes de courant, la
faible connectivité et la charge de travail.

Discussion Un système mobile de gestion de l’information sur la santé mis au point grâce à des processus collaboratifs et intégrés localement
peut appuyer la collecte et l’acheminement de données de qualité ainsi qu’un meilleur suivi des patients. On encourage l’expansion à d’autres
forfaits et zones de services de santé communautaires, ainsi que la formation, le soutien et la distribution appropriés de téléphones pour
répondre aux besoins de santé et éviter d’exacerber les inégalités existantes.

Mots clés santé mobile, agents de vulgarisation sanitaire, étique, Éthiopie, système de gestion de l’information sur la santé, santé maternelle,
tuberculose, genre et collectivités.

Introduction
To ensure universal access to primary health care at commu- provide community-based services.6 HEWs work at health
nity level, Ethiopia launched its flagship health programme posts based in ‘kebeles’ (the smallest administrative unit).
known as the Health Extension Programme (HEP) in Two HEWs are assigned in each kebele with an average
2003.1,2 The programme was designed to provide equitable population of 5000 people (~1000 households). They devote
access to primary health care by bringing services to the 75% of their time to making house-to-house visits.3
community. Female community health workers, referred to Tuberculosis (TB) and poor maternal health (MH) out-
as health extension workers (HEWs), are the key cadre deli- comes contribute to high levels of morbidity and mortality
vering the HEP packages of services.1,3 They are recruited in Ethiopia2 and therefore are national public health priority
from the local community, which is shown to improve rela- areas included as two of the 16 packages delivered by
tionships within communities,3–5 have completed school HEWs. Since the implementation of HEP, coverage of
until at least grade 10, and received training for 1 year to health services has improved. Maternal delivery coverage
ii76 JOURNAL OF PUBLIC HEALTH

rates reached 72.7% and the TB detection rate reached iterative process of intervention development and problem
61.3% in 2016.7 Improving MH outcomes and effective TB solving; quantitative data tracking on the coverage and uptake
care requires early identification and follow up, linkage to com- of the mobile HMIS; and qualitative data collection to examine
munity or facility-based services, and improved data use, which the impact the technology had on HEWs’ experiences.
can be acted upon to support more equitable service delivery.
HEWs use the Health Management Information system
(HMIS) to collect data at the health post, which are routinely The intervention
fed up to the health centre for validation, compilation and fur- The project was conducted in Sidama zone, Southern
ther reporting.6 These data are used to track health outcomes Ethiopia, a densely populated zone with ~3.5 million people6,10
and plan the use and allocation of resources at community and ran from July 2013 to July 2017. The population level
level. The HMIS relies on paper-based reports, transported access to mobile phones is improving and most health workers
upwards from health posts to health centres and higher levels. and policy makers own their own phones. There is adequate
Key challenges to the HMIS include delays in reporting, mobile connectivity in the area (though it can suffer connectiv-
incompleteness or inconsistency of data, inadequate data collec- ity failures) and most health facilities have a power supply. The
tion tools and poor monitoring systems. This results in poor project was implemented in six Primary Health Care Units
data use, delays in patient follow-up, inability to take timely (PHCU) based in six districts, with a minimum of four and a
action and missed opportunities for responsive planning. maximum of seven health posts each. They were selected
In 2012, the Ethiopian Ministry of Health developed a based on their patient load, and via discussion with health sys-
mobile Health (mHealth) (eHealth refers to the use of ICT, tem stakeholders. We identified health facilities that had pro-
especially (but not only) the Internet to enable health and health blems with the HMIS and patient follow-up.
care.8 mHealth is the use of mobile phones and any other com- Prior to the intervention, HEWs relied on a paper-based
munication devices for health services.) strategy providing a HMIS. Informed by a baseline analysis,6 we designed and
framework for action.6,9 The framework discusses how HEWs, established a mobile HMIS for HEWs to register clients for
being the first port of call for remote and rural populations, TB and MH services using mobile smartphones. The inter-
should be the drivers for the first mHealth roll out phase.9 The vention involved providing smartphones to health workers
expectation is that improving data collection and reporting by (HEWs, heads of health centres and focal persons from the
HEWs will improve the quality and timeliness of data for local district and zonal levels), with a preloaded data entry plat-
decision-making, and consequently increase accountability, form, via an adapted CommCare platform11 (see Appendix 4).
transparency and redress inequities in the health system.9 The mobile HMIS intervention was implemented alongside the
In response to the mHealth strategy and to the central existing paper-based HMIS. Programme managers, health
role of HEWs in providing more equitable and responsive workers and HEWs were trained on TB and MH, data col-
health services, we designed and implemented an integrated lection and entry onto the platform; how to use the
mHealth intervention (The intervention was part of a pro- reminder text messages and client follow-up.
gramme funded by the International Development Research The mobile HMIS was developed in close collaboration
Centre (IDRC) and called Strengthening Equity through with the Ministry of Health and using the Ethiopian
Applied Research Capacity Building in eHealth (SEARCH). mHealth strategy as a guiding framework. The use of open
The SEARCH project ran from 2014 to 2017 and built on standards and open source software system enabled more
prior work that began in 2010 under the TB REACH pro- seamless and secure integration within existing HMIS for-
ject implemented by REACH Ethiopia.) using mobile mats and data flow. A total of 97 smartphones were distribu-
phones to support timely capture of quality data by HEWs ted to HEWs, and eight computers to health centres in the
to improve the effectiveness and equity of primary health implementation districts. Sensitization and training was con-
care service provision.2,10 The aim of the present study was ducted with 126 stakeholders at different levels of the health
to assess the impact of the intervention on quality of data system—including focal persons for TB, MH and HMIS,
collection and reporting for TB and MH, as well as to exam- HEW supervisors, heads of health centres, districts and all
ine experiences of HEWs using the technology. 62 HEWs from the PHCUs. Data entered was anonymized
and uploaded to a central database accessible at different
levels.
Methodology The mobile HMIS enables HEWs to register clients dir-
The study was a mixed method evaluation. Data were col- ectly onto the mHealth platform. For each client, data are
lected via three complementary processes: learning from the captured about sociodemographic details, symptoms and
 CAN MHEALTH IMPROVE TIMELINESS AND QUALITY OF HEALTH DATA COLLECTED ii77

duration, laboratory results, treatment, follow up and out- Structured notes were taken from these different interac-
comes (for TB clients) and information on gravidity, parity, tions, which were used to frame and support our analysis
gestational age, laboratory examination for ANC, delivery (see Appendix 1).
and its outcome and post-natal care (for MH clients). To
ensure completeness of data, inbuilt data validation was Quantitative tracking of impact
included so that fields cannot be left blank.
The maternal health and TB registration data captured via
Data was stored and uploaded to the cloud when network
smartphones were immediately available for analysis once
connection was available and immediately synchronized to
uploaded to the main server and was tracked from
the HMIS. This helped supervisors and heads of health cen-
December 2015 to July 2017. The anonymized data were
tres track and analyse data. Reminder messages were sent to
routinely downloaded by the IT professional of the project
HEWs and programme focal persons with the aim of redu-
(W.M.), exported to an excel sheet, and then checked and
cing the number of patients who may be lost to follow up
analysed to monitor numbers of new registrations and alert
(Fig. 1). The messages included reminders for ANC or TB
messages sent. These data were cross checked with the exist-
laboratory tests and due dates for delivery. A minimum of
ing paper-based HMIS for regular updates, completeness,
three messages were sent per client.
the number of follow-up messages received, and action was
taken to ensure follow up if needed. Server-based and paper-
Methods based data were used to evaluate mHealth activity. When net-
work problems arose, the paper register was used to fill gaps.
Iterative learning
The intervention had a strong focus on training and sensi-
tization to lay the foundation for sustained positive change Understanding HEW experiences
among female HEWs. An engagement strategy was used to Qualitative methods included face-to-face semi-structured
ensure ongoing communication and problem solving, for interviews (n = 19) and focus group discussions (FGDs)
example, through the setting-up of a technical working group (n = 8) with a mix of HEWs, supervisors and community
and district and catchment level meetings. Regular activities leaders (Table 1). Interviews were conducted in four of the
enabled dialogue and feedback on ongoing challenges and intervention districts purposively selected for variation in
opportunities emerging from the project, including: geographic location and performance.
Interview topic guides addressed if and how smartphones
• Training and sensitization workshops with health workers helped or hindered the HEWs’ role; use of the phones out-
and key stakeholders to create awareness of the project side of work; and changes in relationships between and
aim and related processes. among health staff since the introduction of the smartphones.
• Regular supportive supervision of HEWs alongside a A female research assistant, fluent in Sidamigna (local dialect)
review of their activities. was recruited and trained over two days. The lead qualitative
• Ongoing stakeholder meetings held at all levels of the researcher (R.S.) was present during the interviews to clarify
intervention in the community. any questions or concerns. Interviews were conducted at
• Regular HEW meetings to discuss the project, share experi- health posts and health centres within private areas, at a time
ences and highlight ways to improve project performance. convenient to the respondents. Following informed consent,
interviews were recorded using digital Dictaphones.
The recordings were transcribed verbatim and translated
to English by experienced researchers. The quality of transla-
tion was checked by a member of REACH Ethiopia (A.Z.
K) using sample transcripts. Transcripts were read and re-
read by the lead researcher (R.S.), informing the develop-
ment of codes for analysis, identifying emerging themes and
areas for further exploration through the iterative develop-
ment of a coding framework in NVivo.12,13

Ethics statement
Ethical approval was given from the Federal Ministry of Science
Fig. 1 Flow diagram of the intervention process. and Technology, National Research Ethics Review Committee
ii78 JOURNAL OF PUBLIC HEALTH

Table 1 Qualitative interviews conducted by participant, district and sex

District participant District 1 District 2 District 3 District 4

HEW 2 × IDIs (Female) 3 × IDIs (Female) 4 × IDIs (Female) 5 × IDIs (Female)

1 × FGD (Female) 1 × FGD (Female)a 1 × FGD (Female)
HEW supervisor 1 × IDI (Female) 1 × IDI (Male) 1 × IDI (Male) 1 × IDI (Male)
1 × IDI (Male) 1 × FGD (Male)b 1 × FGD (Male) 1 × FGD (Male)
Community leaders 1 × FGD (Male) 1 × FGD (Male) 1 × FGD (Male) 1 × FGD (Male)

FGD = focus group discussion; IDI = in depth interview.

a
Merged with participants from District 3.
b
Merged with participants from District 1.

in April 2014. A support letter was obtained from Regional great satisfaction. This is common not only for health
Health Bureau to conduct the interviews. Written informed extension workers but also us. [Supervisor, FGD]
consent was obtained from the participants of the study.
It is advisable to provide the mobile phone for all of us to
give more services for pregnant women as well as for the
community. [Community leader, FGD]
Results
Increased accountability HEWs reported the mHealth system improved in the
Accountability occurs at different levels within the health sys- speed of preparing and delivering messages to their clients.
tem; from the HEWs to the communities they serve; from the The use of smartphones for prompting appointments was
supervisors to the HEWs; and from the supervisors and collea- triangulated with similar responses from health profes-
gues at the health centre to the district and regional level. The sionals. Health-centre heads and kebele administrators indi-
iterative approach taken ensured buy-in from all stakeholders cated that use of smartphone reminders, if extended to
and a sense of responsibility to account for one’s actions. This kebele administrators, may reduce the number of non-
led to interest and accountability in the mobile HMIS, evi- attendees at clinical appointments. The main challenges
denced by continued discussion and engagement at various reported by all respondents was the lack of internet connect-
meetings throughout the project (see Appendix 1). Data can ivity and power failures in rural areas. These issues were
now be collected in real-time and is accessible throughout all solved by supervisors travelling to nearby areas with the
levels of the health system through different reports and dash- mobiles where connectivity and power could be reached.
boards (see Appendices 2 and 3 for more information on
Registrations and population reached
reporting and information flow). Interviews with stakeholders
suggest that this aided decision making and has helped create Prior to the project intervention, only paper-based recording
stronger links with policy makers for action. and reporting existed; while delays and inconsistencies in
reporting were common. Compared to the baseline, data on
It helps to hold clients’ information, to disseminate the TB and maternal health have improved in accuracy, complete-
necessary information from top to bottom and vice versa, ness and timeliness, as confirmed by multiple respondents.
it also helps for decision making purposes by relying on
the data, to track the progress of the service and it also There is an improvement in data quality as compared to
helps as a reminder to give service... [Supervisor, IDI] previous years. ……[P]reviously when pregnant woman
come to first/second visit of antenatal care, we register her
Supervisors, who also receive the alert messages, indicated information and wait several days without reporting [her].
that the alerts encourage their active participation in patient But now we report daily performance within 2–3 h to mul-
follow-up and community leaders expressed a wish to tiple sites…This also helps to avoid data fallacy/wrong
receive the phones to aid follow-up in communities. reports…and avoid missing to report. [HEW, FGD]

There is good motivation toward [the HEWs’] work. The data quality is improving for instance, when they feed
Additionally, when they receive the message alarm for a data and miss something it is difficult to continue to the
particular woman after entering data in mobile, they feel next steps unlike in hard copies. [Supervisor, FGD]
 CAN MHEALTH IMPROVE TIMELINESS AND QUALITY OF HEALTH DATA COLLECTED ii79

In past years, we were giving report to woreda [district] who are lost to follow-up. About 2 700 alerts had been sent
without performing the actual work in that particular for MCH and 500 for TB by the end of July 2017 (Table 4).
kebele and this was cheating the community. But now this Follow-up alerts may directly benefit patients who may
mobile helps to keep and report actual work that is done. have otherwise been lost to follow-up due to health system
[Community leader, FGD] challenges (no follow up/poor record keeping) or inequities
that prevent access to care (by disability, geography, gender
The 62 HEWs involved in the project worked with over inequities and/or financial constraints). Qualitative inter-
200 000 rural women and men. This population benefited views revealed that the follow up alerts that serve as remin-
from more targeted health services resulting from higher ders to the HEWs (and their supervisors) enable them to
quality and more timely use of data collected. Increases were follow up with clients in a timely fashion, promoting equit-
observed in the number of pregnant women identified, pre- able and responsive continuity of care.
sumptive cases referred, and TB cases detected over time,
resulting in improved follow-up and increased service The mobile phone is helping us to give pregnant women
uptake. The number of clients registered for skilled delivery the necessary continuity of care such as antenatal care…It
increased every reporting period (Table 3), demonstrating decreases…dropout rate. [Supervisor, FGD]
increased enrolment in the system.
There are differences over time in numbers of ANC visits After training and handling this mobile I followed many
scheduled through the mHealth system. These discrepancies mothers through mobile message and this makes me so
in reporting can be attributed to the fact that some pregnant happy. Because previously there were occasions where I
women may not reach ANC4, possibly due to delayed initi-
ation of ANC and delivery before reaching the fourth visit. In
addition, some woman who do not follow ANC still deliver Table 2 Sidama zone (intervention zone) report at baseline (July 2013–
June 2014)
at health facilities. Qualitative research with HEWs and super-
visors also revealed variations in reporting due to the nomadic
Module July ‘13– Oct ‘13– Jan ‘14– April ‘14– Total
nature of people in some districts, improved family planning Sept ‘13 Dec ‘13 March ‘14 June ‘14
services and reduced falsification of data. Collectively these
can explain improvements in the accuracy of this measure ANC 1092 1083 1050 1060 4285
and decreases in registrations for ANC and delivery as shown TB cases 66 78 72 38 254
from baseline (Table 2) to endline (Table 3). Delivery 465 554 699 723 2441

Yes, there is a reduction in the performance especially

institutional delivery…The reason for this is most people Table 3 Sidama zone (intervention zone) cumulative report during the
project period (December 2015–July 2017)
in our kebele…frequently move around…and, there is a
tendency to inflate reports. There is a change in attitude
Module Dec ‘15– Dec ‘15– Dec ‘15– Dec ‘15–
following the start of using this mobile. The data being April ‘16 Sept ‘16 Feb ‘17 July ‘17
fact makes the number reported less. [Supervior, FGD]
ANC 1009 2072 2132 2409
We faced a problem after we began using this technology.
TB cases 190 173 295 333
The health centre head and kebele focal asked that reason
TB symptomatic cases 157 455 486 510
for the decline of our kebele antenatal care performance. Delivery 981 1909 2590 2635
If the coverage of family planning is high, it is obvious Post-natal care 901 1098 2076 2410
that the antenatal coverage is low. We informed them the
reason behind for the decline of the service is due to the
improvement of data quality. [HEW, IDI] Table 4 Total number of text alerts received by the HEW, for maternal
health and TB services by district at project endline

Alert messages No. of ANC No. of Skilled No. of TB

At baseline, health officials did not receive alert messages. alerts sent delivery alerts sent alerts sent

Registration onto the mHealth platform triggers alert mes-

Total 1236 1464 500
sages so HEWs can follow up with clients and reduce those
ii80 JOURNAL OF PUBLIC HEALTH

forget to follow up of mothers for several reasons. [HEW, For example, I have gained knowledge on the mobile and
FGD] it makes me able to do quality work and this increased my
participation in different meetings. [HEW, IDI]
They are identifying pregnant women and are feeding the
data easily, it eases their work burden and we are follow- HEWs—regardless of phone ownership—reported receiv-
ing pregnant women to get continuity of care. ing recognition from the community. However, one HEW
[Community leader, FGD] reported that with this comes more expectations from the
community. This requires exploration; the technology may
Conversely, this method of active follow-up can lead to add additional pressure, such as stress from feeling unable to
ethical dilemmas. For example, clients may have reasons for meet expectations, and increased workload for HEWs.
not wanting to follow an approved course of action (e.g. giv-
ing birth in a delivery centre), which should be understood,
discussed and respected. HEWs need to be appropriately Ethical concerns for HEWs
supported in these situations. The phones were only given to one HEW per health post to
Reflections from the iterative learning process revealed avoid duplication of data, however, only one participant
that during the reporting period, HEWs’ capacity to use the reported understanding this to be the reason for this alloca-
software improved. Taking a step back, it is important to tion. HEWs who did not receive phones felt discouraged
explore if and how supplying mHealth technology to the all- that they were not upgrading their skills in the same way:
female HEW cadre poses ethical issues, including putting despite taking part in the initial training, they did not have
them at risk of assault, theft or placing additional profes- opportunity to practice using the phones.
sional burdens (e.g. workload).
The organization provided the phone for routine work
but she considers it as her own property since it was given
HEWs’ experiences free of charge. I am identifying pregnant women using the
Smartphones as a support phone, but I am not practicing on it. [HEW, IDI]
From qualitative interviews the main benefits for the HEWs
pertain to skill building and increased respect from commu- Supervisors also suggested this could lead the HEWs
nity members. Overall, HEWs viewed the technology as an without the phone evading mHealth related reporting, which
aid to their work; enhancing accuracy of reporting, motiv- may inadvertently add to the workload of the HEW with the
ation, skill-building and confidence and support via the SMS smartphone, and cause tensions.
reminders. Since the other HEW is observing her colleague using
I became motivated. For example, previously if data of [the phone] for purposes other than routine work she
particular mother registered and reported once in a may not have good feeling. The one who doesn’t have the
month, but now it helps us to follow up the mothers as phone thinks that other HEW may get incentives…[and]
TB cases [so] they do not miss their appointment date. may consider herself as inferior to her colleague and she
[HEW, IDI] might also say the mobile health issues are not my con-
cern. [Supervisor, FGD]
[The phone] is good to work. For example, I may miss
the appointment date of pregnant woman when I have Notwithstanding this, most participants did not report
work burden…but the mobile reminds me by the alarm… problems working together and some reported relying on
[HEW, IDI] their supervisors to help fill in any skill gaps with regards to
using the technology.
The intervention has also provided an opportunity to One area of consideration that will need to be addressed
upgrade HEW skills and knowledge. Those with a smart- if the technology is rolled out across all essential health
phone at the health post reported increased participation in packages is the duplication of tasks. Participants reported
district level meetings, compared to control districts. These that they must spend time inputting the data into the phones
consequences may indirectly benefit individual HEWs’ car- in addition to fulfilling paper-based reporting duties. Others
eer advancement opportunities as further training is offered however expressed that this additional work burden was
to best performers, however, we were unable to measure acceptable given the benefits to community members and
this impact. themselves.
 CAN MHEALTH IMPROVE TIMELINESS AND QUALITY OF HEALTH DATA COLLECTED ii81

R2:… we are supposed to register information in mobile system. This inevitably produces some teething challenges and
and this caused too much work burden as it takes time to support and follow-up was required. Existing literature demon-
feed information both in mobile and register. I consider it strates that technologies are embedded within existing social,
as simply a fashion but not to help community…I feel it cultural, economic and political structures.14,15 MHealth inter-
added only work burden as we [only] register 10–15 ventions require changes in the behaviours of service providers
mothers in daily basis. (HEWs) and their patients (community members).15 These
changes are driven by social, cultural and environmental factors
R3: Even though it causes work burden; it is important as and as such, they require careful sensitization and customiza-
we are helping the mothers through it. [HEWs, FGD] tion to have the intended positive impacts.
Research on CHWs’ use of mHealth tends to focus on
health outcomes or health system benefits of pilot initia-
Discussion tives.16–19 To this end, mHealth has been piloted for use
with HEWs and midwives as a data collection tool in
Main finding of the study
Northern Ethiopia.20 Reporting on their intervention, Little
Results from this study show that integrating an mHealth
et al. found an improvement in the access of data and that
solution, with a strong engagement and support strategy, can
ownership of smartphones is a strong motivator for
improve the quality and timeliness of data collection by
HEWs.21 Echoing a sentiment that was also expressed by
HEWs for MH and TB in Sidama Zone. The decision to
some HEWs in our study area, a systematic review by White
involve HEWs, their supervisors, as well as focal points at
et al.18 also found that mobile HMIS that do not allow
health centres, district and zonal levels, allowed the interven-
CHWs to leave blanks in registration forms do not speed up
tion to examine different forms of accountability that can be
data entry and rather, add to CHWs work burden as they
strengthened through improved data collection, management
are forced to be more thorough.
and use. For example, use of alert messages and related fol-
low up activities by HEWs strengthened accountability to What this study adds
the community; greater communication and support between
MHealth is arguably an important approach to supporting
HEWs and their supervisors strengthened bi-directional account-
and empowering HEW to better provide integrated services
ability; and engaging with decision-makers at different levels
and feed information and priorities from communities into
of the health system strengthened their ability to use data to
decision making processes. However, the experiences of the
improve accountability of planning activities.
HEWs and the technology influences on their workload and
The mobile HMIS improved the completeness and quality
experiences has been less well examined. Our findings high-
of data in the intervention areas. HEWs using the system
light that HEWs have the technical capacity to use electronic
felt a greater sense of opportunity and power to serve their
data capturing mechanisms to improve client follow-up, but
clients better, as well as additional responsibility to manage
that consideration should be given to addressing the ethics
their time, workload and the smartphones. Findings also
and equity of mHealth interventions to better support
pointed to the need for continued attention placed to the
HEWs. Firstly, efforts are required to ensure mHealth
ethical dimensions of distributing smartphones, understand-
reduces rather than increases workload of HEWs. In this
ing the socio-cultural, gender and financial implications of
case the mHealth process focused on two priority health
their use, and reconciling active follow-up of clients with
areas only rather than the 16 health packages within the
respectful means to understand reasons for non-adherence.
remit of HEWs. Hence, they had to simultaneously use
The Ethiopian HEP has been designed to support and
mHealth and paper-based reporting. Future steps to expand
strengthen the equity and reach of the health system and is
mHealth across the whole package of work should alleviate
widely heralded as a successful approach.2 The HEWs who
this burden. Other challenges in the project related to only
were using the mobile HMIS were overall very positive
one of a pair of HEW situated at the health post receiving a
about the impact of mHealth on their work, the skills they
phone. Again, where possible, ensuring all HEWs are appro-
had developed, their relationships with the community and
priately equipped, so they have an opportunity to practice
in their ability to appropriately follow-up with patients.
their training and build their skills, would ensure the benefits
are appropriately shared. Clear follow up training and shared
What is already known on this topic agreements on what to do in the case of theft or loss or
The intervention introduced a new approach to data capture, leaving the role of HEW, that do not penalize HEWs is
data flow and data sharing at different levels of the health necessary.
ii82 JOURNAL OF PUBLIC HEALTH

Limitations of this study Acknowledgements

This article summarizes results from a mixed method process
We would like to thank Yamrot Haile for conducting the
evaluation. It captured data and perspectives in real time, which
interviews as well as the HEWs, health workers and supervi-
limited its ability to adopt a long-term perspective. We do not
sors who contributed to the success of the project. We
yet know how, or if, the mHealth process will be sustained
would also like to thank Chaitali Sinha and Anne-Marie
through time or what its implications may be on responsive
Schryer for their comprehensive review of the article and
planning and governance. Ensuring ongoing process evaluation
insightful comments.
using both quantitative data, and the qualitative perspectives of
the HEWs (and the health systems and communities they
serve) will continue to be important to ensure equity and ethics
are meaningfully considered and addressed. Funding
The intervention was also implemented on a relatively
small scale, for only two of the 16 essential package areas. This work was supported by the International Development
As such, while the ongoing paper-based reporting is Research Centre Grant: 106229-013.
required, we cannot get a true sense of how the mHealth
technology may support HEWs if it was rolled out. Finally,
problems related to power failure and connectivity varied by References
district and not all districts began the intervention at the
same time and with equal intensity, which resulted in varia- 1 Wang H, Tesfaye R, Ramana GNV et al. Ethiopia Health Extension
tions in data collection, reporting and use. As a result, effect- Program: An Institutionalized Community Approach for Universal Health
Coverage. Washington: The World Bank, 2016.
ive comparisons could not be made. Further collaboration
2 Yassin MA, Datiko DG, Tulloch O et al. Innovative community-
with the Ethiopian government, who dominate the telecom-
based approaches doubled tuberculosis case notification and
munications market in Ethiopia, to improve the network improve treatment outcome in Southern Ethiopia. PLoS One 2013;
coverage would greatly improve the efficiency of this project 8(5):e63174.
and help ensure successful scale-up. 3 Kok MC, Kea ZA, Datiko DG et al. A qualitative assessment of
health extension workers’ relationships with the community and
health sector in Ethiopia: opportunities for enhancing maternal
Conclusion health performance. Hum Resour Health 2015;13(1):1–12.
The mobile HMIS has improved the quality and timeliness 4 Naimoli JF, Frymus DE, Wuliji T et al. A community health worker
of data flow and the responsiveness and impact of the health ‘logic model’: towards a theory of enhanced performance in low-
and middle-income countries. Hum Resour Health 2014;12:56.
system. Female HEWs in Ethiopia play a critical interface
5 Kok MC, Dieleman M, Taegtmeyer M et al. Which intervention
role between the health system and rural communities.
design factors influence performance of community health workers
Providing them with smartphones and a responsive mHealth in low- and middle-income countries? A systematic review. Health
platform, along with focused training and supportive supervi- Policy Plan 2015;30(9):1207–27.
sion, has enabled this important cadre of health workers to 6 Dusabe-Richards JN, Tesfaye HT, Mekonnen J et al. Women health
use new technologies and build skills that are positively extension workers: capacities, opportunities and challenges to use
impacting delivery of health services in their communities. eHealth to strengthen equitable health systems in Southern
HEWs face large workloads, as is common with many Ethiopia. Can J Public Health 2016;107(4–5):e355–61.
close-to-community providers and we need to do our best to 7 Ethiopia FDRo. The 2017 Voluntary National Reviews on SDGs of
ensure that this new technology supports them in their work- Ethiopia: Government Commitments, National Ownership and Performance
Trends. Addis Ababa: National Plan Commission, 2017.
load and interactions with community members rather than
8 Grantmakers in Health. Examining e-health. Issue Brief (Grantmakers
bring new burdens. An equity and ethics lens is critical for
Health) 2002;14:1–28.
mHealth technology to ensure that new approaches appropri-
9 Ethiopian FMoH. Ethiopian National eHealth Strategic Report.
ately support staff whose job it is to use these new technologies;
Addis Ababa; 2014.
and result in better services to the communities they serve.
10 Datiko DG, Yassin MA, Theobald SJ et al. Health extension work-
ers improve tuberculosis case finding and treatment outcome in
Supplementary data Ethiopia: a large-scale implementation study. BMJ Glob Health 2017;
2(4):e000390.
Supplementary data are available at the Journal of Public 11 Dimagi. [cited 2019 27th September]. https://www.dimagi.com/
Health online. commcare/
 CAN MHEALTH IMPROVE TIMELINESS AND QUALITY OF HEALTH DATA COLLECTED ii83

12 Ritchie J, Lewis J. Qualitative Research Practice: A Guide for Social Science households to ensure successful implementation of the
Students and Researchers. London: Sage, 2003. project and learn from the implementers about the contri-
13 Ritchie J, Spencer L. Qualitative data analysis for applied policy bution of the project to improve the health system. This
research by Jane Ritchie and Liz Spencer. In: Bryman A, Burgess provided for further process evaluation through the follow-
RG (eds). Analyzing Qualitative Data. London: Routledge, 1994,
173–94.
ing channels:
14 Pinch TJ, Bijker WE. The social construction of facts and artefacts:
or how the sociology of science and the sociology of technology
• Training and sensitization workshops were conducted
might benefit each other. Soc Stud Sci 1984;14(3):399–441. with key stakeholders to create awareness of the project and
15 Maar MA, Yeates K, Perkins N et al. A framework for the study of its potential to improve health service delivery. Training was
complex mHealth interventions in diverse cultural settings. JMIR also conducted with HEWs and health staff on HMIS.
Mhealth Uhealth 2017;5(4):e47. • Regular supportive supervision of HEWs was con-
16 Källander K, Tibenderana JK, Akpogheneta OJ et al. Mobile health ducted by project staff on a weekly basis alongside a
(mHealth) approaches and lessons for increased performance and review of their activities. The forum was meant to discuss
retention of community health workers in low- and middle-income the implementation and use of mobile phones, perform-
countries: a review. J Med Internet Res 2013;15(1):e17.
ance and challenges encountered. In addition, the enrol-
17 Agarwal S, Perry HB, Long LA et al. Evidence on feasibility and ment of patients and clients to care was monitored in the
effective use of mHealth strategies by frontline health workers in
developing countries: systematic review. Trop Med Int Health 2015;
community and follow up plan was designed to improve
20(8):1003–14. the project implementation.
18 White A, Thomas DS, Ezeanochie N et al. Health worker mHealth util- • Ongoing stakeholder meetings occurred throughout
ization: a systematic review. Comput Inform Nurs 2016;34(5):206–13. project period (20× catchment health centre meetings, 2×
19 Braun R, Catalani C, Wimbush J et al. Community health workers district meetings, 2× province meetings and 2× national
and mobile technology: a systematic review of the literature. PLoS level meetings) to enable experience-sharing, debates and
One 2013;8(6):e65772. discussions on the use of mobile phones to support health
20 Little A, Medhanyie A, Yebyo H et al. Meeting community health care; further customization of the results and lesson learned
worker needs for maternal health care service delivery using appro- and provided fora to discuss future directions on the utiliza-
priate mobile technologies in Ethiopia. PLoS One 2013;8(10):e77563. tion of mHealth on other health care activities.
21 Baqui AH, Rosecrans AM, Williams EK et al. NGO facilitation of a • HEW meetings were held within districts to ensure HEWs
government community-based maternal and neonatal health pro-
were given opportunity to share their experiences in a com-
gramme in rural India: improvements in equity. Health Policy Plan
2008;23(4):234–43. fortable environment. The meetings occurred over twenty
times throughout the course of the project with the aim of:
(a) Evaluating each district’s mHealth activities.
Appendix (b) Sharing knowledge and experience of mHealth activities.
(c) Provide explanation for, or fill any gaps in data
Appendix 1: Further information on process
collection.
evaluation:
(d) Evaluate the strengths and weaknesses of mHealth
The project staff conducted regular supportive supervi- activities for each district.
sion to the districts, health centre, health posts and (e) Check the performance of district supervisors.
ii84 JOURNAL OF PUBLIC HEALTH

Appendix 2: Platform of data generation

 CAN MHEALTH IMPROVE TIMELINESS AND QUALITY OF HEALTH DATA COLLECTED ii85

Appendix 3: Information flow for the Health Extension Programme

ii86 JOURNAL OF PUBLIC HEALTH

Appendix 4: CommCare data entry platform interface