MomConnect

A suite of mobile technologies fostering long-term, supportive dialogue with mothers to improve health in South Africa

How might we deliver a cheaper and more impactful maternal and newborn health at scale in South Africa?

The Story

MomConnect is an initiative of the National Department of Health (NDOH) supported by a range of partners in South Africa including Praekelt.org1. At the heart of the initiative is a national mobile messaging service that directly supports mothers, connects them to the health system and provides efficient access to health information for tracking and managing service delivery. Introduced in South Africa in 2014 as an SMS service, the MomConnect initiative now includes WhatsApp messaging, a health facility-based app to register pregnancies and a Road to Health app providing content around caring for children aged 0 – 5 years, developed by the MomConnect consortium of partners including Jembi Health Systems and Health Information Systems Program (HISP). MomConnect currently reaches 95% of public health facilities in South Africa and is available in 11 languages. The technology platform has also supported similar programs in Nigeria and Uganda.

This Case Study is focused on the design-approach that resulted in WhatsApp being introduced to MomConnect to deliver a cheaper and more impactful service text message service. But it is useful to first establish how design was approached in earlier in the MomConnect’s history.

MomConnect launched in 2014 as a USSD and SMS-based national pregnancy registry, through which every pregnant woman in South Africa could register to receive free, informative, stage-based messaging during pregnancy and for the first year (and up to two years from 2018) of her baby’s life. For users, the service provides access to a text-based Helpdesk where questions, complaints or compliments are sent directly to relevant officials to ensure mothers’ and nurses’ feedback drives service delivery improvements.

MomConnect was based on an earlier mobile health messaging program that developed content using funders’ materials and experience that was tailored to the South African clients and context, and tested with mothers for comprehension in different languages, as well as for comprehension and testing the tone of the messaging. Based on the caring personas of people who would normally give mothers advice, the developers created a composite persona to serve as the brand or voice of MomConnect. That voice is described as projecting a friendly tone as if receiving advice from a trusted source. “We were keen from a design perspective that the voice of MomConnect not be authoritative, but like a companion or friendly nurse.” Design in the form of user interviews and ethnography was also used to shape the service mothers received at health facilities including pregnancy registration and clinic processes.

The use of design in these early stages led the team to create a much bigger and impactful program. MomConnect was originally intended as data capture for pregnancy, introducing digital systems to facilitate data quality, flow and use. The team, applying the user-centered lens, went further to explore the acceptability of a new system with mothers and the nurses, asking: “What would make a mom want to sign up for this registration; What would make a nurse want to register a pregnancy on a mobile phone?” This was a fundamental change in how the project was viewed and in the approach taken by the MomConnect team. The idea of creating value for all the stakeholders in the ecosystem and motivating them to use MomConnect inspired the idea of making the program a mechanism for connecting mothers with the health system through their phones, providing health messages.

The mobile messaging uptake proved attractive to mothers. However, implementers felt that innovative efforts were essential but only if they could be delivered sustainably to every pregnant woman in the public health system in the country. For MomConnect, prohibitive SMS inventory costs to support the 800,000 women using MomConnect by 2017 threatened its sustainability. Sending essential information via SMS to pregnant women and new mothers was costing approximately USD 100,000 per month, over 40% of monthly program expenses, even after discounted rates were negotiated with local mobile network operators. To solve for this, MomConnect applied for and was chosen to be a WhatsApp development partner during the closed-group, beta development phase of their WhatsApp Business API. Through a rigorous service design process, they were able to understand the benefits of WhatsApp and design a simple solution for the mothers who want to use it, while ensuring they were still able to connect with the hardest to reach through SMS. Since the original design work (2017) that helped shape the next generation of MomConnect, the intervention has evolved even further. For example, the design research in 2017 resulted in a pilot with 394 mothers that tested a modified registration process to allow mothers to choose either WhatsApp or SMS. In April 2018, the USSD self-service option for switching from SMS to WhatsApp was introduced. A drive to encourage users to switch from SMS to WhatsApp was conducted in May 2018 and registration to WhatsApp as a default (if the user had an account) was introduced in September 2018.
1 Praekelt supported the DOH to develop a suite of mobile technologies, tools and methodologies that allow governments and other implementing partners to create and manage customize, population-scale health programs, like MomConnect.

Process and Measurement

Amount of Design: End to End

This was a technology project that introduced human-centered approaches in its early stages. The designers gradually steered it toward a full HCD process as the program grew and new channels such as WhatsApp were introduced. There was some resistance in the beginning to taking a user-centered approach and finding suitable times to access clinics, and interview staff and mothers was a challenge. Additionally, in working with health experts, the design processes had to be framed in a language that made sense from a health perspective. This meant talking about how they could strengthen the way in which they implement a program and showing impact early on. Members of the DOH and other stakeholders now participate in the design process whereas before they were outside of it.

An illustration of the design process is most clear in the introduction of WhatsApp and the new WhatsApp Helpdesk. The team took an agile approach, incorporated qualitative and quantitative research and evolved their ideas and implementation based on the insights gathered throughout the process.

To shape the WhatsApp solution and WhatsApp HelpDesk the MomConnect team worked with HelpDesk operators and users, prototyping solutions and piloting them. The team integrated qualitative research alongside A/B testing to learn and iterate on different types of data at once. That research and design process informed what was developed and tested (using A/B testing), leading to the implementation of a WhatsApp service that is now integrated into the MomConnect national maternal health services at clinics nationally.

“Each of the 800,000 plus mothers on our platform do matter, to us, to their families and communities, and we always want to honor their voice in improving our service.” The MomConnect team

Understand the Landscape

The team conducted desktop research to understand the challenges and advantages of a WhatsApp solution in South Africa and how other organizations are using the technology.

Verify Desirability?

Mothers were segmented by age to measure willingness and ability to use the WhatsApp technology instead of SMS. The key barriers that needed to be addressed were identified.

The desirability of WhatsApp was verified, particularly among younger mothers. Older mothers perceived WhatsApp as unreliable, complaining that without data you don’t get your messages. Even though 56% of the existing audience had a WhatsApp account, this insight meant that mothers couldn’t be switched over automatically.

Due to data limits, the team learned it was essential to build a self-service functionality that would allow users to switch between SMS and WhatsApp.

Understand User Perspective

Primary qualitative research was conducted with 20 mothers to thoroughly understand their perceptions of WhatsApp and how they use it. The team also drew on insights about mothers’ contexts and health behaviors from previous primary research with 900 mothers, 40 health workers, 12 maternal health experts. “Our primary and secondary research resulted in the creation of personas and an insights report that guided all decision-making during the project life-cycle.” (MomConnect Team)

Cost Impact

A team of visiting MBA students began modeling the potential cost reduction that could be achieved by moving mothers over to WhatsApp over the next 5 years. Confident they were on the right track, it was then time to prototype the WhatsApp service.

Prototype Design

In 2017, a new registration process was designed that allowed mothers to choose whether to receive their messages over SMS or WhatsApp, understanding was still needed as to how this worked in the context of clinics.

A/B Testing

Hundreds of mothers at eleven selected high-volume clinics were involved in A/B testing variations of the new registration process.

Ethnography

Ethnographies were conducted in clinics to understand how to improve the customer journey and deepen understanding of the quantitative data. After a month, 394 mothers had tested the WhatsApp service.

Address Unanticipated Challenges

WhatsApp users were sending on average 4.6 messages per user, while SMS users were sending 1.8.

The team conducted interviews with the three Helpdesk operators to understand how the increased volume affected their productivity.

User Journeys

In 2018, as the program evolved, implementers tested prototypes and created user journeys for all the aspects of a minimum viable WhatsApp service, which was text only. This included WhatsApp registration, a self-service feature that allows users to switch to and from WhatsApp, and the WhatsApp switch campaign for existing users.

Uptake Measurement

The registration service was built and launched to all 3,500 national clinics. The WhatsApp switch campaign was run with all 800,000 active mothers, encouraging them to switch to receive their messages via WhatsApp. A switching self-service portal was launched to ensure mothers are always in control of how they receive their messages.

Ongoing Iteration

The design process didn’t stop after implementation and the team continues to monitor user data and iterate on the WhatsApp Service.

The Role of Measurement

The NDOH monitors MomConnect through their information system. Since each pregnant woman is linked to a facility, the government can calculate coverage rates by comparing registration data with total antenatal attendance data for facilities, districts and provinces. Monthly reports are sent to provincial and district MCH coordinators. Registration has increased over time from 40% of all facilities in 2014 to 95% in 2017. The University of Stellenbosch evaluated MomConnect, linking MomConnect data with indicators collected through routine data in the government health information system and conducted a qualitative assessment of pregnant women’s views on MomConnect and self-reported changes in attitude and behavior.

Documentation of MomConnect’s performance is found in editorials, publications and some specially designed evaluations. One study in 2018 noted that in a survey of 2,000 women, 98% found the messages helpful, 77% felt better prepared for delivery, 81% shared their messages with family and friends, and 70% wanted more messages per week. Another publication noted iron supplement stock outages by clients that stimulated facilities to address the gap in essential medicines. A third study, in a very small sample of clinics, of the influence of MomConnect on the use of health services found no difference between MomConnect users and the control population using MomConnect. However, the feedback from MomConnect users regarding their specific experiences are encouraging. Users indicated they were happy with the message frequency. They said the information within the messages was useful and allowed them to make better decisions about their health and the health of their infants.

WHO (2016) reports that “MomConnect was rapidly developed and taken to scale without the testing and development stages recommended for digital and behavioral health interventions.” Extensive measurement and learning were not integrated into program development, although design research, user testing, and other forms of measurement as noted above were used to inform adaptations and the transition to WhatsApp. Having reached scale, in 2018 the MomConnect team have now defined an extensive monitoring and evaluation (M&E) strategy to fill this data gap, inform future adaptations and assess the impact of the intervention. The strategy includes theories of change mapping pathways to program objectives and assumptions that inform those pathways as well as data sources and indicators to measure medium and long-term program effect. Interest and investment from funders in measurement of the effectiveness of MomConnect is framed in the context of sustaining the intervention over time and its potential for replication in other settings.

Currently, the research team is busy doing micro surveys to test behavioral, design and content elements and tracking response rates. They are exploring ways to integrate measurement, design and program content and learning. Pieces of the 2018 M&E plan are being executed and funding for additional measurement work is secured including use of key performance indicators (KPI) and dashboards; a responsive feedback system; and an impact evaluation testing the hypothesis that mobile phone messaging influences women’s behavior and health outcomes. Measurement is guided by Praekelt’s own organizational goals and focus as well as the ethical and procedural guidelines that are required by the DOH 2.

Finally, there are no formal studies of the influence of design on MomConnect though there is interest in understanding how integrating design has made a difference. In the early stages, there is limited documentation of the influence of design from an output perspective. However, the team is reflecting on questions that are important for their future practice such as: when does design make sense and the importance of using design to understand motivation, not just the barriers to behavior change.

2Barron, P., Peter, J., LeFevre A.E., Sebidi, J., Bekker, M., Allen, R., Parsons, A.N., Benjamin, P., Pillay, Y. Mobile health messaging service and helpdesk for South African mothers (MomConnect): history, successes and challenges. BMJ Global Health. 2018; 3 (S2): 1-6. DOI: 10.1136/bmjgh-2017-000559

Team Structure and Dynamics


The service design process was implemented by a multidisciplinary team of service designers, content designers, data scientists, engineers and project managers.

Geography and Reach

Two and a half million pregnant women, newborns and infants in South Africa with evidence of scaling to other countries in the region

Objectives

Improving the health of over 2.5 million pregnant women, newborns and infants nationally

Outcomes

Costs have been reduced significantly through the introduction of WhatsApp, in 2019 they are USD 60,000 per month (reduced from USD 100,000 in 2016) even though the growth of MomConnect exceeds 50,000 new mothers per month (60% on WhatsApp) and has extended the duration up until the baby’s second year.

When WhatsApp was introduced in 2017, mothers were 6.7x more likely to reach out than on SMS, and in 2019 engagement on WhatsApp is 14% higher than SMS. Ongoing research is exploring how to optimise this engagement for improved health outcomes.

Resources

https://www.praekelt.org/momconnect

http://www.kznhealth.gov.za/Momconnect/Lines.pdf

http://www.health.gov.za/index.php/mom-connect

Engelhard M., Copley C., Watson J., Pillay, Y., Barron, P., LeFevre A.E. Optimising mHealth helpdesk responsiveness in South Africa: towards automated message triage. BMJ Global Health 2018;3(S2):1-9. DOI: 10.1136/bmjgh-2017-000567

Heekes, A., Tiffin, N., Dane, P., Mutemaringa, T., Smith, M., Zinyakatira, N., Barron, P., Seebregts, C., Boulle, A. Self-enrolment antenatal health promotion data as an adjunct to maternal clinical information systems in the Western Cape Province of South Africa. BMJ Global Health 2018;3(S2): 1-9. DOI:10.1136/bmjgh-2017-000565.

LeFevre, A.E., Dane, P., Copley, C.J., Pienaar, C., Parsons, A.N, Engelhard, M., Woods, D., Bekker, M., Benjamin, P., Pillay, Y., Barron, P., Seegbregts, C.J., Mohan, D. Unpacking the performance of a mobile health information messaging program for mothers (MomConnect) in South Africa: evidence on program reach and messaging exposure. BMJ Global Health 2018;3:(S2). DOI:10.1136/bmjgh-2017-000583

Mehl, G.L., Tamrat, T., Bhardwaj, S., Blaschke, S., Labrique, A. Digital health vision: could MomConnect provide a pragmatic starting point for achieving universal health coverage in South Africa and elsewhere? BMJ Global 2018;3(S2):1-5. DOI:10.1136/bmjgh-2017-000626

Peter J., Benjamin P., LeFevre A.E., Barron, P., Pillay, Y. Taking digital health innovation to scale in South Africa: ten lessons from MomConnect. BMJ Glob Health 2018; 3(S2):1-4. DOI:10.1136/ bmjgh-2017-000592.

Peter, J. Achieving scale, sustainability and impact: a donor perspective on a mobile health messaging service and help desk (MomConnect) for South African mothers. 2018; 3 (S2): 1-3. DOI: 10.1136/bmjgh-2017-000562

Seebregts, C., Dane, P., Parsons, A.N., Fogwill, T., Rogers, D., Bekker, M., Shaw, V., Barron, P. Designing for scale: optimising the health information system architecture for mobile maternal health messaging in South Africa (MomConnect). BMJ Glob Health 2018;3 (S2). DOI:10.1136/bmjgh-2017-000563

Skinner D., Delobelle, P., Pappin, M., Pieterse, D., Esterhuizen, T.M., Barron, P., Dudley, L. User assessments and the use of information from MomConnect, a mobile phone text-based information service, by pregnant women and new mothers in South Africa. BMJ Global Health 2018;3(S2): 1-6. DOI: 10.1136/bmjgh-2017-000561.

Xiong, K., Kamunyori, J., Sebidi, J. The MomConnect helpdesk: how an interactive mobile messaging programme is used by mothers in South Africa. BMJ Global Health 2018;3(S2): 1-7. DOI: 10.1136/bmjgh-2017-000578.