Sweet Life is South Africa’s diabetes community: a space to share information and advice about living well with diabetes, both here on the website and on Diabetic South Africans. But we’re also involved in diabetes education research – a big focus for 2020. Each month, we’ll be sharing the research we’ve done into diabetes education in public clinics, in the hopes that it helps someone else looking for answers to diabetic questions! This is the May report, be sure to scroll to the bottom of the page for the other months.
Now that we’re settling into the ‘new normal’ it seemed like time to pick up the conversation around mHealth for diabetes: a diabetes app similar to MomConnect that could work for diabetes education in public clinics. Alongside that conversation, we have been developing the diabetes food list booklet mentioned last month, and spreading awareness of diabetes and prediabetes in the media.
Relevant diabetes research:
mHealth for diabetes
A handbook on how to implement mDiabetes. Some fascinating excerpts that apply to the South African situation in the context of mHealth for diabetes:
According to WHO, 80% of cases of diabetes, 80% of heart disease and 40% of cancer could be prevented by avoiding tobacco, increasing physical activity and adopting a healthy diet. While better management of diabetes and improving treatment and care are important, it is globally recognized that, in the case of NCDs and especially diabetes, prevention through lifestyle changes is critical and cost–effective.
Use of mHealth for diabetes at a population level could be an effective way to achieve behaviour change, especially in lower- and middle-income countries where mobile phone use is growing exponentially and transmission is inexpensive.
A population-based diabetes prevention programme should account for the wide diversity of individuals and their motivation to change their lifestyle. While the majority of a population does not have diabetes, some individuals may be interested in the programme because they know someone with diabetes; others are at high risk or have undiagnosed diabetes.
Content focus: mHealth for diabetes
Of particular interest, particularly if we look at the low health literacy levels in SA.
In writing the text messages, the following should be considered:
- the language, tone, clarity, health literacy, technological literacy;
- provision of both motivation and information;
- salience of practical tips and strategies for the population;
- tailoring for specific groups (e.g. diabetes and healthy populations, rural and urban, different socio-economic status, ethnic group, age, gender);
- whether the information should be static or dynamic, depending on whether the programme is uni- or bi-directional, fully interactive via an app or based on input from the user; and
- Length of text messages allowed in the country.
The strategy: mHealth for diabetes
- the target audience and the best means of reaching them with promotional material;
- whom the public sees as the “owner” of the programme, as the operational model may well direct the promotional campaign;
- whether the campaign can be linked to or leveraged by organizations or notable personalities currently involved in mass media campaigns for diabetes control or prevention and whether there are lessons to be learnt about effective promotional techniques;
- how people will register or sign up, e.g. directly by text message, online, by telephone, in person, or through a third party, and the amount of data to be collected at baseline to allow tailoring and to ensure follow-up;
- the local mobile network environment, e.g. whether sending unsolicited text messages is allowed (contravenes network codes of conduct in some countries), which may also be an important consideration in the potential effectiveness of or engagement in the programme: whether a population that has been desensitized to receiving unsolicited health-related text messages will be likely to read mDiabetes text messages;
- use of incentives to encourage participation;
- the effectiveness of mobile outreach for enrolment: the response rates to messages asking people to subscribe to a service are < 1% even if it is free;
- whether marketing and promotional campaigns can be leveraged by e.g. stakeholder agencies, including technical partners; and
- the timing of the campaign, perhaps with a “soft launch” before the promotional campaign, to ensure that all processes are working well before many participants sign up.
Lessons learnt from the field:
- The success of the mHealth for diabetes programme is measured by how well the population adheres to it. Factors for success include: wide media coverage (e.g. advertising on local radio and buses, leaflets distributed at points of care) and simple enrolment (e.g. a free phone number, free SMS, local contact at points of care or pharmacies or patient association office).
- In many countries, radio and TV advertising are important means of promoting an mHealth for diabetes programme but may be very expensive. Initial underestimation of the cost of promotion is common and can be difficult to remedy later.
Focus for SA mobile app: mHealth for diabetes
- The mHealth for diabetes journey: If we only give this app to people at a clinic who have just been diagnosed with Type 2 diabetes, there’s a consistent journey, rather than trying to map out every journey of every person with diabetes in SA.
- One entry path: a route to take just diagnosed diabetics and educate how their life needs to change using the TEEL model to improve their outcomes in the first 6 months.
- TEEL model (proposed mnemonic):
The scope: mHealth for diabetes
- How could we create good habits in the first 6 months of being diagnosed with diabetes using a WhatsApp chat tool?
- Core journey: what happens in the first week?
- Specific habits to develop when you’ve just been diagnosed with Type 2 diabetes at a public clinic. For example: Feeling sick because of medication – metformin makes you feel worse at first. Get into the habit of taking it at the end of the day rather, after dinner.
- Suggestions for what to eat for breakfast / lunch / dinner.
- Advice / tips for testing / eating / medication.
Praekelt call with Sanofi:
Praekelt developed MomConnect. We spoke with them in March about potentially developing DiabetesConnect, and Sanofi reached out to them in the interim. They put together a proposal for how DiabetesConnect could work.
Two streams: Diabetes on WhatsApp and upskilling MomConnect and NurseConnect
Diabetes on WhatsApp
- Type 2 diabetes patients specifically. Target the patients as they engage with healthcare workers at the public clinics, ask whether they want to register for the diabetes service on WhatsApp. Specifically Type 2 and prediabetes. Start an engagement and support service on WhatsApp. Rich media, interactive environment on a standard platform.
- Tapping into normal behaviours they have on WhatsApp, incentivise healthy behaviour change.
- Think through a menu based service. Overarching content areas that we believe might be useful for prediabetic and diabetic patients (link to the WHO Be Healthy Be Mobile programme). Also access a text based help desk which we would build with turn.io platform – tool that leverages machine learning and natural understanding. Matching FAQ with answers, include behavioural milestone journey within the platform. Goals to be achieved.
Machine learning and help desk operator
- Help desk operator human in the loop, so the answer is pulled up and they edit if necessary (more efficient). Allows the user to know that there is someone there, not just engaging with a chatbot, there’s a person on the other end during the journey.
- Ask questions for additional support – also direct to other online resources.
- Natural learning – geared up to 11 official languages. We would need to do the translations of the content into the relevant national languages. Potentially content available in top 5 most spoken languages, like COVID19 Connect.
- In terms of the content set, there is a content set available from the WHO Be Healthy Be Mobile programme. It’s SMS based, so would need to be adapted to WhatsApp. Sweet Life also has content available, so that won’t be a stumbling block.
Iterations over a year
- Two critical things: Praekelt thinks it’s really important when they deploy and launch a platform to not just leave it, to come back and see how people engage with it, iterate it and make sure it’s improved in line with how people use it.
- Also, when you’re looking at platforms like this, include monitoring and evaluation – research and data science teams to assess it over a period of 12 months.
- 3 months to build – design, build a prototype, launch.
- 12 months to monitor and evaluate, make relevant platform improvements and optimisations.
Marketing and promotion
- What is usually the uptake and how long does it take to build up on this? What is the realistic time for 1000 people to start using it?
- At a high level, we would need to make sure that there’s the relevant and right amount of marketing and promotion necessary to get the users on to the platform and make them aware that it exists and the value it can provide. For example, COVID-19 Connect launched in mid-March and has millions of users.
- Extensive recruitment from NDoH, important to be a project with NDoH. Advertised on TV, radio. To get scale and numbers, they need to be aware of it with the right amount of marketing effort. Look at different levels of penetration, number of Type 2 diabetics and clinics.
- We need to get the same kind of recruitment going as MomConnect: every time a mom comes to the clinic they’re asked: do you want to sign up?
- Around monitoring and evaluation, is there any data? How many on MomConnect now, how many have registered for the service. Regionalisation: do you have any data regarding that that we can use as a benchmark?
- 3 million users that have registered since 2014. 1 million active every month.
- M&E hasn’t been through an impact evaluation, but data science teams have done high level. 10,000 messages a day. 3 help desk operators based at NDoH, managing that number.
- COVID19 Connect: 7 million users, 500,000 a day.
- Automated as much as possible, only one help desk operator.
- Really useful guidelines, what are people asking about that we haven’t addressed in the content on the platform, how should we be prioritising content going forward? Content shouldn’t just be static that you leave and let it run. Give NDoH a good feeling for the type of questions that have been asked.
- What happens after the first 18 months when Praekelt finishes their monitoring?
- Long term sustainability. If we ensure that it’s an NDoH programme, that ensures the longevity of the programme as well.
- Diabetes is a long-term engagement, adherence is important. That’s something that we will bear in mind in terms of content that we’ll produce for the platform and the type of service that it flows and the design of the programme.
- Long-term who’s in charge: help desk operator employed by NDoH. Dedicated help desk operator would be necessary for DiabetesConnect, ideally sit with the NurseConnect and MomConnect operators. Praekelt provides strategic and technical support, message flows, design, content, service design.
- Who’s in charge of making sure it doesn’t stagnate? A programmatic lead would be necessary. Different streams coming together to support different aspects.
- Convening power of WHO.
How to approach NDoH
- NDoH has programmatic leads for NCDs. Ability to sell anything to them has to be inclusive – diabetes, hypertension, oncology.
- Itumeleng suggested diabetes education, but if we could encompass hypertension as well and pitch it a bit higher that would be helpful.
- This is a pilot of the first segment, add hypertension and other NCDs as a bolt-on. To package it as a holistic programme and start with diabetes. Stages of incremental value.
- Step back and look at the overarching vision, which we’re aligned with as an organisation. NDoH want to expand to other diseases and other citizens and expand this more broadly, and to NCDs more broadly. Make sure it’s aligned with their vision.
- To get NDoH buy in we need a prototype / demo model. Pitch to Diabetes Alliance and then get Diabetes Alliance to pitch to NDoH. Get input and then see if other stakeholders want to contribute.
- Diabetes Alliance approach SEMDSA and get their buy-in.
- WHO Be Healthy Be Mobile initial discussions with Praekelt about a WhatsApp diabetes project in SA. Test case for their toolkit on WhatsApp. Need to see where they see their involvement.
Strengthen MomConnect and NurseConnect around diabetes
- The other stream is trying to strengthen the help desk as is – MomConnect (gestational diabetes) and NurseConnect (diabetes). Look at what exists and build a content bank of diabetes content and train the nurses in that content.
- Isn’t that an easier start? However, not everyone on MomConnect has diabetes, a very small segment of the diabetes platform.
- Still need a standalone diabetes service for patients on WhatsApp.
Diabetes food list booklet
As mentioned last month, food education is a critical part of diabetes education. We are in the process of developing a very simple food list booklet that will illustrate portion sizes and recommended food to choose. Aimed at public clinic patients. Our illustrator has been illustrating almost 100 kinds of food, and we are currently moving into the design phase, with the goal to print 15,000 copies at the end of June / beginning of July. We will then iterate based on feedback from diabetic Pick n Pay employees and other core groups, with the aim of doing a massive print run in November for National Diabetes Month.
Diabetes in the media
One of the Sweet Life mandates for 2020 was to get diabetes in the media more, not just in November but year-round. We put together a prediabetes press release which resulted in an interview on Channel Africa Radio and a live TV interview on ENCA on Sunday 24th May, as well as various online exposure. The press release can be found online here, a video of the TV interview is here.
Next steps: June
- Further develop DiabetesConnect concept with Praekelt and pitch it to Diabetes Alliance.
- Design diabetes food list booklet.
The 2020 Diabetes Education Project by Sweet Life is supported by Pick n Pay and BD.