How data analytics is assisting with the fight against malaria

Despite long-term efforts to eradicate malaria, the disease still remains one of the top 10 killers in low-income countries. A huge 40% of the world’s population is at risk of the illness and, according to World Health Organization estimates, there were about 216 million cases worldwide in 2016 – the most recent figures available – 90% of which were located in Africa. This number does represent a drop in incidence rates of about 18% globally though, from 76 per 1,000 people in at-risk populations in 2010 to 63 cases six years later.

But despite this improvement, a huge 445,000 individuals were still estimated to have died from the disease during 2016, 91% of whom were, once again, living in Africa.

Another point to bear in mind in this context is that, while malaria tends to be associated with mosquitos, the insects themselves are merely carriers of a parasite that causes the illness, the most prevalent types in Africa being plasmodium falciparum. Therefore, preventative measures have traditionally focused on stopping mosquitos from biting people, for example by providing them with mosquito nets and spraying their homes with insect-killing chemicals.

But what has really helped numbers of malaria victims to “crash” over the last 15 years, and even to completely eliminate the disease in some countries such as Sri Lanka, has been the addition of other new tools, according to Jeff Bernson, senior director of results management, measurement and learning at PATH, a not-for-profit global public health organisation funded by the Bill & Melinda Gates Foundation. He explains:

As well as preventative methods, we now have better diagnostics that allow health workers to do a blood test for the disease in under 20 minutes by means of a simple pinprick on the finger. There are also more effective drugs to drive the parasite out of the human gut. So these things together have been a game-changer.

Even in places like Zambia, which is still considered a malaria-endemic country, there have been “huge advances”, Bernson says. While the risk of infection is still high in some areas, rates are now below 5% in the Southern Province, at which point it becomes about “starting to chase individual cases and create areas that are malaria-free”, he adds.

Major breakthrough

Put another way, an 80% drop in malaria cases has been witnessed in the south over the last three years, accompanied by a 90% fall in mortality rates. This breakthrough has been achieved in a combination of ways.

On the one hand, the number of community health clinics in the region has been expanded threefold to more than 600 in order to make them more accessible, while the 1,500 health workers assigned to them have also been trained to test and treat the condition locally. On the other hand, there has been much more effective use of the data that health workers are tasked with collecting and entering into a dedicated district health information management system. Bernson explains:

We have dashboards so we can see weekly and monthly data and the frequency of infection rates. It’s quite powerful as it’s allowed us to do more targeting around hotspots in order to reduce the burden of the disease. Having that information helps us plan and mobilise resources, which are limited in Zambia, so it’s all about directing things to where they’re most needed. Data is the lifeblood here, especially during the rainy season when transmission takes place and you have flare-ups.

As to how the data is kept clean, he says this is down to efforts of the Minister of Health Chitalu Chilufya who recognised from the outset that his Ministry was “sitting on a goldmine of data and information”. Bernson says:

There are a number of practical measures that district teams take such as asking ‘has this facility or those health workers reported on time and was the data complete?’ It’s actually a really good proxy as to whether quality is good or not. We also do data quality audits by sending out a team to sample information at different health facilities and check if what was written down in the register is the same as what was reported electronically.

PATH’s role, meanwhile, has put together an informal consortium of partners to work on its ‘Visualize No Malaria’ campaign. This campaign aims to help Zambia meet its ambitious goal of being the first large sub-Saharan African country to fully eradicate local transmission of the disease by 2021 – although the fact that it is land-locked makes it difficult to promise that the illness will not be brought in from outside.

Leapfrog technology

Members of the consortium consist of volunteers from tech vendors such as Exasol, which provides its in-memory analytics database to process the data for data modelling purposes, the Tableau Foundation, which contributes data visualisation tools and Mapbox, which enables the production of custom maps. Bernson says:

It’s leapfrog technology because it’s not just about how people collect and store data. It’s enabling individuals without a degree in computer science to grab and manipulate information in real-time so they can create data products and dashboards and infographics with confidence. This enables them to communicate what’s happening, ask why and also plan more effectively.

In his opinion, what has been achieved so far is a “stunning example” of what can happen when people and organisations come together to work towards a common aim. He continues:

Over the last two years, we’ve been training members of district health management teams to use the tools and they’re are now getting better access to data and are asking better questions as a result. They’re not just producing product requirements for use cases but are producing their own dashboards and are now increasingly looking at using the technology for other diseases.

The ‘Visualize No Malaria’ campaign has already started operating in the capital Lusaka and Zambia’s Central Province and is now expanding into the Western Province too. Bernson says:

Our focus this year is on how to scale up to ensure the tools get into the hands of more district health management teams. So we have a training cohort that is working on it and they’ll start to become a centre of excellence around data use that will, over time, extend beyond malaria. We’re also trying to create a mentor/mentee network and group of trainers at the national and sub-national as we need local talent to sustain things.

Another goal for the future is to start taking advantage of artificial intelligence technology in order to predict areas where malaria is likely to flare up again to ensure resources are mobilised there in advance.

My take

It is amazing what can be achieved when people choose to work together towards a common goal and are given the right tools to do so. But the fact that the ‘Visualize No Malaria’ campaign’s model is repeatable across Zambia and other malaria-endemic countries in sub-Saharan Africa such as Botswana, Mozambique and Swaziland has to be a major plus too.

Image credit – wikipedia