Unleashing the power of alternative data to transform India’s battle against malaria

Too many malaria cases, too little data

Today, India is fighting one of its toughest battles against malaria. With a malaria prevalence rate 1.5 times greater than that of Southeast Asia, India accounts for over 67% of malaria cases in the region.

What’s even more worrying, however, is that this deadly disease has plagued the country for years. India launched its first national malaria vector control program more than 60 years ago, but malaria still runs rampant today. Why?

Well, a quick look at rural India’s health infrastructure gives you all the answers you need. Health workers are missing in the hospitals that need them most. On-ground disease surveillance is almost non-existent. Mosquito nets lie dusty and unused in village households. Pools of water, an ideal breeding ground for mosquitoes, stagnate in cooking pots and communal tanks. Meanwhile, malaria runs rampant, free and uninhibited.


Where are these problems greatest? Where is malaria most common? The truth is that India just doesn’t know. The country’s disease data is outdated and hard to extract and, more importantly, drastically incomplete. The WHO estimates that approximately 12 million cases of malaria in India were unreported in 2017. This means that India’s fight against malaria is worse than the current data suggests.

Without the data they need, how can governments and organizations possibly prepare for these outbreaks?

Building a granular malaria prediction model

Imagine this. What if organizations had real-time predictions of when and where malaria was going to break out in India? What if regional governments knew, in advance, which sub-districts in a state were at the highest risk of experiencing malaria outbreaks? The national fight against malaria would be revolutionized. For the first time in history, India would be one step ahead of the game in its fight against malaria.

To help solve this problem, we decided to build a real-time, granular malaria prediction model. After all, national health stakeholders face several challenges in their fight against malaria. A challenge they definitely shouldn’t face is a lack of data.


Malaria is complex — how and where it spreads is determined by weather, geography, people’s actions, healthcare infrastructure, and countless other variables. However, most malaria models use just a couple variables, usually temperature and rain.

Instead, at Atlan, we use 500 variables to pinpoint the sub-districts where malaria is most likely to break out with less than 5% error. Our model uses alternative data from 14 data sources, combining environmental factors, such as temperature, rainfall and humidity, with geographical factors, such as population density, economic status and healthcare infrastructure. This data is updated in real time and is sourced from PDF files, web pages and Excel sheets.

To test the accuracy of our model, we compared its predictions to the number of actual cases in Orissa for January and February 2018. The error in both predictions was less than 5%. This predictive power can revolutionize the fight against malaria.

We were shocked by two key data points from the model:

  • Only 3 states (Chhattisgarh, Gujarat, and Odisha) account for 55% of India’s malaria cases
  • 5% of India’s districts account for almost 50% of its malaria cases

Let that sink in for a moment.

Only 5% of India’s districts account for almost 50% of its malaria cases. That means that just 30 districts account for over 50,000 cases of malaria. It also means that India’s malaria problem is highly concentrated in some parts of the country.

While these insights are scary, they’re also a shining ray of hope. Since malaria cases are so concentrated, efforts to prevent malaria can be geographically targeted, and even small efforts can have a huge impact. With targeted interventions by key stakeholders in a few volatile regions, we can drastically reduce potential malaria cases and deaths.


The Indian government: target drives, collect better data, and send critical malaria alerts

The Indian government is invested in the fight against malaria. From supporting a national vector control program to launching a national plan to eliminate malaria in India by 2030, the government has taken significant steps to reduce malaria in India.

However, for its efforts to bear fruit, the government needs to answer critical questions like:

  1. Which sub-districts are most likely to experience malaria outbreaks in the future?
  2. Where does health care infrastructure need to be developed most urgently?
  3. How should health budgets be allocated to reduce high disease burden?

With high-quality granular data, government leaders can analyze past trends and gauge future forecasts to answer these questions. They can use our prediction model to:

  1. Target insecticide spraying drives: Spraying insecticides on the walls and ceilings of homes is one of the most important ways to prevent malaria. However, these indoor residual spraying (IRS) drives often don’t reach remote villages in hilly areas. This can be changed by targeting spraying drives according to risk of future outbreaks. Spraying professionals and ASHA workers can be assigned to impacted regions to ensure that the right number of people are allocated based on need.
  2. Issue national travel warning alerts: Reducing the spread of malaria is an important step in preventing major outbreaks. By issuing national travel warning alerts based on accurate real-time data, the government can ensure that outbreaks remain contained and aren’t transmitted to high-risk groups like women and children.
  3. Increase disease surveillance: Today, a majority of malaria cases go unreported and undiagnosed. If the government knows the regions that are most vulnerable to outbreaks, they can increase disease surveillance (the practice of monitoring the spread of disease through data) in those areas. This will also help in increasing prompt early case detection, which is crucial for the successful treatment of malaria.
Photo by Marc Shandro on Flickr

FMCG businesses: keep insecticides stocked and customers healthy

The role that healthcare businesses play in combating vector-borne diseases, such as malaria, can not be overstated. As new pockets of malaria outbreaks rapidly emerge across the nation, businesses need to ensure that insecticides are available in the right place at the right time.

To avoid shortages so people can protect themselves against mosquitos, businesses need to optimize their supply chains and distribution networks by asking hard questions:

  1. Where should new healthcare products be launched to combat outbreaks effectively?
  2. How can distribution networks and supply chains be expanded to reach vulnerable areas?
  3. Where are the insecticide shortages and how can they be prevented?

FMCG companies can use our prediction model to spot trends and patterns in the frequency and severity of malaria outbreaks across the country. They can use the data to distribute resources efficiently and:

  1. Prevent insecticide shortages: In India, news of mosquito repellant shortages are as frequent as news of malaria outbreaks. The gaps in coverage are glaring and businesses need to ensure that demand is being met when and where it arises. The WHO recommends the use of mosquito nets and insecticide sprays to control transmission. One way businesses can prevent shortages of these prevention tools is by using monthly predictions of outbreaks to drive supply chains.
  2. Increase product distribution to high-burden rural areas: While creating long-term product distribution strategies, businesses need to ensure that locations of new distribution centers align with regions of high disease burden. There are over six hundred thousand villages in India that are spread across millions of kilometers with low physical connectivity. By targeting specific regions with high rates of transmission first, businesses can expand into rural areas effectively.
  3. Ensure consumers at risk are safe and healthy: When using media sources and other communication strategies for advertising, healthcare companies can play a key role in driving behavior change and ensuring rural communities take safety precautions specially during high-risk seasons like the monsoons. Businesses can adapt national campaigns to resonate with specific rural communities and help in the larger fight in reducing malaria transmission.
Image by Pippa Ranger/Department for International Development

Philanthropies: focus awareness drives and behavioral campaigns in the right places

The easiest way to eliminate malaria is to prevent it. In fact, last year on World Malaria Day, the WHO placed a special focus on closing gaps in accessing malaria prevention tools.

For philanthropists and donors committed to preventing malaria, it is important to raise awareness in regions where knowledge gaps exist. As a result, they need to answer hard questions like:

  1. Where should targeted awareness campaigns be run to maximize impact?
  2. Which population groups are at the highest risk of contracting malaria in the future?
  3. How can strategic advocacy and communication be used to prevent malaria outbreaks?

By using our malaria prediction model, philanthropies can make targeted investments to:

  1. Improve awareness drives: The spread of malaria is highly dependent on environmental factors, and its transmission varies around the year. Philanthropies can target the frequency, intensity and locations of awareness drives to the times and places that most need them. Additionally, they can target schools and other communal groups to ensure all community members are being reached.
  2. Focus on high-risk groups: By focusing on vulnerable groups, philanthropies can reduce malaria for groups with the highest risk of disease contraction. Apart from focusing on traditional high-risk groups like young children and pregnant women, donors can use our data to find other low-immunity, high-risk groups. For example, our model shows a correlation between high malaria prevalence and high percentages of tribal populations. Such groups can be targeted during medical interventions.
  3. Influence behavioral change: One of the hardest problems healthcare organizations face in rural areas is driving behavioral change. For example, despite having mosquito nets, many villagers do not sleep under the nets during the monsoons. Additionally, many people store water in pots, coolers and tanks, and they don’t regularly clean these areas to prevent mosquitoes from breeding in them. Philanthropies can target their communication strategies in such areas to encourage malaria prevention practices.

All these use cases make it hard to deny that the secret weapon to combating malaria will be targeted outbreak predictions. So, what are you waiting for? Join a revolutionary group of philanthropists and business leaders who are unleashing the power of our malaria prediction model today!

Photo by Julie Johnson on Unsplash


We are reimagining the human experience of working with data. Join us on our mission of helping data teams around the world do their lives’ best work! Check out our open positions.

Write A Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.