Dr Ann Aerts, head of the Novartis Foundation, highlights how the COVID-19 crisis has thrust artificial intelligence in healthcare into the limelight, and makes three bold predictions for the future of the field, especially in emerging countries.


As a true believer in the power of technology to transform health, I am convinced that Artificial Intelligence (AI) is today’s defining technology and the greatest opportunity to transform health systems from being reactive to proactive, predictive, and even preventive.

This September, we launched a new report from the Broadband Commission Working Group on Data, Digital, and AI in Health, entitled Reimagining Global Health through Artificial Intelligence: The Roadmap to AI Maturity. The report aims to give low- and middle-income countries a roadmap for how to fully integrate AI into their health systems. When we started working on this report in 2019, I never imagined we would be launching it two years later in the middle of a pandemic. It turns out the pandemic underlined the importance of our findings, as the speed of innovation in digital and AI solutions has been exponential this year.

So at this time of unprecedented change, what have we learned, and what might change in the near future?


My top three learnings from 2020


1. A crisis can create a step change in technology use

Before the pandemic, in most countries telemedicine was a niche interest pursued by a few tech-obsessed health professionals. Now it is the norm in many health systems. We are seeing a big increase in reimbursement of telemedicine through health insurance – something the Novartis Foundation has called for over many years. The advantages are obvious (besides those of reducing in person visits during the pandemic): increased access to health, reductions in travel, less need to take time off work, and a reduced need for childcare. These are important in a high-income country; but if you live in a distant location in Africa for example, walking to and from the nearest clinic can take most of a day. When I was a recently-qualified doctor working in Mozambique many years ago, I saw how many more lives could have been saved if we had known which patients were most ill, so we could travel to them first. Widespread deployment of telemedicine in lower-income countries should be possible and cost-effective. People in such countries are often highly tech-savvy: for example, the number of active SIMs in Kenyan mobile phones exceeds the total population, and 70% of adults use mobile-based money transfer services.


2. AI technologies can be implemented more easily than many think

As part of the Foundation’s initiatives in Brazil, we work with the local government and University Hospital of São Paulo – along with many other partners – to help develop a national ecosystem for innovation, data science and AI in health. The arrival of the pandemic saw a rapid shift in priorities: an AI platform to help doctors diagnose and treat COVID-19, using a vast database of patient chest x-rays and CT scans from 40 hospitals in Brazil, was created in a few weeks. And between May and August 2020, it analysed over 10,000 lung images, identifying 70% of patients as positive for COVID-19. As a fast, simple and efficient tool, the government aims to develop and adapt the platform for other illnesses, such as cardiovascular diseases.


3. Pandemic preparedness requires better data – and the tools to analyse it

A Canadian start-up called BlueDot achieved fame for alerting its clients at the end of December 2019 to avoid traveling to Wuhan. The firm uses an AI-driven algorithm that scours foreign-language news reports, animal and plant disease networks, and official proclamations to warn of new disease risks. But better pandemic preparedness requires much earlier surveillance – in the case of zoonotic diseases, identifying them as soon as they make the jump from animals to humans. In a much less connected world, and with slower routes of transmission, pandemics can take decades to evolve – the first case of HIV probably occurred in the Congo in the 1920s. But in today’s world, novel viruses must be identified and isolated at the source. This means that the entire globe needs to have the capacity to spot new illnesses. My hope is that this pandemic will lead to significantly increased support for universal health coverage in all countries, including heightened funding for data collection and data science capability building, to enable more people to analyse and use data for better decision making. This would not only help us prevent the next pandemic, but also better address the unprecedented global health challenges around the world, such as the burgeoning burden of chronic diseases such as cardiovascular disease and cancer, the rapid urbanization and the enormous shortage of skilled health workers.

So we have seen really big changes in 2020, and strong advances in the use of data and digital in health and care delivery, but what might happen next year?


My top three predictions for 2021


1. Innovation born of the pandemic is here to stay

Some might think this is self-evident, but many innovations have failed to take off as one would expect. When home video conferencing tools were in their infancy, poor sound and image disappointed people using them. In the years that followed, internet connections improved massively and by December of 2019, home video conferencing had become more popular with Zoom – the market leader – logging 10m video calls a day. Most quality issues had been resolved, yet it took the pandemic for most of the world to realize this. In April 2020, daily Zoom usage had risen to 300m, and virtual working and meeting has become the new reality. Similarly, having experienced remote consultations, many patients will not want to return to sitting in a doctor’s reception room. I believe the example of São Paulo using AI to diagnose COVID-19 will only be a beginning and that collaboration will soon be extended to other disease areas.


2. Lower-income countries realize they need AI and understand they need to invest in data and infrastructure, their people and workforce and the right governance and regulatory systems to increase their readiness to deploy AI in health

Our report details the case study of Rwanda, where a company called Babyl offers remote consultations with AI-enabled triage and consultation support. Their subscription rates have increased to the point where a third of Rwanda’s adult population uses the system to get better access to the limited number of health providers. In a country where one doctor may serve as many as 60,000 people in rural areas, shortages of trained health professionals puts tremendous pressure on the health system. Being able to get prescriptions and referrals through your phone has proved to be an immeasurable benefit – Babyl works with over 450 health clinics and pharmacies to make sure it has a strong footprint in the real world. Other African countries are taking note and beginning to follow. Post-pandemic, this type of service may start expanding across Africa, and could be a model for all countries around the world.

The other post-pandemic trend will hopefully be increased funding for health – once recovery from the economic impact of COVID-19 starts. AI can help allocate resources in the most rational way, identifying where funding will have the greatest benefit. Rational health funding has often been a problem in high-income countries, with money not necessarily allocated on the basis of real-time evidence – leading for instance, to the perennial underfunding of prevention. Any mechanisms that lower-income countries can put in place to stop this happening would mean better population health over the long term.


3. Unlocking private sector innovation in health AI for lower-income countries

While overall, everyone assumes much of the AI in Health innovation happened in high-income countries, our report showcased that AI has already had significant impact in lower-income countries too. However, many entrepreneurs in low- and middle-income countries lack the networks and support to commercialize their ideas. To help with this, the Novartis Foundation worked with the Harvard Global Health Institute and the Massachusetts Institute of Technology to set up the inaugural Data Science & AI Summit (DASH) in Africa. The Summit, which was set to take place in Kigali, turned into a virtual platform due to the pandemic, delivering a series of 10 webinars. DASH aims to support the growing African community of data scientists, entrepreneurs and health workers who are exploring how AI can improve health outcomes in Africa. Here participants can learn how to transition great ideas into real-world solutions; to connect across disciplines, sectors and geographies; and to build a repository of evidence and lessons learned that can support innovation.

In Brazil, the Foundation is supporting an initiative by the government of São Paulo to tap into private sector innovation. The new government innovation platform Ideia Gov has launched public calls for proposals for innovative solutions to address its most pressing health challenges. Its first calls sought proposals for better monitoring of vital signs, affordable and replicable COVID-19 diagnostic tests, and the AI solution for supporting COVID-19 diagnosis. The evidence is that there is no shortage of good ideas, but the platform, networks and funding needed to make them happen are often not there. At the Foundation, we try to spot such gaps and provide the support needed to help the development of the HealthTech ecosystem.

In conclusion, AI allows us to reimagine how we deliver health and care, improve health outcomes, and accelerate the global push towards universal health coverage. From global pandemics to health worker shortages, the world is facing growing challenges that call for the extraordinary capabilities that only AI can offer.