COVID-19’s global impact has accelerated the adoption of artificial intelligence in healthcare, with MedTech companies and governments laying the groundwork for a data-driven future. Whether this will translate into a timely prediction of future outbreaks, better diagnosis, faster drug development and efficient care is the billion-dollar question.
By Jisha Krishnan
Nine days before the World Health Organization (WHO) released its statement alerting people to the emergence of a novel coronavirus outbreak, a startup based in Toronto, Canada, had identified an unusual increase in pneumonia cases in Wuhan, China.
In December 2019, BlueDot – using natural language processing (NLP) and machine learning (ML) to track, locate, and report on the spread of infectious diseases – was the first to alert government agencies and public health bodies on the early signs of the impending viral outbreak.
Artificial intelligence (AI) – the simulation of human intelligence in machines that are programmed to think like humans and mimic their actions – is the buzzword. At present, AI is being used not just to predict the speed and pattern of the pandemic, but also to facilitate swift research for a vaccine and cure.
Global efforts to fuel AI adoption in healthcare have never been more urgent. From predicting future outbreaks to revolutionising care, here’s looking at the potential and pitfalls of building a data-driven healthcare system, powered by AI.
Getting on the bandwagon
When it comes to research and development in the AI space, Asia is leading the world. AI in healthcare, analysts estimate, will be worth $6.6 billion by 2021. Right from managing hospital workflows, medical imaging and diagnosis to therapy planning, virtual assistants, and drug discovery, AI is fundamentally changing healthcare as we know it.
With healthcare taking centre stage in the pandemic world, the focus today is as much epidemic prevention as it is on making optimal use of limited resources. “Microsoft Taiwan, for example, collaborated with healthcare centres to deploy AI masks and infra-red temperature detection devices to automatically scan individuals entering the premises for face masks and body temperature,” says an engineer working with a leading MedTech company.
“The challenge with most Asian markets is making these technologies affordable right away,” he admits.
However, that hasn’t deterred tech companies from harnessing the power of ML, NLP and data analytics to drive the healthcare narrative in the new world order.
In China, Alibaba’s AI algorithm claims to diagnose suspected COVID-19 cases within 20 seconds with 96 per cent accuracy, while a South Korea-based life sciences company Seegene came up with testing kits in less than three weeks, courtesy its AI expertise.
Even in Singapore, the government has developed a system that traces the movement of COVID-19 patients using facial recognition and public transportation records, so it can assess the risk of the infection spreading. Over the last few months, AI has been playing a significant role in helping countries cope with the global outbreak.
Early diagnosis
Google’s DeepMind AI platform collects data on patient symptoms to help doctors make a quick and effective diagnosis on the basis of its massive dataset for comparable symptoms. India-based healthcare analytics firm Tricog uses AI to predict a patient’s risk of heart disease and suggests personalised treatment plans based on the patient’s vital health indicators.
Today, there are algorithms to spot malignant tumours; AI-enhanced microscopes to scan for harmful bacteria in blood samples; deep neural networks used to interpret medical scans and pathology slides at high speed; CT image analytics and genome sequencing for the coronavirus. Early diagnosis can be a life-saver.
“Yet these technologies largely remain within the confines of super-speciality hospitals in urban areas. How can we make them accessible to the majority of the population in countries like India with dismal healthcare budgets and inadequate infrastructure?” asks Dr K Sujatha, a consultant physician, who has spent the better part of her life practising medicine in rural India.
“In theory, AI can be a potent tool to reduce socioeconomic health inequalities. The challenge is to translate it into practice,” she rues.
Research and drug development
Did you know that 90 per cent of the drugs never make it to the market? Given the high costs of research and development and the lengthy timelines and processes tied to discovering the drugs and taking it to market, it’s important to invest in more efficient and accurate models.
“The pandemic has opened up a first of its kind large scale, multi-country, multi-player, global collaboration on healthcare data. AI has been helping us identify a lot more patterns of disease transmission, mechanism of action, treatment and new drug and vaccines development,” says Suresh Ramu, Co-founder, Cytecare Hospitals, Bengaluru.
“Usually it would take 8-12 years to develop a vaccine for a novel virus…with global collaboration and data sciences we should have this within 1-1.5 years since the pandemic began,” he maintains.
Epidemic intelligence – spotting and assessing epidemics in their earliest stage – is an area where AI shows immense promise. Its innate ability to break down data silos and analyse complex data sets can also help streamline the development of new drugs.
Besides, researchers are already using predictive analytics to identify suitable candidates for clinical trials and helping medical practitioners to gain better insights into new treatments.
Automated operations and personalised care
In March 2016, Johns Hopkins Hospital partnered with GE to use predictive AI techniques to improve the efficiency of patient operational flow. A review found that since the implementation of the programme, there has been a 60 per cent improvement in the ease to admit new patients and 30 per cent faster bed assignments in the emergency department.
Further, there was a 21 per cent increase in patient discharges before noon, resulting in more efficient operations and positive patient experience.
In other words, AI solutions can help automate routine tasks and solve operational challenges at healthcare centres to streamline the patient experience and provide personalised care.
In many parts of the world, AI-powered chatbots are now reviewing patient symptoms and recommending healthcare solutions.
It’s not just about automation though. Take robot-assisted surgeries, for instance. They help surgeons perform complex procedures in a minimally-invasive manner, with fewer surgery-related complications, lesser pain and faster recovery time.
“Given the shortage of healthcare professionals and growing cases of burnout among doctors, smart automation tools could be the solution to a complex problem,” admits psychologist Sonali Gupta, who is quick to point out that doctors are not always good at managing their own health.
“Besides, with AI, there’s always the fear that the machines may replace us. Only if we are convinced that we have no reason to feel insecure can we actually work together and make a difference,” she notes.
Accountability and trust issues
Arguably, the biggest concerns with the increasing use of AI in healthcare is accountability and trust. When doctors rely on AI algorithms to make clinical decisions, it’s often hard to interpret them. There’s no obvious reasoning or explanation as to why a scan has led to a particular diagnosis. So, when things go wrong – as they are bound to at times – there will be difficult questions with no easy answers.
This can seriously hamper trust in the doctor-patient relationship, causing several legal and ethical dilemmas. Is the doctor liable for a diagnosis that was not made by him/her? Can there be complete transparency in such situations? Will patient privacy become more contentious?
“The medico-legal liabilities are still being passed on to the clinicians who use AI currently to support their decisions. In the future, we could have some of the liabilities being held by the software or medical equipment companies that own the AI tools,” opines Suresh of Cytecare Hospitals.
All said and done
According to the World Economic Forum, by 2030, AI will access multiple sources of data to reveal patterns in disease and aid treatment and care; healthcare systems will be able to predict an individual’s risk of certain diseases and suggest preventative measures; AI will help reduce waiting times for patients and improve efficiency in health systems.
However, for that to happen, we need the right policies and regulatory frameworks in place – an important element that is currently missing in most countries. The United Kingdom (UK) is a noteworthy exception. It was among the first to establish a multi-disciplinary, expert, independent body – the Centre for Data Ethics and Innovation – to advise the government on data-related policies and technical solutions that facilitate effective development and deployment of new technologies.
In the meanwhile, in Asia, AI technologies that can help cities control future outbreaks are becoming a central element of ambitious smart-city projects. And that brings us back to the original question: Can AI prevent the next pandemic? It can, but whether it will or not, depends on us. As someone rightly said, “The future belongs to those who prepare for it today.”
