Vishal Doshi looks at how AI is already revolutionising healthcare in Asia-Pacific and the work still to be done on the foundational elements – infrastructure, regulations, and the availability of Asia-specific data – to ensure AI can be effectively utilised to improve patient outcomes in the region.
Healthcare, manufacturing, finance, education – what do all these sectors have in common? Over the last couple of decades, these industries – and more – have benefitted from the advent of artificial intelligence (AI).
For us here in Asia to reap the benefits of AI, there is a significant need to increase the availability of Asian-specific data
In healthcare, AI has made great strides particularly in aiding diagnosis for rare and unknown conditions. In the Asia Pacific region, AI is also being tested in the management of health conditions. For example, AI Singapore, a national program driven by government-wide partnerships, granted SGD$35 million in research funds to three multidisciplinary teams to pioneer innovative AI solutions to lower the risk of diabetes, high cholesterol and high blood pressure progression and complication development. Initiatives such as this signal the greater proliferation and acceptance of AI as part of the spectrum of solutions to address healthcare needs.
Another element in healthcare where AI has been strongly utilised is the further integration of AI into drug discovery and clinical trial designs in oncology. AI’s potential to track patterns will enable us to identify toxicity signals, potential combinatorial treatments as well as indication expansion models. This not only helps to speed up clinical trial times but can also identify the best possible treatment and outcomes for patients.
AI has certainly made an impact by turning data into actionable insights, allowing us to solve complex, intricate problems. However, for AI to deliver what it’s capable of in healthcare we must be better focused on creating a stronger platform for its potential to thrive. In turn, we can make smarter decisions quickly and bring better outcomes to more patients at the right time.
For drug development specifically, these foundational elements include a better generation of diverse and bias-free Asia-representative data, as well as the improved agility of our infrastructure and regulatory landscape to adapt and scale-up.
AI is only as good as the data it gets
The use of AI is largely driven by the availability of Big Data. Healthcare, being a highly regulated industry, is a data-rich industry. However, for us here in Asia to reap the benefits of AI, there is a significant need to increase the availability of Asian-specific data. At the moment, existing genetics and clinical trial databases are mainly made up of Caucasian data. Our first step in overcoming this less than simple feat is recruiting Asian patients to participate in any studies or clinical trials. Recognising that this involves massive recruitment efforts given the vastly dispersed geography and patient accessibility in Asia, our efforts are best focused on offline and in-person recruitment drives.
Here at AUM Biosciences, together with mandating biomarkers for the selection of molecules for development, Asian-specific data is the central piece of the puzzle to make our Asia-focused solutions a reality.
It’s also important to understand how the AI algorithm works. In each and every situation, there should be clarity on the kind of data that is being fed to the algorithm, otherwise, we risk what is known as algorithm bias. This phenomenon occurs because data is being fed and created by humans, who have conscious or subconscious biases. Without a systematic or collective review of data that is being fed into the algorithm, these biases may go undiscovered and potentially affect the outcomes.
Agility to regulate and innovate
As we reach more individuals to contribute data sets, and the volume of data continues to increase, we will need to scale up rapidly. This means developing a next-generation infrastructure that will allow us to harness such large quantities of data to derive actionable insights.
Although this is a relatively new area for us here in the Asia-Pacific region, our regulatory systems will also need to approach these new databases and systems with speed and agility. A great example would be the Intellectual Property Office of Singapore (IPOS) offering fast-track AI-related patient applications in a push to further develop into a digital economy. The application will now take as little as six months to be processed rather than the usual two to four years, making Singapore the fastest country in the world for such procedures.
While AI is meant to reduce, if not eliminate the margin of human errors, disparities and incorrect diagnoses can still occur. Accountability thus becomes another area for regulators to consider – who is responsible for the information produced by the algorithm? If there is no trust in the technology, social adoption could prove to be a spanner in the works of furthering the cause of AI. Regulatory sandboxes that allow companies to test innovations in a controlled environment and under regulatory supervision is one initiative that could help to enhance trust in new technologies.
As the saying goes, “You accomplish victory step-by-step, not by leaps and bounds”. By taking steps to ensure our foundational elements – infrastructure, regulations, and the availability of Asian data – are solid, only then will we be able to elevate the role of AI in drug development and bring affordable cancer therapies to patients in Asia.