How can AI and data take us into the next age of healthcare?

By Shirley Tay

Experts from Singapore and New South Wales discussed at a recent GovInsider webinar.

It was in 1923 that American surgeon Elliot Cutler carried out the world’s first successful heart valve surgery. Since then, cardiac surgeons have made breakthroughs — researchers recently discovered a way for AI to determine the exact valve shape that best would fit a patient’s heart.

Technologies such as AI and data analytics have unleashed a realm of possibilities — from remote monitoring to predictive healthcare. How can professionals best harness these tools to address challenges during the pandemic and beyond?

Five leading healthcare experts came together at a GovInsider webinar last week to discuss the opportunities in the next frontier of healthcare.

Accessible remote care

“Remote patient monitoring and virtual care is here to stay,” said Ian Schrader, Program Director of the Virtual Care Accelerator at eHealth New South Wales. In its most recent budget, the state allocated AU$500 million (US$379 million) to digital health initiatives, including virtual care and telehealth.

The agency had already been running remote monitoring for patients with diabetes and cardiac rehabilitation, but ramped up virtual care when the pandemic hit. It monitored the recovery of Covid-positive patients and vital signs like blood oxygen saturation levels, Schrader said.

To make virtual care accessible, eHealth NSW is loaning devices and is looking into subsiding internet costs for patients who may not be able to afford them.

For Singapore’s National Healthcare Group, primary care teleconsultations ballooned and hit about 40,000 a month at the peak of the pandemic, said Thomas Lew, its Group Chief Data & Strategy Officer.

But he emphasised that “the end-to-end process is as important, if not more important, than the consult itself”. Healthcare providers need to ensure patients can also receive medication quickly and access a pharmacist immediately after the consultation, for instance.

While virtual care adoption was high in the initial stages of the pandemic, both Schrader from eHealth NSW and Singapore’s National University Health System (NUHS) found that adoption rates dropped as time passed.

Healthcare is a high-touch industry, with every patient having different needs, said Ngiam Kee Yuan, Group Chief Technology Officer of NUHS. He believes this is an opportunity for hospitals to extend the same level of hospital care to patients at home, especially with the help of better technology and equipment.

eHealth NSW taking a human-centric design process to ensure patients are part of the virtual care journey, Schrader shared. “We're trying to make sure that we get those key foundations right and focus on the things that we can do well, and then start to build on those.”

AI in healthcare

NUHS has spent “a number of years” developing AI tools, said Ngiam. It’s testing the use of AI to predict the risk of patients falling in a ward, and machine learning to automate the characterisation of thyroid lumps.

It has learnt that there is “no space for legacy systems”, he added. Healthcare providers will have to build cloud platforms that allow them to “scale from one to 100 AI tools” to reap the technology’s full potential.

The MOH Office for Healthcare Transformation (MOHT) has built a tool that helps patients with multiple chronic conditions decide which risk factors to address first - such as to lose weight or lower their sugar intake, said Praveen Deorani, Senior Data Scientist.

“If you tell them to improve everything at the same time, it may get quite overwhelming. So that personalised insight for each individual patient, based on this AI tool, is quite powerful.”

In encouraging clinicians to adopt AI tools, Deorani believes AI should be seen as an assistant, not as a replacement. The technology can help to speed up MRI imaging and radiologists can come in to verify the results, for instance.

Clinicians are driven by evidence and are not always adverse to AI tools, said Ngiam from NUHS. AI tools need to be sufficiently tested, and should not be “accepted to be standard of care unless it passes the test of evidence, as with all other medical devices” .

Protect patient privacy

As healthcare providers adopt more technology, data privacy and security is paramount.

Healthcare providers are dealing with many different types of data sources today, from wearables to patient health records, said Cindy Maike, Vice President, Industry Solutions & Value Management at Cloudera.

They need to ensure data is not compromised and understand who’s done what with the data. Techniques such as data anonymisation and tokenisation, which turns sensitive data into a string of characters, will be helpful, she added.

Lew from NHG said providers need to establish a level of transparency and explicit consent - this should not be deemed or implied. Hospitals must be able to justify the right to access patient information, and be very specific about how they are trying to help patients.

Establishing this trust and relationship is crucial as Singapore needs to move from a “healthcare model to a wellness model”, he added. Data will help authorities understand what motivates citizens to keep well, and the relevant behaviour modifiers that will encourage them to maintain their health.

Remote care and the use of AI in healthcare will be here to stay. Healthcare professionals will need to ensure care remains accessible and secure for citizens.

Watch the full replay of the webinar here.