August 28, 2023
SINGAPORE – Singapore is pushing for new technology and artificial intelligence (AI) to achieve good, or even better, healthcare with lower manpower or cost.
This is critical as the rapidly ageing population will increase demand for healthcare, says Singapore’s chief health scientist Tan Chorh Chuan.
In one pilot, the Institute of Mental Health (IMH) is using mobile phones and wearables to identify patients with schizophrenia who are more likely to suffer a relapse, so that intervention can be started early.
Professor Tan said: “We now have the ability to get data on so many aspects that we can bring together to drive population health, public health, hospital care, in a way that was not possible before.”
Digital phenotyping uses digital devices such as smartphones and watches to provide continuous glucose monitoring for diabetics, or even assess the mental health of patients – all with approval from those being monitored.
In an exclusive interview with The Straits Times, Prof Tan said that the proliferation of tools today, for uses ranging from genetic sequencing to monitoring of social behaviour through wearables, makes it possible to diagnose problems and intervene early, and to increase compliance with treatment.
“We now have the ability to capture really high-quality data that was not possible before,” he said.
For example, how people use their mobile phones – the number of messages sent each day, the way they type those messages, and even how often they leave home – can be used to reflect their mental well-being.
When someone is agitated, for instance, the cadence and the number of mistakes made when typing a message is different, Prof Tan added.
On IMH’s pilot plan, senior consultant Jimmy Lee said: “We aim to get a small group of about 40 persons in recovery to try the platform and provide us feedback to improve it. When it’s ready, we plan to enrol a larger group of individuals to examine its effectiveness.”
This follows an earlier study to design an algorithm to identify those at risk.
Schizophrenia, which causes delusions, hallucinations, disorganised speech, trouble with thinking and lack of motivation, usually manifests in one’s late teens or early 20s. It can be controlled, but not cured.
An IMH study released in 2021 showed that 0.86 per cent of the population suffer from the condition. That works out to more than 25,000 people.
Dr Lee said: “In the early phase, the illness can be quite well controlled. Relapse prevention is very important to reduce the risk of illness progression.”
Helping case managers identify patients who are on the verge of a relapse so that they can intervene could make a major impact on disease progression and allow the patients to continue living successfully in the community.
Dr Lee said that currently, about 30 per cent of people with schizophrenia suffer from a more severe condition, clinically known as treatment-resistant schizophrenia. “They experience persistent psychiatric symptoms and impairment in functioning,” he added.
Prof Tan said this pilot uses seven parameters – sleep, physical activity, heart rate, social indexes, changes in mobility patterns, keyboard tapping speeds and ambient light – to develop a workable algorithm to monitor possible relapses in schizophrenia patients.
If this works well, it can be expanded to include other mental conditions such as depression, he added.
The areas that Singapore is looking at first are those where there is a high disease burden, more people are affected or where interventions could be very effective, said Prof Tan.
One such condition is diabetes, which affects about one in 10 people here. As diabetics are at significantly higher risk of going blind, regular eye checks for diabetic retinopathy is important as spotting and treating it early can save their eyesight.
But such screening is manpower-intensive. After a picture of the eye is taken, it has to be read by two different people. If they give different results, it has to be looked at again.
To address this issue, researchers at National University of Singapore School of Computing and the Singapore Eye Research Institute (Seri) invented a deep-learning AI software called Selena+ that is capable of reading eye images. It was rolled out in July 2021 in 20 polyclinics.
It flags abnormalities in the eyes, such as a microscopic bulge in a blood vessel in the retina, which could indicate diabetic retinopathy.
Associate Professor Daniel Ting, who heads AI and digital innovations at Seri, said the analyser is in full implementation, and it has halved the manpower needed.
It reads the images of 134,800 people screened each year at polyclinics, hospitals and general practice clinics.
The quality of AI in reading images of the eye for diabetic retinopathy meets the Ministry of Health requirements for both sensitivity – the ability to correctly identify patients with a disease – and specificity – the ability to correctly identify those without the disease.
Prof Ting said that as Selena+ is designed to be more sensitive in picking up positive cases, it might include those who do not have the problem.
This is why the final implementation model will be hybrid, with a human reviewing all positive cases and making speedy referrals for those in more urgent need of treatment. Human readers would also audit 10 per cent of cases that the analyser says are normal.
Prof Ting said switching to the semi-automated model would result in 20 per cent cost savings.
“By 2050, Singapore is projected to have one million people with diabetes. At this time, the estimated annual (cost) savings would be $15 million.”
These are just two of several projects in the pipeline using genetics and technology to improve health outcomes.
Others include identifying people with a gene mutation that significantly raises their risk of getting a heart attack at a young age, and allowing patients to be “hospitalised” at home instead of in a hospital through remote 24-hour monitoring.
Said Prof Tan: “In the future, we will have greater ability to identify at-risk groups, high-susceptibility groups, earlier, and then intervene earlier and in a more targeted manner.
“There are also exciting new opportunities created by remarkable advances in science and technology. The ability to collect, curate and analyse large amounts of data will enable innovative new ways to improve population health and prevent and manage disease.”