Associate professors Bee Yong Mong (far left) and Rinkoo Dalan worked on the project using HealthVector Diabetes to predict the risk of chronic kidney disease onset in diabetes patients over three years. ST PHOTO: SHINTARO TAY
The Straits Times (5 August 2024)
Doctors from two healthcare clusters here will soon be using digital twin technology to detect and manage chronic kidney disease (CKD) due to diabetes, a condition that could lead to kidney failure if left untreated.
A digital twin is a virtual counterpart of an individual based on his or her biological data.
This virtual replica of the patient’s condition can be used to simulate the progress and probable direction of a chronic disease, thus allowing doctors to make better decisions.
The technology will be trialled under a pilot programme in early 2025 at the Singapore General Hospital (SGH), Tan Tock Seng Hospital (TTSH) and selected polyclinics under Singapore Health Services (SingHealth) and National Healthcare Group (NHG) for potential clinical adoption at no extra cost to patients.
It has already received regulatory approval from the Health Sciences Authority.
Digital-twin tech is expected to help Singapore move away from having the third-highest incidence of treated end-stage CKD due to diabetes globally, said Associate Professor Bee Yong Mong, who heads the department of endocrinology at SGH.
According to the 2023 annual report by the US Renal Data System, Brunei in 2021 topped the list of countries with the highest rates of CKD due to diabetes, with 370 per million population (pmp). Jalisco, a state in Mexico, followed in second place with 297 pmp and Singapore placed third at 250 pmp.
Diabetes is the No. 1 cause of kidney failure in Singapore, accounting for 67 per cent of new cases.
The number of people with kidney failure here has gone up by 40 per cent over the past 10 years. Currently, more than 470,000 adults in the Republic live with diabetes, and about half of them have CKD.
With the number of Singapore adults with diabetes projected to double to one million by 2050, doctors here want to be able to detect CKD early to manage it effectively.
“When you look at (patients with) Stage 5 or end-stage kidney failure among all the dialysis patients, about 70 per cent have diabetes as the underlying contributing factor.
“When they reach Stage 5, the lifespan is shortened by up to 16 years,” Prof Bee said.
“In terms of the trend from 2010 to 2019, we see a large increase in the prevalence of end-stage kidney disease. I think part of this is because our diabetes patients are surviving longer with better treatments,” he added.
Dr Andrew Wu, chief executive of Mesh Bio, a health technology start-up focusing on the digital transformation of care delivery, said: “We believe this technology will be instrumental in reducing the burden of diabetes complications and advancing personalised care strategies.”
Using data from two hospitals and selected polyclinics under SingHealth and NHG, Mesh Bio developed HealthVector Diabetes, the world’s first foundational digital twin model of human biology.
The metabolic digital twin software uses data analysis and artificial intelligence to predict the risk of CKD onset in diabetes patients over three years.
At present, patients with diabetes need to be monitored for diabetic kidney disease through yearly blood and urine tests, as early kidney disease usually does not have any symptoms. Testing is the only way to know how well the kidneys are working.
Data of 7,000 patients from SGH, TTSH and selected polyclinics was assessed to derive a risk score to predict worsened kidney function.
Using information such as fasting blood sugar levels, cholesterol measurements, body mass index and blood pressure, the software estimates the three-year risk for CKD development in Type 2 diabetes patients who do not have pre-existing kidney issues.
Associate Professor Rinkoo Dalan, an endocrinologist at TTSH, a member of NHG, said that work done using HealthVector Diabetes has shown that the predictions of the software outperformed other prediction models such as the annual blood and urine tests.
The study into the development and validity of HealthVector Diabetes was published in npj Digital Medicine journal in May 2024.
“(From there), we developed this algorithm further and subsequently fine-tuned and validated it by adding the SingHealth diabetes registry,” Prof Bee said.
Prof Rinkoo said the improved software can potentially transform the care of patients with diabetes, as doctors use the information to prioritise and intensify treatments to delay the progression of kidney failure.
Once the pilot programme is launched in early 2025, Prof Rinkoo said its predictions will allow doctors to create personalised treatment for sugar and blood pressure control, which will in turn delay the onset or progression of CKD and its complications.
“This way, patients with diabetes will have an additional 15 years free from having to deal with end-stage kidney failure, and will have a better quality of life,” she added.