Senior Principal Specialist Dr. Goh Han Leong (left) and Principal Specialist Dr. Fu Zhiyan (right), as part of Synapxe’s Data aNalytics & Ai (DnA) team (Image credit: Chen Lai Fu)
There are more than 400,000 patients with diabetes in Singapore, and chronic kidney disease (CKD) is one of the most common complications. Synapxe, Singapore’s national healthtech agency under the Ministry of Health, has developed an artificial intelligence (AI) system to estimate the risk of worsening of chronic kidney disease in patients with diabetes in the next five years. This serves to help doctors better provide advice to patients, including early intervention treatment and disease control.
Diabetes is one of the common causes of CKD. CKD will gradually cause patients to lose kidney function, but its symptoms are not obvious in the early stages. Once CKD develops into end stage renal failure, dialysis or kidney transplantation is required to sustain life. Data shows that up to two-thirds of renal failure is caused by diabetes.
CKD is divided into five stages. The first nationwide CKD prediction model for diabetic patients developed by Synapxe uses data from patients with stage 1 and 2 CKD in the national diabetes database. It analyses whether these patients would progress to stage 3 and above, with more serious conditions, in the future.
The AI prediction model will analyse various data, including the patient's age, sex, medication records and blood test results, and classify the patient into 3 categories – namely low risk, high risk, and extremely high risk.
Dr. Fu Zhiyan, Principal Specialist at Synapxe’s Data aNalytics & Ai (DnA) team, said during the interview that the prediction model has been integrated into the National Electronic Health Record (NEHR). Doctors can view their patient's risk category from the system and provide further advice. Generally, low-risk patients only need to be continuously monitored. Dr. Goh Han Leong, Senior Principal Specialist at Synapxe’s DnA team, highlighted that in the process of data collection and model design, they communicated with clinicians consistently to obtain feedback and improved the model. This is to ensure that the model can be used effectively as "gaining the confidence of doctors is crucial".
According to both specialists involved in the study, this research aims to aid doctors with making medical decisions. Data has shown that early intervention in medical treatment can help to slow down the deterioration of CKD in patients.
The clinical model, which has been endorsed by the Ministry of Health (MOH), is part of Singapore’s Holistic Approach in Lowering and Tracking Chronic Kidney Disease (HALT-CKD) programme.
In 2017, MOH launched a comprehensive programme to delay kidney disease. Through methods such as conducting kidney function tests and consuming drugs to protect kidney function, it allows patients with proper illness management to slow down the deterioration of CKD and reduce the probability of renal failure.
Dr. Fu Zhiyan said that at present, this AI prediction model can only predict whether CKD will deteriorate further. At the same time, Synapxe is studying how to use this type of AI disease prediction model for other chronic diseases.
Research results on the AI prediction model have also been published in academic journal Frontiers back in January this year.
Early prediction of chronic diseases and measures to prevent them
Dr Alan Wong, a family doctor at Vitacare Family Clinic, said that with ageing population, general practitioners are seeing more patients with chronic diseases in their daily work, and one common chronic disease would be CKD. He said that AI tools that can help identify or predict diseases can help family doctors like him who provide medical services on the front line with decision-making during consultations and provide guidance for treatment planning.
Jacob Ho, 54, chose to retire early 3 years ago from his high-pressure sales job to focus on treating his CKD. Ho developed diabetes in 1999 and was diagnosed with CKD in 2019. At that time, it became clear to him that he had to make a drastic change in his lifestyle to prevent his condition from deteriorating.
"I started watching my diet, reading nutrition labels, and avoiding foods high in sodium and calcium. I also started exercising for at least an hour per day and looked for ways to reduce stress."
Ho's health condition gradually improved, but he highlighted that had he known that diabetes would lead to a high risk of CKD, he would have made changes from the beginning.
"For those who are going through the same situation, don't give up, keep working hard, nothing is difficult if you set your mind to it."
