IT Brief Asia - Technology news for CIOs & IT decision-makers
Story image
Microsoft is helping to reshape healthcare in Asia with AI
Tue, 24th Jan 2023
FYI, this story is more than a year old

Artificial Intelligence (AI) and Machine Learning (ML) are becoming integral parts of various industries as the world continues to evolve technologically. One industry already seeing practical applications of such technologies is healthcare.

AI is helping to revolutionize multiple health systems in Asia, including clinical intelligence, drug discovery and population health.

Microsoft is becoming an integral partner with health systems in countries across Asia, helping to evolve how health care is approached.

One example can be found at the China Medical University Hospital (CMUH) in Taiwan. In collaboration with Microsoft's Azure Machine Learning platform, the hospital developed and deployed an "intelligent antimicrobial system". The system helps detect drug-resistant bacteria. 

In a real-world application of the system, a patient with an infection required a lab test to identify what caused it. The test was sent to the lab for processing and was received three days later, where it identified a drug-resistant bacteria. However, this test was made redundant because the AI modelling identified the same bacteria within an hour.

This is just one example of AI being used practically in the hospital. Evolutions in the way wait times in healthcare are reduced come very rarely, but since CMUH has been using the model, patient mortality has fallen by 25%.

CMUH has developed and deployed hundreds of models hosted on Microsoft's Azure cloud platform, and they are used every day across the systems' 12 hospitals. The custom-built AI models are helping to diagnose diseases such as cancer and Parkinson's, treat stroke and heart patients quicker, and ease the paperwork load on staff.

"The major goals of every AI tool we develop are to save patients' lives and save doctors' time," says Dr. Kai-Cheng Hsu, Director of CMUH's Artificial Intelligence Center for Medical Diagnosis.

"Our vision is to be a world-class platform provider for the AI hospital of the future."

The foundations for the use of AI in the hospital have been around since 2015, with the hospital launching the AI Center for Medical Diagnosis launching in 2017. 

Now, ten different AI clinics operate at CMUH. The clinics involve collaboration between doctors and AI researchers on a monthly basis, where ideas for future models are discussed, and implemented models are reviewed. 

Eight of the models developed in practice have received regulatory approval from Taiwan's Food and Drug  Administration, and eight more have been submitted for review. All of the models have been developed using Microsoft's Azure Machine Learning platform.

"Azure is at the core of our digital healthcare transformation," says Dr. Jiaxin Yu, Director of CMUH's AI Innovation Center. 

"The platform lets us quickly develop new AI tools, deploy them into the clinic  and improve the daily practice of our physicians." 

Part of the accessibility of the service is the integration of the models into software that doctors use daily. This can even be, in some cases, the application of a model through the push of a button.

The models are also being used to treat heart attack patients. CMUH developed an algorithm that can detect the likelihood of a heart attack based on a patient's ECGs and send a message to an on-duty specialist for confirmation. This model has been used for two years at CMUH, and has cut in half the time between when a patient arrives and when they are treated. 

The algorithm was so successful that CMUH partnered with a medical device manufacturer to bring it into ambulances as an early-warning system.

It is important to acknowledge that not all of the models developed have had similar successes.

A model designed to predict sepsis in patients was using organ failure, which is not always a sign of sepsis, to correlate it with the illness. Yu says that the team could train future models with additional labels to help it better distinguish between cases that do or do not have the illness.

Importantly, Hsu and Yu explain that these models are only being used to help support the decisions which medical experts best make by providing deeper insights into data.

"We want to deploy our AI models to hospitals and clinics around the world, and it is much more efficient and cost-effective to use a cloud platform," says Yu. 

"Patient privacy is always a top concern, and a critical reason we chose Azure is because it can ensure the safety of sensitive health data. Azure really reduces the barriers to bringing AI into the hospital."