Hong Kong health care and hospitals

Hospital Authority, University of Hong Kong team up for big data project to help treat stroke victims 

The project, which began last August, involves sharing anonymous data of 7,000 patients admitted to public hospitals in 2016

PUBLISHED : Tuesday, 08 May, 2018, 10:32pm
UPDATED : Wednesday, 09 May, 2018, 10:28am

The Hospital Authority has teamed up with University of Hong Kong scientists on a pioneering big data analysis project that uses patient records to help doctors better predict severe strokes and prioritise treatment, it was revealed on Tuesday.

The project, which began last August, involves sharing the anonymous data of 7,000 patients admitted to public hospitals in 2016.

One of the leaders of the project, Dr Gilberto Leung Ka-kit, a professor with the university’s Li Ka Shing Faculty of Medicine, said a quick diagnosis was crucial for victims of acute ischemic stroke because the “golden window” for treatment is six hours.

There are 7,000 acute ischemic strokes every year in the city, including 900 caused by large vessel occlusion (LVO), the most severe form of acute ischemic stroke. The death rate of patients with LVO within 30 days is 40 per cent.

To speed up assessment and help doctors decide which patients to treat first, researchers analysed patient data, including Computed Tomography (CT) brain scan images provided by the Hospital Authority, to build an assessment model.

How dementia is caused by mini strokes, how to detect it early and nine ways to reduce your risk

Doctors currently use CT scan images and other data of individual patients to analyse risk.

The data of 300 patients was used in the first phase of the study while the data of 6,700 more could be used in the second phase. 

Another researcher, Professor Philip Yu Leung-ho, estimated that when the tool is ready, it would take doctors just 20 seconds to predict LVO.

“It can help doctors to identify LVO and decide which patients should be treated with priority,” Yu said.

The model, which Leung said would take a few years to be applied in hospitals, was able to identify 95 patients out of 100 with LVO in the testing period.

Eva Tsui Lai-hing, chief manager of statistics and workforce planning with the Hospital Authority, said it might do more extensive research on patients in the second phase, to test the model.

While the university conducted the first phase of the research for free, Leung said he hoped the city would provide funding for the second phase, which is expected to cost millions of dollars.

Meanwhile, Polytechnic University opened its University Research Facility in Big Data Analytics (UBDA) on Tuesday.

Will ‘zero waiting time’ make a difference for Hong Kong security guard who lost ability to speak after a stroke?

According to professor Cao Jiannong, who directs the facility, the HK$50 million (US$6.37 million) project will provide analysis to PolyU departments as well as other institutions.

Cao said UBDA had already partnered with private companies such as GoGoVan and Huawei, as well as public bodies such as the Hong Kong police force to fight cybercrime.

“The police want to know how to identify cybercrimes and fake accounts on social networking sites,” Cao said.

The facility would help van service provider GoGoVan predict demand under different weather conditions and at different times to get to customers quicker, he said.

Cao said the facility would in the near future focus research on health care, finance and smart city design.