AI helps hospitals treat patients

Xu Lingchao
Hospitals in Shanghai are using artificial intelligence technology to manage the health of patients.
Xu Lingchao

Hospitals in Shanghai are using artificial intelligence technology to help manage the health of patients.

Some of these technologies will be showcased for the first time at the 2019 World Artificial Intelligence Conference to be held in the city next month.

The Shanghai Cancer Center of Fudan University admitted more than 1.44 million patients last year, 70 percent of whom came from other cities.

Wu Jiong, deputy director of the cancer center, said many patients didn’t know which department to register for.

“We thought if we could analyze their condition before registration and find the right expert, it would solve the problem,” Wu said.

The cancer center worked with digital giant Tencent and developed an intelligent registration system.

Patients fill in their personal information in a mini-program the hospital developed in WeChat, and the data is sent to the hospital’s AI engine.

After evaluation, the system makes an appointment with the right specialist.

For example, a patient with an urgent need for thyroid surgery will see a specialist, whereas a patient who needs a check will see a general doctor.

The hospital launched the service in March last year. 

So far, 83 specialists from 11 departments have been using it.

The time a patient has to wait before seeing a specialist has also been cut down by 7.4 days.

Another problem with the large number of patients is that doctors get tired, especially those working in the radiology department who must examine hundreds of x-rays and other images.

Both the cancer center and Shanghai No. 10 People’s Hospital looked to AI technology for a solution.

“It takes a doctor, senior or junior, at least five minutes to examine a CT scan of a patient’s lung to detect nodules,” said Tang Guangping, head of the radiology department of the hospital. “Sometimes there are so many it can take more than 30 minutes.”

Tang has lost count of how many images he has to look at every day.

“A single missed nodule could lead to a fatal tragedy,” said Tang.

Last year, an AI system was developed, which can detect possible nodules in an instant.

“The only problem now is that the system is too sensitive,” said Qiu Yuyou, a doctor from the hospital’s radiology department. “More than often the ‘nodules’ are blood veins.”

Qiu said their job then became to eliminate wrong detections.

“But it’s better a false alarm than an actual nodule being missed,” said Qiu.

The hospital has also deployed a patient first contact mini diagnosis lab in its emergency lobby. 

A patient sits in the egg-shaped machine which carries out basic preview checks for such things as blood pressure and body temperature.

Yesterday, a 24-year-old man surnamed Wang felt chest pain and started to sweat. He went to the No. 10 hospital and a nurse guided him to the machine. The machine asked some simple questions such as where does it hurt and how long has it lasted.

This information was relayed to a doctor who was prepared for the consultation with Wang.

Peng Hu, head of the hospital’s emergency treatment department, said the hospital is planning to showcase these technologies at the AI conference so that more people will come to understand them.

“The cost of these technologies is quite high,” said Peng. “The more people know about the artificial intelligent medical system, the more likely it is for us to further promote it.”

Special Reports