Artificial intelligence algorithms can predict patient death accuracy by up to 90%

A research team at Stanford University uses artificial intelligence algorithms to predict patient death, hoping to improve the timing of hospice care in critically ill patients. In the test, this system proved to be very accurate, correctly predicting the death outcome in 90% of cases. However, although the system can predict when a patient may die, it still cannot tell the doctor how it came to a conclusion.

Predicting death is very difficult. Doctors must consider a range of complex factors, from the patient's age and family history to the response to the drug, as well as the nature of the disease itself. To complicate matters, doctors must struggle with their own conceit, prejudice or unconsciousness to assess how much the patient has. Sometimes doctors can accurately predict, but sometimes patients may delay for months (if not years), whether it is premature or late to predict death, is not conducive to hospice care.

This brings problems to the precise arrangement of hospice care. Often, when a patient is unlikely to live for more than a year, treatment is transferred to a hospice group who tries to keep the patient as far away as possible from the pain in the last few days or months. To this end, they strive to manage the patient's pain, nausea, appetite and confusion, provide psychological and spiritual support, while respecting the social, cultural and spiritual needs of patients and their families.

But if a patient transitions to hospice care too late, they are likely to miss this important stage of care. If it is too early, it will bring unnecessary pressure on the medical system.

Artificial intelligence algorithms can predict patient mortality with an accuracy rate of up to 90%

Often, advanced disease can turn into a medical crisis and patients will eventually be in the intensive care unit. There, the situation has developed wildly, leading to more and more interventions, often unable to provide good services to patients and their families,” Ken Jung, a medical research scientist at Stanford University and co-author of the new study, told Gizmodo. The hospice team One of the goals is to have a conversation with the patient so that the patient can think carefully and express their preferences before the crisis. It is worth noting that even if the patient does not die in the future, these behaviors are appropriate. The purpose of hospice care is to allow patients to benefit from these conversations.

Jung said that this unmet need was first discovered decades ago, and the survey showed that 80% of Americans want to die at home, but only 35% do so. He said that the situation has improved, but we "have a long way to go."

In China, according to the data, about 7 million people go to the end of life every year, but the hospice care provided by the society can only meet about 15% of the demand.

The report of the British "Economist" Information Department also believes that the supply of hospice care services in China cannot keep up with the aging of the population. According to the 2015 "Death Quality Index" of the Economist Intelligence Unit, China is in 80 countries. Ranked 10th in the bottom.

The timing is just right, which is why Stanford University Anand AvaTI and his team developed an AI-based system. The death prediction algorithm is not intended to replace the doctor, but rather provides a tool to improve the accuracy of the prediction. In addition to improving the timing of hospice care, the system can alleviate the burden on doctors in predicting patient outcomes, a time-consuming and laborious process.

Artificial intelligence algorithms can predict patient mortality with an accuracy rate of up to 90%

“The problem we have to solve is that only a small percentage of patients who can benefit from hospice actually accept it, in part because they are found too late, partly because of the shortage of [human resources] in hospice care services. Early detection.” Avati told Gizmodo, “We are trying to solve this problem.”

The system uses an artificial intelligence called "deep learning," in which neural networks learn from large amounts of data. In this case, the system is based on data from the Electronic Health Record (EHR) of adult and child patients, either Stanford Hospital or Lucille Packard Children's Hospital. After analyzing 2 million records, the researchers identified 200,000 patients for this program. Researchers are "agnostic" about the type of disease, stage of disease, degree of admission (intensive care unit, and non-intensive care unit). All of these patients have relevant case reports, including diagnosis, number of scans ordered, type of procedure performed, number of days of hospital stay, medication used, and other factors.

USB Car Charger

The usb car charger can plug into your car's cigarette lighter jack to become a powered USB port to fit all electronic appliances with a USB charging Cable. It is very convenient to use in the car when you go outside for travel or business.


Yidashun offers dual usb car charger 5V 2.4A and 5V 3.1A, it is aluminum case with good-looking, and it can be as a hammer to break the window of the car when in emergency.


Yidashun also offers plastic usb mobile car charger 5V 4.8A, and same one with extension usb-c or lighting cable, it is high AMP with fast-charging of your device. And one more usb Car Charger makes you convenient; it is with 6-port and 5-port usb to charge 6 and 5 usb devices at the same time.

usb carcharger

usb car charger

USB Car Charger,Car Charger,Car Phone Charger,Car Charger Adapter

Shenzhen Yidashun Technology Co., Ltd. , https://www.ydsadapter.com

This entry was posted in on