This AI is able to predict the future psychological needs of cancer patients

This AI is able to predict the future psychological needs of cancer patients
This AI is able to predict the future psychological needs of cancer patients

Scientists at the University of British Columbia have developed artificial intelligence (AI) capable of predicting the future psychosocial and psychiatric needs of cancer patients.

Cancer is a very common tumor. This disease is the second leading cause of death per year worldwide. In France, it is estimated that 3.8 million people are living with a cancer diagnosis. In addition to the heavy treatments and physical consequences that this tumor causes, cancers of all types, can also cause depression and anxiety in patients. However, these psychological aspects are not trivial. They can influence the recovery of patients. To address this, researchers at the University of British Columbia and the cancer center have developed AI to help detect future psychological support needs. Their research was made public in the scientific journal Nature.

An AI to predict psychosocial and psychiatric needs

Faced with illness, some patients may feel psychological distress, which can greatly influence their recovery. But thanks to this new AI, oncologists will be able to predict whether or not Patients will need to see a specialist within a year. This technology would even be capable to identify the type of care they will need: psychosocial or psychiatric. “AI could prompt the oncologist to talk to their patient about psychiatric services or trigger sending patients an email with a personalized list of available services,” said lead researcher Dr. John-Jose Nunez, a psychiatrist and clinical researcher at the University of British Columbia’s Mood Disorders Centre.

How does this AI work?

But how is this possible? Generally, in a treatment pathway to detect cancer, patients meet an oncologist. The latter, during a first meeting, draws up a consultation report in which is mentioned, among other things, medical and psychosocial history and treatment options to follow. It is this document which will therefore be analyzed by AI using “natural language processing”. This intelligence will therefore scan the content of these reports and will detect clues relating to the psychology of the patient such as: alcohol or drug consumption habits, type of cancer or previous psychological disorders. By storing all this information, this device will be able to predict the psychological or psychiatric monitoring needs of patients. This technique has already proven itself, particularly in the analysis and summary of medical reports. The device was able to understand how words are linked to each other using a system of so-called “convolutional” neural networks.

70% success and improvements in progress

Implausible, isn’t it? To develop this AI, scientists carried out tests on 47,625 patients from the British Columbia cancer center, whose medical data remained anonymous. These experiments showed an efficiency of more than 70%. However, the AI ​​missed a number of referral opportunities to a psychologist.
Scientists continue to improve the precision and efficiency of this device and see these results as “l“AI’s enormous potential to essentially act as a personal assistant for oncologists, helping identify mental health needs earlier and ensuring more patients get the support they need.” They also plan to extend the use of this method to other areas of health in which psychosocial factors have a significant role.

You may also be interested in:

⋙ How artificial intelligence tools can help us predict future pandemics

⋙ That’s it, AI can predict your death!

⋙ We could detect the signs of cancer several years before it appears according to this study

⋙ Cancer: how to detect it with a blood test?



PREV Overconsumption of salt increases the risk of stomach cancer by 41%
NEXT great benefits in screening undiagnosed cases and educating patients