What if artificial intelligence could help doctors better manage depression? This would make it possible to better diagnose this disease and to plan treatments. The details of Martin Ducret.
Martin Ducret, doctor and journalist at Doctor’s Daily, talks to us today about AI and depression. What if artificial intelligence could help doctors in their diagnosis and management of depression? This is the hope aroused by a recent French study published in the journal Scientific reports.
franceinfo: What does this study tell us?
Martin Ducret: The researchers of this study, piloted by a Parisian medical technology start-up, followed around sixty variables – sleep, physical activity, heart rate and breathing to name a few – which were recorded daily for 3 months, in 26 patients with depression.
Then, an artificial intelligence analyzed all these physiological data, and determined for each patient a specific profile, called a biosignature. And each biosignature made it possible to accurately predict the evolution of depression over the following 3 months, with an accuracy of 90%.
What prospects do these results open up?
For Professor Pierre Geoffroy, co-author of the study, psychiatrist at Bichat Hospital and at the ChronoS center at GHU Paris Psychiatry Neurosciences, “CThese results, still at an intermediate stage, show that artificial intelligence, by identifying individualized profiles of patients suffering from depression, could help doctors improve their care, in a much more personalized way.”
Depression is very variable from one patient to another, hence the importance of personalizing its care?
Yes, absolutely, depression is a disease that presents different clinical forms. Postpartum depression (which occurs after pregnancy), differs from seasonal depression or melancholic depression. And for the same type of depression, the evolution of the disease, such as the risk of suicide or relapse, the response to antidepressants and to psychotherapy are not comparable from one patient to another.
To date there is no objective criterion to evaluate all these parameters, it is based only on a set of subjective criteria, collected during exchanges between the doctor and the patient in consultation. “Therefore, having precise and individual biomarkers, analyzed by an artificial intelligence, would make it possible to better diagnose this disease, to predict the responses to treatments and its evolution”explains Professor Geoffroy.
Although further work is needed to validate the results of this study, artificial intelligence-assisted medicine seems very promising to better manage depression, a disease that affects 1 in 5 people during their lifetime. .