Researchers have developed an algorithm that can predict whether people with trauma will develop post-traumatic stress disorder, the same as PTSD. The algorithm, which makes the diagnosis based on the collected medical data, will assist medical personnel in the early detection and treatment of the disorder.
According to health officials, 10-15% of traumatized patients suffer from long-term symptoms of PTSD, and they mostly recover within a year of the injury.
Although there is a method that can effectively reduce the risk of PTSD, strategies for its prevention in its early stages have not been implemented because there is no proven way to predict which patients are at high risk.
Too many biological or psychological signs – be it high-stress hormones, increased inflammatory signals, high blood pressure, or hyperexcitability – often precede PTSD. However, none of them were completely reliable for a preliminary diagnosis of PTSD.
In this new study, in which several teams of researchers participated, machine learning was used to make a preliminary diagnosis to create an algorithm that would assess the risk of PTSD from a combination of 70 clinical data.
The researchers created the algorithm from data from 377 adult traumatized people and tested the results in 221 adult traumatized adults in New York.
Among people who were put at risk for PTSD by the algorithm, 90% showed symptoms lasting a year. However, 29% of people who did not put the algorithm at risk also showed prolonged symptoms of PTSD (this is an indicator of a false negative outcome).
Various tests are required before introducing the equipment into reality. In future studies, the team plans to test how well the algorithm can detect PTSD in patients experiencing potentially traumatic health problems.