Researchers have developed artificial intelligence (AI) capable of identifying women suffering from post-traumatic stress disorder related to childbirth.
- Researchers have developed an artificial intelligence (AI) tool capable of identifying women with PTSD linked to traumatic childbirth.
- The researchers used an OpenAI model, known as text-embeddings-ada-002, which converted women’s birth stories for the AI to analyze.
- Ultimately, this could make it possible to more quickly screen women at risk of PTSD linked to traumatic childbirth.
25 to 34% of women describe their childbirth as traumatic, depending on the Prevention and Treatment of Traumatic Childbirth (PATCH). This means that the birth of the child turned into a very difficult ordeal – due to different causes: pain, emergency cesarean section, fetal distress, gynecological violence, feeling of being close to death, etc. – and that mothers have significant after-effects.
These women can, for example, develop post-traumatic stress disorder (PTSD). THE Reims university hospital center estimates that 3 to 6% of mothers are affected after giving birth. In a new study, published in the journal Scientific Reportsresearchers have therefore developed an artificial intelligence (AI) tool capable of identifying women suffering from PTSD linked to traumatic childbirth.
An AI to diagnose post-traumatic stress disorder linked to childbirth
“Current assessment of PTSD related to traumatic childbirth currently relies on extensive clinical assessment”, explains Sharon Dekel, author of this study, in a communicated. A process that is too long and does not meet the urgent needs of women. AI could therefore be a quick and, according to the authors, inexpensive strategy for making a diagnosis.
During their work, the scientists used 1,295 short childbirth stories, told by women with or without childbirth-related post-traumatic stress disorder (CB-PTSD).
Then, a model OpenAI known as text-embeddings-ada-002 converted this narrative data so that the machine learning algorithm, the AI developed by the team of researchers, could identify women with CB-PTSD.
PTSD: better screening of mothers at risk of suicide
Results: The AI identified women with childbirth-related post-traumatic stress disorder with 85% accuracy, and those without with 75% accuracy.
“The model we developed could perhaps improve accessibility to screening and diagnosis of CB-PTSD by integrating seamlessly into routine obstetric care“, says Sharon Dekel.
Early screening of women who may be suffering from post-traumatic stress disorder after childbirth could have important and life-saving consequences. Indeed, according to a Public Health France survey carried out published in 2021, suicide is, after heart problems, the second cause of maternal mortality. Of the 262 maternal deaths (up to one year after childbirth) identified between 2013 and 2015 in France, 13.4% were caused by suicide.
