Two studies have shown the value of AI tools for helping to predict mental health conditions in women during the vulnerable time after they have given birth. The findings could be used to create tailored care and target prevention strategies or resources towards women who might be at particularly high risk of mental health conditions postpartum.
The first machine-learning model could help predict who might be at risk of postpartum depression at hospital discharge.
The other AI tool calculates the familial contribution to the risk of postpartum psychosis and shows that women whose sisters have experienced this are at particularly elevated risk.
Both sets of findings are published in the American Journal of Psychiatry conjunction with Maternal Mental Health Awareness Month and coincide with the 2025 Annual Meeting of the American Psychiatric Association.
“Maintaining maternal mental health and well-being is critical for not only mothers’ health but also for their babies’ normal social, emotional, and physical development,” said the journal’s editor in chief Ned Kalin, MD.
“Psychiatric illnesses such as anxiety and depression are common during the perinatal period with the most severe problems presenting as postpartum psychoses. Because of the potential negative impacts, sometimes severe, on mother and baby, these illnesses need to be identified and effectively treated.”
Postpartum depression involves extreme sadness, anxiety, apathy and issues with sleep, appetite and energy.
It affects an estimated one in seven women after they have given birth and is associated with an increased risk for suicide and self-harm. The treatable mental health condition is estimated to be responsible for at least 10% of all pregnancy-related deaths.
Most attempts to develop a risk stratification tool have been limited as they lack external validation or include individuals who are currently depressed or treated for depression.
Mark Clapp, MD, from Massachusetts General Hospital, and co-workers therefore set out to develop and externally validate a simple risk model for postpartum depression. This risk model was based on machine learning using electronic health record information known to the clinical team for women who gave birth in the hospital and had no recent history of a depressive disorder.
Among 29,168 women, 9.2% met at least one of the criteria for postpartum depression in the six months after delivery. The model was able to predict the onset of postpartum depression in 28.8% of women who had it and could predict the lack of postpartum depression in 92.2% of people who did not develop it.
“This tool could help identify patients within a practice at the highest risk and facilitate individualized postpartum care planning for the prevention of, screening for, and management of postpartum depression at the start of the postpartum period and potentially the onset of symptoms,” the investigators suggested.
In the second study, Behrang Mahjani, PhD, from Mount Sinai School of Medicine, and colleagues examined the familial risk of postpartum psychosis, one of the most severe psychiatric conditions with high risks of suicide and infanticide if untreated.
They note that although genetic factors contribute to this risk, the extent of familial risk remains to be determined.
Among 1,648,759 women on Swedish nationwide registers, 2514, or 0.15%, experience postpartum psychosis within three months of giving birth to their first child. Women were 10.69 times more likely than others to experience postpartum psychosis if they had a full sister who had the condition compared with a woman whose sister did not have the condition, after adjusting for year of and age at giving birth.
Women with a sister who had bipolar disorder and postpartum psychosis were at greatest risk, with their risk increasing approximately 14-fold.
The authors noted that, despite the higher familial risk in full siblings, the absolute risk among these women remains low, at an estimated 1.6% in the entire population.
Nonetheless, they added: Our findings suggest that female siblings of women who have experienced postpartum psychosis should be counseled that their own risk is elevated, but that the absolute risk for the disorder remains small.
link
