Hippos at Risk: Data Deficiencies Threaten Conservation Progress

A new database of African hippo populations has revealed huge gaps in our knowledge of where the megaherbivores live and thrive, with populations fragmented and reliant on protected areas. A new database of African hippo populations has revealed huge gaps in our knowledge of where the megaherbivores live and thrive, with populations fragmented and reliant
HomeBabyAI Model for Accurate PTSD Assessment in Postpartum Women: A Game-Changer in...

AI Model for Accurate PTSD Assessment in Postpartum Women: A Game-Changer in Mental Health Care

analysis of women’s stories who went through traumatic childbirth and women who had uncomplicated childbirth, researchers have developed an AI model that can effectively identify those who are at risk of experiencing childbirth-related PTSD. Massachusetts General Hospital (MGH), a member of the Mass General Brigham healthcare system, conducted a study that revealed the potential of this generative artificial intelligence (AI) model to accurately screen for post-traumatic stress disorder (CB-PTSD) by analyzing the narrative accounts of recent childbirth experiences.The study examined the abilities and limitations of various OpenAI models, including ChatGPT, and pinpointed a specific version that can provide valuable insights into the mental health of mothers after experiencing a traumatic childbirth.

This particular model can be seamlessly integrated into regular obstetric care and has the potential to be used for evaluating other mental health conditions. The findings of the research were released in Scientific Reports.

Sharon Dekel, PhD, who is the director of MGH’s Postpartum Psychiatric Disorders Program, stated, “Currently, the assessment of PTSD related to traumatic birth heavily relies on extensive clinician evaluation, which does not address the urgent need for a quick, cost-effective assessment strategy.”with combat veterans, victims of violent crimes, and survivors of natural disasters. New research has revealed that AI-driven analysis of brief patient narratives of childbirth could potentially help in the detection of childbirth-related post-traumatic stress disorder (CB-PTSD). This method may prove to be a cost-effective and patient-friendly approach for identifying women at risk for CB-PTSD before the condition fully develops. Dr. Seng, the lead author of the study, believes that with further research, this tool has the potential to be a valuable resource in identifying and aiding women at risk for CB-PTSD.The article discusses how childbirth can be a significant trigger for PTSD, leading to detrimental effects on the health of the mother and child, as well as significant societal costs if left untreated. Previous research has shown that brief psychological interventions after traumatic childbirth can reduce maternal PTSD symptoms. The latest study, conducted by Dekel and Alon Bartal of Bar-Ilan University, focused on the effectiveness of artificial intervention.intelligence and machine learning (ML) techniques to identify CB-PTSD were used. The researchers specifically assessed the effectiveness of different large language models (LLMs) and variations of ChatGPT in extracting new insights from text-based data sets taken from short narrative descriptions provided by postpartum women about their childbirth experiences. The team gathered narrative accounts from 1,295 recently postpartum women for the study, with a focus on the OpenAI model known as text-embeddings-ada-002, which converted the personal narrative data of women with and without CB-PTSD.The team converted probable CB-PTSD into a numerical format, which was then analyzed by a trained machine learning algorithm. The researchers found that this model performed better at identifying postpartum traumatic stress compared to other ChatGPT and large language models. These models are usually trained on large amounts of data to understand and interpret natural language. The ML model relies on childbirth narrative input from the Open AI model as its exclusive data source, providing an efficient mechanism for data collection during the vulnerable postpar.The study showed that it had an 85% sensitivity and 75% specificity in detecting CB-PTSD cases,” Dekel says.

“In addition, the model we created has the potential to make CB-PTSD screening and diagnosis more accessible by integrating seamlessly into regular obstetric care and laying the groundwork for the development and widespread use of commercial products.”

Dekel, whose research focuses on women’s mental health after traumatic childbirth, emphasizes the clinical advantages of using a pre-trained large language model to evaluate potential PTSD in new mothers.

“Timely intervention is crucialIn order to avoid this disorder from becoming chronic and making treatment more complicated,” says the MGH researcher.

“Our new method could provide a fresh and affordable way to screen for high-risk women and help them get treatment early. It might also be useful for diagnosing other mental health conditions, which could improve patient results.”

The use of artificial intelligence in healthcare has been revolutionary and has the potential to transform the way care is provided. Mass General Brigham is at the forefront of this innovation as one of the top academic health systems in the nation.The largest innovation enterprises, such as MGH, are taking the lead in conducting thorough research on new and emerging technologies to responsibly incorporate AI into care delivery, workforce support, and administrative processes. Dekel, a psychologist at MGH and assistant professor of Psychology at Harvard Medical School, along with Bartal, an assistant professor of Information Systems at Bar-Ilan University in Israel, are collaborating on this research. Other co-authors in the Dekel Laboratory include Kathleen Jagodnik, PhD, a Harvard research fellow, and Sabrina Chan, a clinical research coordinator. Dekel’s work has been supported by funds from the NIH (Eunice Kennedy Sh.The river National Institute of Child Health and Human Development provided grants R01HD108619, R21HD109546, and R21HD100817.