In a groundbreaking discovery, researchers at Brigham and Women’s Hospital in the US have unveiled the potential of artificial intelligence (AI) in identifying individuals at a heightened risk of postpartum hemorrhage, a prevalent complication during pregnancy. Utilizing the advanced Flan-T5 language model, the team successfully extracted medical concepts from electronic health records, offering a more refined understanding and identification of populations susceptible to postpartum hemorrhage—a condition characterized by excessive bleeding after childbirth that can pose life-threatening consequences.
The large language model, Flan-T5, has been trained extensively on vast textual datasets, granting it the capability to process, manipulate, and generate textual content with remarkable accuracy. The researchers found this tool to be a game-changer, achieving an impressive 95 percent accuracy in identifying patients with postpartum hemorrhage. Notably, the model outperformed the standard method by identifying 47 percent more patients, as reported in their study published in the journal npj Digital Medicine.
Corresponding author Vesela Kovacheva, associated with the Department of Anesthesiology, Perioperative and Pain Medicine at the hospital, emphasized the necessity for improved methods in identifying patients susceptible to this complication. Kovacheva stated, “We need better ways to identify the patients that have this complication, as well as the different clinical factors associated with it,” as per a report in Economic Times
The research team conducted a comprehensive analysis, leveraging the Flan-T5 model to extract postpartum hemorrhage-related concepts from discharge summaries of over 1.3 lakhs patients who gave birth at Mass General Brigham hospitals between 1998-2015. By prompting the language model with lists of known concepts associated with the medical condition, they achieved rapid and accurate results.
First author Emily Alsentzer, a research fellow in the Division, highlighted the model’s efficacy, stating, “We looked at all of the patients that Flan-T5 identified as having postpartum hemorrhage and looked at what fraction of those also had the corresponding billing code. It turns out that Flan-T5 was 95 percent accurate and allowed us to identify 47 percent more patients than we would have from the billing codes alone.”
Alsentzer expressed optimism about the potential applications of this tool, stating, “Ideally, we would like to be able to predict who will develop postpartum hemorrhage before they do so, and this is a tool that can help us get there.”
The researchers intend to continue exploring the capabilities of this approach by examining other pregnancy complications. Kovacheva remarked, “This approach can be applied to many future studies, and it could be used to help guide real-time medical decision-making, which is very exciting and valuable to me as a clinician.”
The integration of AI, particularly the Flan-T5 model, has demonstrated remarkable effectiveness in rapidly and accurately identifying individuals at risk of postpartum hemorrhage. This breakthrough not only enhances the understanding of pregnancy complications but also opens avenues for real-time medical decision-making. As researchers delve deeper into the potential applications of this approach, the future holds promise for more precise and proactive healthcare interventions in maternal care.
Also Read: Embrace The Earth: The Surprising Benefits Of Barefoot Walking Unveiled
