Researchers have developed a groundbreaking AI model capable of accurately assessing maternal risk in pregnant women diagnosed with pre-eclampsia, potentially offering a lifeline in the fight against this life-threatening condition.

The study, involving over 8,800 women across 11 countries, has yielded a risk-prediction model that swiftly classifies the severity of pre-eclampsia-related complications into five distinct categories within just two days of initial assessment. Pre-eclampsia, affecting between 2 to 4 percent of pregnancies worldwide, stands as a significant contributor to maternal morbidity and mortality, resulting in an alarming number of maternal and newborn deaths annually, particularly in low- and middle-income countries.

While most cases of pre-eclampsia are mild and resolve postpartum, approximately one in ten women in the UK face severe complications, including life-threatening conditions such as stroke. Named PIERS-ML (Pre-eclampsia Integrated Estimate of Risk – Machine Learning), the newly developed model integrates Machine Learning, a subset of Artificial Intelligence, to enhance predictive accuracy.

The collaborative efforts of researchers from the University of Strathclyde in Glasgow and King’s College London have culminated in the creation of a versatile tool intended for international application. With plans underway to develop a user-friendly app for personalized risk assessment post-diagnosis, the model promises to revolutionize clinical decision-making and improve patient outcomes.

Lead author Tunde Csobán emphasizes the urgency of effective risk assessment in managing pre-eclampsia complications, underscoring the potential of the model to save lives. Dr. Kimberley Kavanagh, a co-author of the paper, highlights the model's adaptability to diverse clinical settings, offering tailored risk predictions aligned with individual patient circumstances.

Professor Peter von Dadelszen, the project's Principal Investigator, underscores the model's global relevance and its ability to dynamically adjust based on geographic and socioeconomic factors. The incorporation of variables such as countries’ GDPs and maternal mortality ratios enhances the model's applicability across various healthcare contexts.

The study's comprehensive validation process, involving over 8,800 women from diverse geographical locations, affirms the model's robustness and reliability. External validation exercises further validate its performance, solidifying its status as a pioneering tool in pre-eclampsia risk assessment.

Funded through STRADDLE (Strathclyde Diversity in Data Linkage), Tunde Csobán's PhD research has been instrumental in advancing this groundbreaking initiative, underlining the vital role of interdisciplinary collaboration in driving innovation in healthcare.

As the research continues to garner acclaim for its universal applicability and lifesaving potential, it stands as a testament to the transformative impact of AI in revolutionizing maternal healthcare worldwide.