Becoming a parent is one of the most exciting things in life. Unfortunately, sometimes pregnancy complications occur, remaining a serious challenge, that is not often talked about.
Ahead are interviews with leading researchers in Dublin aimed to understand the impact of data and analytics on maternal health.
According to the Irish Neonatal Health Alliance, pre-eclampsia causes persistent high blood pressure during pregnancy or the postpartum period, potentially having damaging effects on organs. If untreated, pre-eclampsia could be fatal for mothers and for babies.
A 2021 study in Pregnancy Hypertension, an international Journal of Women's Cardiovascular Health suggests that nearly 6% of all pregnant women in Ireland were diagnosed with hypertensive disorder of pregnancy (HDP), of which 79% with HDP had a diagnosis of pre-eclampsia, making it roughly 5% of all maternities in Ireland.
Challenges of diagnosing pre-eclampsia
Dr. Patricia Maguire is a professor of Biomedical Science, Principal Investigator in the research lab, and Director of the Institute for Discovery at University College Dublin. One of her recent research projects focuses on pre-eclampsia, working closely with both the Rotunda, the National Maternity, and the Coombe Hospitals in Dublin.
According to Maguire, some women can have very mild symptoms but others can have very severe symptoms and become ill, very fast. For pre-eclampsia, there are no clinical tests that tell a woman if she has the disorder.
"There is no actual test that says yes, this is it," Maguire said. "In the last 60 or 70 years, there has been no breakthrough ‘rule in’ diagnostics at all. We're still testing these women by looking at proteinuria, which is the amount of protein in their urine or their blood pressure. It’s totally nonspecific.”
Using AI-based ML models to identify pre-eclampsia
Using AI and machine learning, researchers worked to identify and compare blood biomarkers between women with pre-eclampsia and healthy pregnant women. The technologies have helped researchers to compile huge data sets and turn them into clinically relevant insights.
“After about four years of really in-depth investigation using omics technologies, we found some differences. However, we just had so much data. So, we used some machine learning algorithms to help us understand that data and extract the real differences," Maguire said.
This has led to an AI system that can help to diagnose pre-eclampsia in sick pregnant women. Maguire says researchers were able to identify 30 differences and using ML and categorized the top three determinants that could both predict pre-eclampsia severity and presence. Maguire is hoping to open it up to investigate further differences due to new funding to take the AI system to the next level.
Building a project from the ground up
Maguire explained that her team has been able to expand in the last 18 months since getting new funding.
Her team has grown from three principal investigators to over 20 people.
"We are incredibly lucky to have an amazing group of researchers right across the spectrum from basic biomedical scientists to machine learning experts, like Associate Professor Brian Mac Namee, health economists, including Professor Gerardine Doyle from the Smurfit Business School and then of course Dr. John O’Loughlin at Rotunda. It has been incredible to build this team to be able to accelerate this project. We’re now working across the three Dublin maternity hospitals, which means that the project covers 50% of all births in Ireland.”
Forming a foundation for the future
Maguire is particularly excited by the interdisciplinary nature of the project, especially given her role at the Institute for Discovery. She hopes that it will be the foundation for other work across the university and in its associated institutions.
She believes the role of the Institute for Discovery is to cultivate an interdisciplinary community. Seeing the project take shape along with the support of machine learning and statistics "makes the magic happen," Maguire said.
“The goal in the University College Dublin (UCD) AI healthcare hub, is to make these advanced technologies available to all researchers across the university," Maguire said. "But also shows how valuable education, machine learning, AI and good analytics can be and how transformative that can be to any data-rich research project. For me, it’s all about solving complex problems, like pre-eclampsia, in a really impactful way.”
Stay tuned for part two of a blog on the topic of maternal health, including a deep dive into the analytical and machine learning side of the maternity project in Dublin.