UCD team using AI to help diagnose pre-eclampsia and save lives

Diagnostic test powered by machine learning to improve outcome for pregnant women

Although it kills 50,000 women and 500,000 babies globally every year and results in a further five million premature births, the medical condition of pre-eclampsia remains difficult to diagnose. Clinicians still rely on high blood pressure and the presence of protein in urine as possible indicators, an approach that has not changed in the last 200 years.

University College Dublin’s new AI_PREMie diagnostic test looks set to change everything. This combination of patented biomarker testing and risk assessment, powered by machine learning, not only accurately diagnoses pre-eclampsia but also predicts the patient’s outcome. And with patient studies initiated in the three largest maternity hospitals in Ireland, the AI_PREMie test will cover 50 per cent of births in the country.

The idea was born six years ago when project lead Prof Patricia Maguire grabbed a coffee with new colleague Prof Fionnuala Ní Áinle, who talked about her experiences as a consultant haematologist at the Rotunda, seeing women with severe pre-eclampsia. They discussed the potential for biomarkers in the blood to serve as an early warning system and went on to confirm this with a study showing significant differences in the platelets of pregnant women suspected of having pre-eclampsia and healthy women of the same age.

Armed with this evidence and a large amount of related data, the team was able to take its research to the next level when it applied for Science Foundation Ireland’s AI for Societal Good Challenge funding. Using an interdisciplinary approach, it could unlock new insights with data analytics and machine learning techniques.


This interdisciplinary thinking has been incubated in the UCD Institute for Discovery, of which Maguire is director and where she launched the AI Healthcare Hub in September 2020. Although AI_PREMie is the flagship project, she says there is a wealth of data and information generated in an academic environment that’s not being leveraged, and the goal is to nurture new projects in this space.

Concept phase

As one of the 12 teams picked for the SFI challenge in 2020, AI-PREMie won €20k for its concept phase. It has since gone through a Dragons’ Den-style pitch to an international panel of industry, Government and academic experts to make it to the second-round prize of €200k in seed funding, which saw the team grow to 10 people including John Curran, head of technology for SAS Ireland, the Irish division of global data analytics firm SAS Institute.

“Working with the SAS Institute’s Viya platform is what has enabled us to get our machine learning model so quickly from the lab into production,” explains Maguire.

“We have free access to powerful software that enables us to bring our prepped data and produce an algorithm on the back of that. We can quickly generate machine learning models, take the champion model and put that into real-life production.”

And the pandemic has not been a deterrent throughout the last 15 months of intensive research and development. The team has managed to continue to collect vital interview data from those at the frontline: women who have experienced pre-eclampsia and the healthcare workers who treat them.

“The restrictions of the pandemic actually helped; we have this great team that has been able to meet religiously on a weekly basis despite how busy we are because we can just Zoom in. Under ordinary circumstances we probably wouldn’t have been able to meet regularly in person,” Prof Maguire notes, adding that it has meant the continued contribution from valuable team members including Prof Mary Higgins, consultant obstetrician at the National Maternity Hospital, Dublin and Dr John O’Loughlin, laboratory manager at the Rotunda Hospital.

One message Maguire wants to get across is that AI-powered diagnostic tools like this are not being designed to replace medical experts or make their jobs obsolete. Her mantra is that AI should mean “augmented intelligence” rather than “artificial intelligence”.

“That’s key for me as a basic research scientist: if we can augment the intelligence of the clinician by giving them powerful information that pops up on their screen in front of them in real time, we can hopefully help them make a better decision,” she explains.

“It gives me goosebumps to think that we can help a clinician make that critical decision: should I deliver this baby? Should I do this emergency C-section, or should I wait till tomorrow? It’s a life-or-death decision – even one more day in the womb can have a significant developmental impact on a premature baby.”

Holistic information

In addition to the new diagnostics for pre-eclampsia – for which Maguire and her fellow team members Ní Aínle and Dr Paulina Szklanna won the 2021 NovaUCD Invention of the Year Award – AI_PREMie combines holistic information from the patient, and this combination is key.

“Using these insights, if we can augment decision-making in a hospital environment in real time. I think that’s where AI is going to be truly transformative in medicine,” she adds.

Although the image of algorithms crunching data can lead to some people thinking that it removes the human touch from medical practice, it has had the effect of including personal stories and experiences from women who have suffered with pre-eclampsia.

These stories were an important part of the holistic data used to create machine learning models for both diagnosing pre-eclampsia and suggesting a path to treatment by predicting the progression of the disease in the individual patient.

“The amazing thing was that even though some of these women had lost their child, they wanted to make sure that it didn’t happen to anybody else. They told their story so that they could help others,” says Maguire.

Right now, the biomarker testing is being carried out in a lab in UCD, but the next phase of AI_PREMie involves developing a standalone kit for hospitals. Beyond this, Maguire and her team have their sights set on global deployment, aiming to reach even small clinics in remote areas with a solution that might involve a blood prick test in tandem with a mobile app that can quickly process the information in the cloud.

“When I say one in 10 pregnant women will develop pre-eclampsia and 500,000 babies die each year, this is probably an underestimation because it is likely that it is under-reported in lower-income countries. The big dream is to reach every person who needs this test across the world.”