A team from the University of Coimbra develops a system to predict birth type

A multidisciplinary team from the University of Coimbra is developing a computer analysis system to predict the likelihood of an induced birth occurring vaginally, as announced this Monday.

So far, researchers involved in the project have analyzed data from 2,600 women followed at the Hospital and University Center of Coimbra (CHUC), “which indicates promising results,” the higher education institution said.

“The next step is to analyze the collected ultrasound images and then create a tool with all this data and test it on real people,” the Faculty of Science and Technology of the University of Coimbra (FCTU) said in a statement sent to Lusa. agency.

The work involves researchers from FCTUC’s Department of Computer Engineering (DEI) and the Faculty of Medicine (FMUC), who “seek to develop a system that aims to predict, through computer analysis, the possibility of vaginal birth.” after its introduction.”

“Induction of labor is being done more and more often, but does not always result in vaginal birth,” he said.

This was the “starting point” of the study “Prediction of vaginal birth after induction of labor using machine learning”, authored by Yolanda Ferreira, PhD student in Medical Sciences, under the scientific supervision of Ana Luisa Areya, Professor at FMUC, with the assistance of João Nuno. Correia, a professor at the same college.

“All inductions have a 30 to 35% chance of ending in a C-section, so we know in advance that 70% of women give birth vaginally. Caesarean section, we could provide appropriate and proactive advice on the need to induce labor, a difficult process for mother and fetus, which, in fact, can increase the emotional and economic burden associated with this procedure,” explained Yolanda Ferreira.

Because it’s “such a common procedure that generates so much data, we thought maybe we could use a method that analyzes it to help doctors understand whether—or when it’s worth it—to invest in induction for vaginal birth “, – he said.

The doctoral student quoted in the note emphasized that “obstetricians currently invest in inducing labor in all women, knowing from the very beginning, due to certain characteristics, in whom this may or may not occur vaginally.”

According to João Nuno Correia, “the idea is to come up with something that, by combining data (tables and images), forms a support module that provides personalized information about each pregnant woman’s high probability of vaginal birth after induction.” .

“If this indicator is high, induction will be carried out with greater confidence. If not, then there is a very high probability of caesarean section, the pregnant woman can be advised otherwise,” he added.

In this study, the innovation involves “predicting the type of birth using ultrasound image data.”

“The doctor is based on the clinical history of the person and his characteristics after this pregnancy, (…) we want to see whether the system, when analyzing this combination of clinical data and images, perceives it in a way that will help later or not.” draw any conclusions,” the researchers emphasized.

This collaboration between DEI and FMUC “is fundamental because in the future it will support physician decisions before birth and therefore improve both neonatal outcomes and women’s experiences during childbirth,” they concluded.