Preterm babies, also known as premature babies, are infants who are born before completing 37 weeks of gestation. They are born earlier than expected, which can pose various challenges and health risks. Just to put things in perspective, a normal pregnancy typically lasts around 40 weeks. But of course, if babies are born before that then it’s not without any consequences. Early Births are more vulnerable to a range of health issues like respiratory or developmental problems, feeding difficulties, infections, etc. They may experience long-term cognitive and developmental impairments, such as learning disabilities or even motor skill challenges.
Now, with an estimated 3.5 million Early Births each year, preterm birth is a major health concern in India. But with over 15 million newborns affected yearly, it is also a significant worldwide health problem.
New Development In The Field
In an ongoing quest to bridge the gap and reduce the prevalence of preterm births, researchers worldwide are tirelessly exploring innovative interventions to ensure healthier outcomes for both mothers and babies. Along the same lines, researchers Arye Nehorai and Uri Goldsztejn developed an algorithm based on deep learning to predict preterm deliveries as early as 31 weeks of pregnancy. The study’s findings were released in PLoS One on May 11.
Nehorai and Goldsztejn used the power of belly buzz to develop a method for predicting preterm births. They used a noninvasive technique called electrohysterograms (EHG) to detect uterine electrical activity through electrodes on the abdomen. They also considered important clinical information like age, gestational age, weight, and any bleeding during the first or second trimester, sourced from public databases.
This method, used around the 31st week of pregnancy, performs just as well as the standard clinical procedures in identifying imminent labor for women experiencing preterm labor symptoms.
In their study, Nehorai and Goldsztejn used a deep-learning model to analyze data from 30-minute electrohysterograms (EHGs) of 159 pregnant women who were at least 26 weeks into their pregnancies. Among these women, nearly 19% gave birth prematurely. By feeding the EHG recordings into a deep neural network, which automatically identifies important patterns in the data, they successfully predicted the outcomes of the pregnancies. The researchers discovered that certain aspects of the EHG measurements, particularly the higher frequency components, played a significant role in determining the likelihood of preterm births.
What Does This Translate To?
The model developed by Nehorai and Goldsztejn is user-friendly, cost-effective for clinical use, and has the potential to be used at home. Moving ahead, they aim to create a device for recording EHG measurements and gather data from a larger group of pregnant women to enhance their method and validate the results. This advancement could lead to better healthcare outcomes by enabling timely interventions, specialized care, and effective monitoring for both mothers and babies.