Male hand and baby hand touch.

Intelligent prediction of preterm birth using AI-empowered electrohysterography sensing

Funder

NIHR

This project is funded by the National Institute for Health Research (NIHR) Artificial Intelligence in Health and Care Award. The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

National Institute for Health Research logo

Team

Dr Yuhang Xu, Professor Dingchang Zheng and Ms Nikki Holliday

Collaborators

Dr Elizabeth Bailiey (Birmingham City Universityand University Hospitals Coventry and Warwickshire)

Duration

1 Spetember 2021 - 31 August 2022

Project overview

Preterm birth increases not only the risks of neonatal mortality, but numerous health risks for surviving babies. Pregnant women have been defined by the NIHR as an under-served group whose inclusion needs to be improved in clinical research.

The public sector costs of preterm delivery in England and Wales have been estimated at £3.4bn annually. Prediction of preterm birth is crucial to deliver interventions that are time-dependent for neonatal survival. The main device currently available for routine uterine contraction detection is the tocodynamometer (TOCO), which suffers from several drawbacks and provides limited information for the prediction of premature birth. The more comfortable technique called electrohysterograhy (EHG) offers an alternative for uterine contraction monitoring.

Project objectives

Using available EHG signals from existing datasets (one online database, one from our previous research), we aim to develop an AI solution to: 1) translate EHG to TOCO-like waveforms that could help facilitate the clinical acceptance for EHG adoption into maternal monitoring practice; and 2) predict preterm birth by an AI algorithm.

Impact statement 

This is a proof of concept study.  Our proposed technology is ultimately anticipated to reduce neonatal mortality, improve patient outcome and experience, and save costs for the NHS.

Follow-up work will include further development and clinical evaluation of the technology, in collaboration with health economists, clinicians and commercial partners to support future implementation and dissemination of our technology.

  • If the preterm birth is predicted, the patients can be offered interventions or immediate treatments in time, thereby reducing the risks of neonatal mortality and health risks for surviving neonates. According to the NICE Clinical Guidelines, the financial burden can be reduced by £1.14bn if the delivery is delayed by 1 week using some timely interventions. It can also help the families to avoid the extra costs associated with neonatal intensive care and long-term survival of the premature child.
  • If no preterm birth is predicted, the unnecessary interventions and administration of agents can be avoided, which may introduce additional risks, such as infection. 
  • The reliable identification of preterm birth provides reassurance to the pregnant. It would have clinical as well as emotional benefits by relieving the stress and anxiety that is brought by the worries about unpredictable delivery. 

Outputs

  • One paper published in peer-reviewed journal.
 Queen’s Award for Enterprise Logo
University of the year shortlisted
QS Five Star Rating 2020
Coventry City of Culture 2021