Transforming complex and large maternity data into research-ready infrastructure

Induction of labour is common in UK maternity care, but its real-world outcomes and fairness are hard to study, because maternity records are large, complex, and inconsistently recorded. This matters for pregnant women, babies, maternity teams, and the NHS, because induction can affect how a baby is born, the risk of heavy bleeding, and whether a newborn needs neonatal care. 

Using the Dexter platform, our BRC team turned routine BadgerNet maternity records into a standardised, research-ready format. Data from three NHS trusts were processed: Sandwell and West Birmingham (SaWB) and University Hospitals Birmingham (UHB) inside the West Midlands Secure Data Environment, creating a database of 310,000 pregnancies; and Shrewsbury and Telford (SaTH), processed locally and transferred in. As an early demonstrator, we studied induction of labour using SaWB and SaTH data.

The study found that induction was linked to higher rates of emergency caesarean, assisted birth (forceps or ventouse), heavy bleeding after birth, and neonatal unit admission. It also showed who was more likely to be induced: clearly by age and when pregnancies continued beyond the due date, and, in the more diverse SaWB population, more often among Asian and Black women. 

This work has improved the NHS’s ability to use routine maternity data for trustworthy research and service evaluation. What once took months of manual processing now takes weeks, and the pipeline can be reused for future data refreshes and additional trusts.

It also highlights where access to induction varies between groups, supporting fairer maternity care. The biggest measurable gains are scale, speed, and reuse: 310,000 pregnancies processed into a standardised resource, about 32,600 analysed in the demonstrator study, and a foundation for future work such as gestational diabetes and interpreter-support studies.