Birmingham’s Dexter spins out to unlock healthcare data for research
The University of Birmingham has launched Dexter AI Ltd, a new spin-out commercialising a web-based platform designed to extract analysable datasets from electronic health records and other healthcare data sources.
The platform is intended to reduce the time and effort required to turn large healthcare datasets into usable evidence for medical and population health research. It can work with electronic health records across primary and secondary care, as well as claims data, private prescriber records and other datasets.
Developed by researchers in Birmingham's Department of Applied Health Sciences, Dexter has already been used in more than 150 peer-reviewed studies, including research published in Nature Medicine, The Lancet journals and the BMJ. Applications have included identifying the full spectrum of symptoms associated with Long COVID, investigating thyroid disorders and atrial fibrillation, and supporting clinical trials from planning and feasibility through to recruitment, outcome evaluation and safety monitoring.
The platform addresses a fundamental problem in working with electronic health records: extracting the relevant information from large and complex datasets can itself be a labour-intensive research task.
Rather than requiring data to be transferred between systems, Dexter is a web-based application that users can take to their data. This allows datasets to be extracted and analysed within the existing data environment, reducing the need to move potentially sensitive healthcare information.
“Dexter was originally developed at the University of Birmingham to make health data research faster, more accessible and more reproducible,” said Professor Krishnarajah Nirantharakumar, who developed the software with Dr Krishna Gokhale and Professor Joht Singh Chandan. “Since then, it has enabled more than 150 studies, including publications in Nature Medicine, The Lancet journals and the BMJ.”
The technology has also been developed and tested across different data sources through partnerships with organisations including IQVIA, Cegedim Health Data and Adelphi Real World. Applications include real-world evidence studies, identifying potential clinical trial participants from electronic patient records and population health research.
Dexter can support studies ranging from incidence and prevalence analysis and cohort building to clinical trial emulation and pharmacoepidemiology, while the team is also exploring automation of clinical audits.
Commercialisation follows several years of development through the University of Birmingham Enterprise programme. Christian Billinghurst, who has worked with the team during this period, will lead Dexter as CEO. Nirantharakumar will become Chief Scientific Officer, Gokhale Chief Technology Officer and Chandan Chief Medical Officer.
The platform has expanded beyond its original focus on primary care electronic health records, with projects involving secondary care datasets, European primary care data, claims and private prescriber records.
Dexter AI is now exploring potential uses beyond healthcare, including justice data and social inequality, where its ability to work with diverse datasets could be applied.
Related Content
Addressing the rheumatic disease diagnostic gap via the lab
Early intervention benefits the treatment of rheumatic condition. Too often, delays are caused by lack of access to sophisticated testing …
Latest Lab Talk News in Brief: digital expansion, new build, tech collab, new markets, financing…
The National Pathology Imaging Co-operative has expanded its national digital pathology network to 15 hospital sites following the addition of …
Lab gloves could inflate microplastics readings suggests study
Scientists may be overestimating their assessments of microplastics pollution, with the cause lying literally in their own hands, suggests a …
