T'ng Chang Kwok



Enrichment Student

Cohort year


Partner Institution


Dr Kwok is a current Action Medical Research training fellow who is exploring approaches to incorporate artificial intelligence into routinely recorded national neonatal electronic patient records to support clinical decision making for high-risk preterm infants. He is working with Professor Don Sharkey (Professor of Neonatal Medicine and Technologies), Professor Jon Garibaldi (Professor of Computer Science) and Professor Carol Coupland (Professor of Medical Statistics) at the University of Nottingham to deliver the proposed project. His research interest includes the use of healthcare technology and artificial intelligence to improve the care of babies, with a focus on reducing mortality and improving respiratory outcomes. 

Dr Kwok is a neonatal subspecialty clinical doctor in the East Midlands with a strong passion for neonatal academia. He is also the trainee representative for the British Association of Perinatal Medicine data and informatics working group as well as the UK Resuscitation Council’s Newborn Life Support and Advanced Resuscitation of Newborn Infant subcommittee.

Research interests

Every year, about 8,000 very premature UK babies need intensive care. Sadly, nearly 900 will die and 2,300 develop bronchopulmonary dysplasia (BPD) due to abnormal lung development. BPD is a disabling severe breathing condition unique to premature babies and often causes life-long disability and educational problems. Parents and healthcare professionals have identified BPD prevention as a top research priority.

Preventative treatments such as steroids are frequently used to improve survival and reduce BPD but have side effects including cerebral palsy. Clinicians may need to make complex medical decisions for these treatments using a just a few birth characteristics. This is not ideal as every baby has unique factors affecting their outcome which changes daily. Presently, 400 data items are routinely recorded daily onto the national neonatal electronic patient record (EPR) for each baby’s three to six months neonatal unit stay, alongside outcome data after discharge.

Current prediction tools provide static risk predictions at fixed time points. Dynamic machine learning (ML) and statistical approaches are needed to predict the evolution of personalised clinical trajectories as each baby’s condition changes, providing an objective measure to support clinicians making difficult personalised treatment plans. This has never been attempted in neonatology.

Dr Kwok aims to develop a clinical decision support tool (PRIOR tool) using daily EPR data and explore/compare dynamic ML and statistical approaches to predict personalised trajectory for BPD or death as part of his Turing related research, funded by Action Medical Research.