1. Faculty of Engineering and Physical Sciences
  2. School of Computer Science
  3. People
  4. Profile
Allan Pang

Allan Pang

  • Email: ugm5a2p@leeds.ac.uk
  • Thesis title: Development, validation and implementation of an adaptive machine learning model for perioperative mortality prediction and its adoption into combat casualty care
  • Supervisors: Owen Johnson, Dr Marc de Kamps, Dr Alwyn KotzĂ© (Leeds Teaching Hospitals NHS Trust), Prof Geoff Hall (Leeds Teaching Hospitals NHS Trust)

Profile

Allan is a military anaesthesia speciality trainee and a PhD student at the UKRI Centre for Doctoral Training in Artificial Intelligence for Medical Diagnosis and Care at the University of Leeds.  Allan graduated from Leeds University Medical School (MBChB) in 2010 and continues to serve as a Regular British Army Medical Officer.  Allan was selected for military anaesthesia speciality training in 2015 within the Northern School of Anaesthesia and Intensive Care Medicine. He gained his Fellowship from the Royal College of Anaesthesia (FRCA) in 2019.  Allan’s research interest is applying established Artificial Intelligence/Machine Learning techniques to the noisy clinical real-world environment.  His PhD, funded by the Academic Department of Military Anaesthesia and Critical Care (ADMACC), has concentrated on applying Time-Series Machine Learning techniques to develop temporally aware risk scoring systems.  Allan co-leads the Alan Turing’s Clinical AI Supra-Interest Group in Anaesthetics and Intensive Care.

Research interests

Allan's PhD research concentrates on predicting peri-operative (i.e. surgical) outcomes. The current state-of-the-art prediction tools produce only point probability before the surgery itself, which has limited clinical utility beyond the surgical event itself.  During his PhD, Allan has developed a surgical inpatient mortality prediction tool using routinely collected physiological data within the Electronic Health Record from Leeds Teaching Hospitals NHS Trust.  Allan is using Machine Learning techniques to capture the temporal relationships of routinely collected physiological signals to improve the predictive performance of clinical deterioration warning systems, similar to the National Early Warning Score (NEWS2).  

 

Qualifications

  • MBChB - University of Leeds
  • FRCA - Royal College of Anaesthetists