麻豆精选

• PhD in Mathematical Statistics, University of Exeter, 2000
• BSc (Hons) in Mathematical Methods for Information Technology, Nottingham Trent University, 1996

Gary is a statistician with over 25 years experience in both methodology and applied research. He is an expert in the development and validation of clinical prediction models/predictive AI tools. Gary’s current major research activity focusses on their evaluation, with a particular interest on fairness considerations.

Before joining the University 麻豆精选, he was at the University of Oxford between 2006 and 2025 working alongside Professor Doug Altman. At Oxford he was the Director of the Centre for Statistics in Medicine and made a Professor of Medical Statistics in 2016. During this period, he also became interested in tackling how to improve the quality and transparency of health research, evaluating published research and driving/supporting initiatives to improve how research is designed, conducted and reported. In recognition of this strand of his research, in 2018, he took on the directorship of the activities in the UK, which he still leads.

Gary is a steering group member, and former co-chair of the international , which aims to provide accessible and accurate guidance in the design and analysis of observational studies, and currently sits on the external advisory board for the Centre for Open Science . He is also on the advisory board of the . Gary has had multiple involvements with NIHR, including being a member of the NIHR HTA Commissioning Board from 2016 to 2020, and currently on the NIHR doctoral research fellowship committee. He has been an NIHR Senior investigator since 2024.

Gary is interested in supervising doctoral research students in the following areas:

• methodology for clinical prediction models
• predictive AI/machine learning for healthcare
• methodology for AI evaluation and fairness
• reporting standards and guidelines
• meta-science and open science methods and evaluation
• research transparency

Gary works on numerous international research initiatives to improve the methodology, and develop standards and frameworks for research involving predictive AI/clinical prediction models for healthcare. Improving the quality and transparency of predictive AI research drives his research, so that the tools used in practice are robust, safe and improve clinical workflows.

Gary leads the international TRIPOD consortium which he established alongside Carl Moons, Hans Reitsma and Doug Altman. He led the initiative to develop the  statement; a set reporting recommendations for studies developing and validating clinical prediction models, which were published in eleven leading medical journals in . He led an update to these recommendations which were published in 2024 to cover models developed using artificial intelligence and machine learning methods () and large language models ().

Gary is also involved in other guidance for artificial intelligence and machine learning studies in healthcare, including

•  - for AI based diagnostic test accuracy studies
•  - for early-stage live clinical evaluation of AI systems
• /  - for AI intervention studies
•  - for evaluating chatbots providing medical advice 

Gary is involved in development of risk of bias tools for machine learning diagnostic test accuracy studies () and prediction model studies (), and the  principles for trustworthy and deployable artificial intelligence in healthcare.

Gary is involved in the development of numerous other reporting guidelines for other study designs, including the

•  statement for studies providing global health estimates
•  statement for mediation analyses
• 2025 update of  and  guidelines for clinical trials
•  and  for trials with surrogate endpoints