Gary Collins

NEW PAPER out today in @BMJ_latest TRIPOD+AI: reporting recommendations for studies developing or validating prediction models for use in healthcare that use #machinelearning methods bmj.com/content/385/bm… #ArtificialIntelligence #AIstandards #OpenAccess Please share 🙏

New preprint 'A general sample size framework for developing or updating a clinical prediction model' - generalises & extends existing ideas - anticipate posterior distributions - target n for expected performance or degradation - target n for assurance arxiv.org/abs/2504.18730

This article provides an overview of the current state of handling continuous variables in healthcare research. It discusses the potential limitations of assuming a linear relationship between independent and dependent variables bmj.com/content/390/bm…

Thrilled to see our work published in @bmj_latest. A practical guide on why categorizing continuous variables or assuming linearity often fails and how to model non-linear relationships using splines. Code for Case Study available on GitHub! @Richard_D_Riley @GSCollins

Outstanding paper! Especially suited for non-statisticians who want to get started with relaxing linearity assumptions without resorting to dreaded categorization techniques ... @Statistics

NEW PAPER in @bmj_latest "Dealing with continuous variables and modelling non-linear associations in healthcare data: practical guide" --> bmj.com/content/390/bm… #methodologymatters

🚨 Applications are now open for the NIHR Senior Investigators Cohort 19. Our Senior Investigators are among the most prominent leaders in applied health and social care research. We welcome applications from all professions across health and care in England and Northern…

NEW PAPER in @JClinEpi looking at the content of open peer-review reports of RCTs in the field of oncology sciencedirect.com/science/articl… Peer review reports [...] were short (median 276 words), superficial, and rarely constructive #metascience #peerreview #RCTs

🌏 Advancing international partnership for governing generative #AI (#GenAI) models in #medicine and #healthcare. We introduce POLARIS-GM initiative: a scenario-based, consensus-driven framework for GenAI #governance and #regulation. @NatureMedicine nature.com/articles/s4159…

NEW PREPRINT "Critical Appraisal of Fairness Metrics in Clinical Predictive AI" -> arxiv.org/abs/2506.17035 We identified 62 fairness metrics (& growing) - unsurprisingly it's all a bit of a mess...with most metrics not fit for purpose #predictiveAI #fairness #machinelearning

** New preprint led by Joao Matos & @GSCollins "Critical Appraisal of Fairness Metrics in Clinical Predictive AI" - An important area but much confusion exists - We identified 62 fairness metrics so far - Better standards & metrics needed for healthcare arxiv.org/abs/2506.17035

Maternal early warning scores shown to be methodologically weak and at high risk of bias - Journal of Clinical Epidemiology jclinepi.com/article/S0895-…

ICYMI - our new paper explaining how small sample sizes negatively impact the quality and performance of clinical prediction models in AI-based healthcare

NEW Viewpoint: Importance of sample size on the quality and utility of #AI-based prediction models for healthcare. Read it here: thelancet.com/journals/landi…

Thinking of applying for the NIHR Doctoral Award? If you hold a 1st or 2:1 bachelor's degree (or equivalent) and are passionate about research in health, public health or social care, this fully funded PhD opportunity could be your next big step. Make your application stand out…

This updated SPIRIT 2025 (Standard Protocol Items: Recommendations for Interventional Trials) explanation and elaboration document provides guidance on protocols of randomised trials for researchers bmj.com/content/389/bm…

The SPIRIT 2025 statement provides guidance on items to address in trial protocols, reflecting methodological advances and feedback from users bmj.com/content/389/bm…