Development and implementation of clinically relevant, ethically grounded causal models in critical care


Project dates (estimated):

October 2024 - October 2027


Name of the PhD student:

Stella Prizeman-Green


Supervisors:

Annemarie Docherty – School of Population Health Sciences

Nazir Lone – School of Population Health Sciences

Sohan Seth – School of Informatics

Emily Postan – Edinburgh Law School 


Project aims:

This project combines applied data science with ethical analysis to explore how causal models might be developed in the critical care setting. Stella is applying novel methods aimed at gaining causal insight from observational data to compare the effect of different blood transfusion strategies on mortality among critically ill patients with cardiovascular disease. Intertwined with this work is a philosophical analysis of the ways that causal methods are promoted in academic literature, and the ethical ramifications of causal methodology as compared to other prevalent techniques applied to observational clinical data. 


Disciplines and subfields engaged:

  • Philosophy

  • Data Science

  • Bioethics

  • Epidemiology 


Research Themes:

  • Ethics of algorithms

    • Ethics of Algorithmic Decision-Making

  • Ethics and politics of data

    • Ethical Data Science and Data Practice


Related Outputs:

Publications:

  • Prizeman-Green, Stella, and Annemarie Docherty. ‘Causation Implies Correlation: A Practical Guide to Target Trial Emulation in Critical Care’. Anaesthesia 81, no. 6 (2026): 875–79. https://doi.org/10.1111/anae.70187.

Presentations:

  • (upcoming) October 2026 Causality and Causal Inference in Medicine - oral presentation: Reframing causal modelling using observational data: from actionability to evidence.

  • January 2026 Isaac Newton Institute Foundations of Causal Inference workshop and conference – poster presentation: Delivering ethically sound causal models in the ICU – S. Prizeman-Green, N. Lone, S. Seth, E. Postan, A. Docherty

Previous
Previous

Limits and Paths Forward in AI Alignment

Next
Next

Examining the ethical implications of natural language processing