The University of Edinburgh
Ethics of artificial intelligence in healthcare resource allocation: an empirical bioethics project
Abstract
dc:description.abstractWhat should we do when we don’t have enough of a healthcare resource—like a vaccination, ventilator, or organ for transplant—for every patient who might benefit from it? Given the persistent reality of limited resources, decisions must be made about who and what is prioritised in conditions of scarcity. How might we make these decisions? This thesis considers the ethics of involving machine learning models in healthcare resource allocation decision-making. Machine learning is a type of artificial intelligence that involves the prediction of outcomes without explicit programming. The promise of machine learning is that it might be able to use large amounts of patient data to make more accurate predictions of who might benefit most from a particular intervention. But there are a number of ethical issues with using machine learning in healthcare allocation decisions. What do we mean by ‘benefit most’, and how can this be accounted for in the design and implementation of a machine learning model? What should we tell patients about the use of these models? And is it possible to trust their involvement in clinical care? This thesis addresses these questions using empirical bioethics methodology, combining philosophical analysis and qualitative interviews to arrive at recommendations for an ethical deployment of machine learning in healthcare resource allocation. It considers three main philosophical themes in relation to this topic: distributive justice, patient consent, and trust. The thesis then analyses a case study in the ethics of algorithmic resource allocation. It does this through a series of interviews with transplant patients and staff about their experiences with the transplant benefit score, an algorithmic approach to allocating livers on the UK transplant waiting list. The analysis of this case study focusses on participant experiences relating to the three key philosophical themes of distributive justice, patient consent, and trust. It then uses reflective equilibrium to synthesise the philosophical and qualitative findings and make recommendations for the ethical deployment of machine learning in healthcare resource allocation. It concludes by proposing a role for deliberative democracy in further developing the recommendations of this thesis.
Degree
thesis:*- Grantor dc:publisher
- The University of Edinburgh
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Webb, Jamie
- Advisors dc:contributor.advisor
-
- Harrison, Ewen
- Wigmore, Stephen
- Cunningham-Burley, Sarah
- Gill, Michael
Subjects
dc:subject × 3Rights
- Language dc:language.iso
- en
Identifiers
dc:identifier.*- Identifier URI
- http://dx.doi.org/10.7488/era/6648