Managing uncertainty in medical decision making

ESRC Grand Union Doctoral Training Partnership (DTP) collaborative doctoral studentship

The Department of Experimental Psychology, the University of Oxford with the Oxford University Hospitals NHS Foundation Trust.

SUPERVISORS

Professor Nick Yeung, Department of Experimental Psychology, University of Oxford

Dr Helen Higham, Consultant Anaesthetist and Director of the OxSTaR Centre; Nuffield Department of Clinical Neurosciences, University of Oxford

Dr Ben Attwood, Divisional Director of Clinical Support Services, Oxford University Hospitals NHS Foundation Trust

PROJECT OVERVIEW

This translational research project will apply insights from psychological research to analyse and improve medical decision making. The specific focus will be the role of uncertainty as it impacts the use of automated (Artificial Intelligence - AI) tools in medical decisions.

Uncertainty is inherent in complex human decisions across diverse economic and social domains. Research in psychology has made significant progress in characterising the role of uncertainty in individual and group decision making. This research has shown that decision makers’ experience and expressions of confidence provide useful but imperfect indications of their uncertainty and the quality of their decisions. In group settings, expressions of confidence affect how advice is given and received, and teams have been shown to make optimal decisions only if team members communicate confidence and uncertainty effectively. This research has been primarily conducted in the context of carefully-controlled lab studies, but provides powerful tools with which to analyse and potentially improve real-world decision making. This project aims to do so in the context of the use of AI tools in critical care. Many such tools provide numerical estimates (e.g., of disease risk) with no accompanying information about uncertainty such as confidence intervals, potentially leading to suboptimal use of the information provided (e.g., over-reliance due to perceived precision of an estimate that is in fact highly uncertain). This project will investigate whether clinicians’ use of AI tools can be increased and improved via effective communication of algorithmic uncertainty.

The research will adopt a mixed methods approach. Possible research strands include: Observational studies of the use of AI tools in hospitals and in immersive simulation training conducted at OxSTaR; semi-structured interviews with clinicians; quantitative experimental work (e.g., using vignette studies to probe clinicians’ decision making processes and evaluations).

The project would be particularly appropriate for applicants with a strong background in psychology and a keen interest in applying that background to practical problems. The project would also be relevant to applicants with a clinical background and prior experience with psychology or AI research.

PARTNERSHIP

The partner organisation is the Clinical Support Services (CSS) division of the Oxford University Hospitals NHS Foundation Trust (OUHT, https://www.ouh.nhs.uk). The Divisional Director of CSS, Dr Ben Attwood, will provide expert advice and guidance within his role on the project’s supervisory team. Via CSS, the student will have the opportunity to work with multidisciplinary healthcare professionals, conduct observational studies of the use of AI tools in critical care settings in hospitals, and run studies to probe clinicians’ decision making processes and evaluations, for example to provide proof-of-concept evidence for proposed interventions for improving the use of automated (AI) tools. The student would also be a member of the OxSTaR team (https://www.oxstar.ox.ac.uk), which provides opportunities for research involving immersive medical simulation training, and the benefits of expert input from this collaboration of clinicians, medical professionals and researchers.

STUDENTSHIP DETAILS

The PhD studentship will be funded by the Grand Union DTP for an expected 3.5 years from October 2024. The award length offered may differ depending on the candidate’s prior training and how they meet ESRC training requirements. .

Information about Grand Union DTP ESRC studentships and eligibility can be found on the Grand Union DTP website.

APPLICATION DEADLINE

12:00 midday UK time on 15 December 2023

HOW TO APPLY

To apply for the studentship, you must submit an application to study for a DPhil in Experimental Psychology at the University of Oxford. Details about applying can be found via https://www.ox.ac.uk/admissions/graduate/courses/courses-open-for-studentships#RD_EP1.

Your application should include a research proposal that indicates how your proposed research would engage with the core elements of this project if you were offered the studentship. Your proposal may draw on the description of the research detailed above.

In addition you must complete a Grand Union DTP Application Form and upload it, together with your graduate application form. Please ensure you also select 'ESRC Grand Union DTP Studentships in Social Sciences' in the University of Oxford scholarships section of the University's graduate application form.

The admissions process is in two parts: applications will be assessed by the institution and the selected candidate will then be assessed as part of the applicant pool for ESRC Grand Union DTP funding.

QUERIES

Queries about the studentship should be addressed to: Professor Nick Yeung (nicholas.yeung@psy.ox.ac.uk)