What we do

We study the psychological underpinnings of judgements and decisions in healthcare. We focus predominantly on the study of decisions made by clinicians; for example, whether to refer a patient urgently for suspected cancer, what type of surgical intervention to choose for a patient, whether to prescribe antibiotics and when to stop antibiotic treatment. We also study the cognitive processes of clinicians when they assess risk and make diagnoses. We employ predominantly quantitative methodologies, such as controlled experiments and surveys, to study the influence of a number of variables on decisions, e.g., clinician characteristics, patient characteristics, features of disease presentation, as well as non-clinical factors, such as patient expectations and the decision context. By identifying weaknesses in the reasoning process, we develop tools to support it, with a view to improving diagnostic accuracy, patient safety, system efficiency and doctor-patient dynamics. We also investigate how existing decision support tools interact with clinical thinking and how they could be better designed for increased uptake.

We actively support Open Science via study preregistrations and Registered Reports, and by making our study materials, data, and code publicly available in the Open Science Framework (OSF).

Why it is important

Suboptimal decisions can introduce errors and delays in the diagnosis, prognosis, and treatment of disease, with potentially serious consequences for patients, practitioners, and the NHS. By identifying ways that decision pathways can fail, and devising tools to support them, we can improve the safety and quality of patient care.

How it can benefit patients

Timely and accurate medical care is the bedrock of patient safety. Effective support for clinical decision-makers helps to ensure that patients get the treatment they need when they need it, and spares them unnecessary treatment that could cause harm.

Summary of current research

Our current research includes:

• Understanding how clinical algorithms interact with clinical judgement. Specifically, we are studying how General Practitioners integrate scores from cancer risk calculators in their judgements and referral decisions; and how such calculators can be introduced to clinical practice for better use.
• Examining risk assessment and antimicrobial prescribing decisions in children presenting to the GP with cough; and how they compare to a clinical prediction rule (CPR – ‘STARWAVe’, Hay et al. Lancet Respir Med 2016). The research aims to inform current and future trials of using CPRs as decision aids for antimicrobial prescribing.
• Understanding why critical care doctors persist with antimicrobial treatment, despite a validated (biomarker-based) recommendation to discontinue.
• Developing a decision tool to support arthroplasty surgeons in evaluating different options for knee surgery (total vs. unicompartmental knee replacement), which have very different implications for the patient (likelihood of complications, reoperations, mortality) and the NHS (cost).
• Exploring how a diagnostic support tool that we developed in previous work influences how clinicians search for and evaluate clinical information, and its interaction with clinicians’ confidence about their diagnoses.
• Testing how a diagnostic support tool that we developed in previous work can be employed in clinical practice and its impact on the consultation, doctor-patient interaction and patient satisfaction.