Clinical decision quality is typically assessed cross-sectionally (audit, peer review, outcomes). What is missing is a practical method to quantify how a clinician’s judgement evolves over time—whether it improves through calibration and learning, or decays through drift, fatigue, overconfidence, or persistent cognitive bias.
The Clinician Judgement Trajectory Predictor addresses this gap by augmenting an existing cognitive bias model with a longitudinal inference function that estimates a time-varying latent judgement state for each clinician, and produces actionable outputs: calibration trajectory, discrimination trajectory, bias drift indicators, and a learning/decay rate.
There is no widely deployed, model-based approach that:
- tracks clinician judgement as a time series rather than a one-off assessment;
- distinguishes true improvement (better calibration/discrimination) from artefacts (case-mix changes, documentation effects);
- quantifies decay/drift in a way that triggers earlier support and governance interventions.