Overview
Prediction models can help quantify risk and expected recovery, but only when they are well calibrated, externally validated and interpreted in the clinical context. NORA develops and evaluates models for selected neurosurgical outcomes with a focus on reproducibility and clinical utility.
Objectives
01. Build parsimonious, clinically interpretable prognostic models where possible.
02. Compare conventional statistical modelling with machine-learning approaches when justified by the question and dataset.
03. Prioritise calibration, external validation, missing-data handling and decision-curve or clinical-utility assessment.
04. Evaluate transportability across centres and patient populations before considering implementation.
Research direction
Models are intended to support — not replace — clinical judgement and shared decision-making. The programme therefore places as much emphasis on validation and interpretability as on predictive performance.
Collaboration
Project-specific collaboration depends on protocol maturity, governance, ethical approvals and data-sharing arrangements. Centres or investigators interested in aligned research questions are welcome to contact the NORA team.