Structured expert elicitation for long-term survival outcomes (NICE DSU TSD 26)
This method helps improve predictions of how long patients might live after treatment when real-world data is not yet available. It does so by systematically collecting and using informed opinions from medical experts in a transparent and structured way, reducing guesswork and increasing trust in the results used for health technology assessments.
At a glance
Use when
Long-term survival data are immature (e.g., in oncology trials with short follow-up); extrapolation of survival curves beyond observed data is required; high uncertainty affects cost-effectiveness results; regulatory or HTA bodies require external validation of survival assumptions.
Avoid when
Sufficient long-term empirical data are available; expert availability or expertise is limited; resources for rigorous elicitation are lacking; the process would be used informally or without documentation.
Inputs
Clinical context, survival data (e.g., Kaplan-Meier curves), expert panel selection criteria, elicitation protocol, predefined survival parameters (e.g., median survival, tail percentiles).
Outputs
Probabilistic estimates of long-term survival outcomes (e.g., survival curves, hazard ratios, uncertainty distributions), documented rationale and expert reasoning, inputs for cost-effectiveness models.
How it works
Structured expert elicitation (SEE) is a formal process to quantify expert judgment when empirical data on long-term survival outcomes are immature or sparse. It involves defining relevant survival parameters (e.g., median survival, tail probabilities), selecting qualified experts, using structured interviews or workshops to elicit probabilistic estimates, and incorporating these into survival extrapolation models. The method supports evidence synthesis in HTA, particularly in oncology, and aims to improve transparency, reproducibility, and defensibility of survival projections in submissions to bodies like NICE.
- HTA domains
- Clinical Effectiveness, Costs & Economic Evaluation
- Assumptions
- Experts have relevant clinical or epidemiological knowledge; their judgments can be reliably quantified; the elicitation process minimizes cognitive biases; expert uncertainty reflects true parameter uncertainty.
- Strengths
- Improves transparency and defensibility of survival extrapolations; allows incorporation of real-world clinical insight; supports decision-making under high uncertainty; aligns with HTA requirements for evidence synthesis.
- Limitations
- Resource-intensive to conduct rigorously; results depend on expert selection and elicitation design; risk of cognitive biases if not properly managed; often used for qualitative validation rather than direct model input.
- Also known as
- Structured Expert Elicitation, SEE, NICE DSU TSD 26, Expert Elicitation for Survival Extrapolation
Questions this answers
- › How can long-term survival outcomes be estimated when clinical trial data are immature?
- › How can expert opinion be incorporated into survival extrapolation models in a transparent and systematic way?
- › What are best practices for conducting and reporting structured expert elicitation in HTA?
- › How can uncertainty in survival projections be quantified using expert judgment?
- › When and how should elicited expert judgments be integrated into economic models?
- › How can bias and subjectivity in expert judgment be minimized during elicitation?
References & sources
Similar by meaning
Beta record. Based on the original catalogue summary; primary-source enrichment pending.