Palmer flexible survival model selection algorithm for immunotherapies
This method helps researchers choose the best statistical models to predict long-term survival for cancer immunotherapies, especially when standard models don't fit well. It uses expert guidance to decide if more flexible models are needed and which ones to test.
At a glance
Use when
Modeling long-term survival for cancer immunotherapies with non-standard hazard patterns, especially when trial data show tailing hazards or potential cure fractions.
Avoid when
When dealing with therapies showing proportional hazards and simple survival curves; when data are immature or too sparse to support complex modeling; when computational resources or expertise are limited.
Inputs
Clinical trial survival data, external real-world evidence, expert clinical input, observed hazard functions, data maturity indicators, evidence on treatment effect heterogeneity.
Outputs
A justified selection of flexible survival models for extrapolation, a set of plausible alternative models for sensitivity analysis, and a transparent rationale for model choices in economic evaluations.
How it works
The algorithm provides a structured 8-step process with 4 key questions to guide the selection of flexible survival models for economic evaluations of cancer immunotherapies. It emphasizes clinical expert input, assessment of hazard function shapes, evaluation of long-term survival or cure potential, integration of external evidence, and consideration of heterogeneity in treatment effects. The approach supports robust extrapolation beyond clinical trial data by systematically identifying plausible models and reducing model selection bias.
- HTA domains
- Clinical Effectiveness, Costs & Economic Evaluation
- Assumptions
- Flexible models are better suited for immunotherapies due to non-proportional hazards and potential long-term survival tails; clinical expertise improves model relevance; external data can inform plausible survival trajectories.
- Strengths
- Incorporates expert judgment and external evidence systematically; addresses limitations of standard parametric models for immunotherapies; promotes transparency and robustness in model selection; supports regulatory and HTA acceptance through structured justification.
- Limitations
- Requires access to clinical experts and external data; may increase complexity of analysis; not all flexible models are implementable in standard software; subjective judgments may introduce bias if not documented.
- Also known as
- Palmer algorithm, Flexible survival model selection algorithm for immunotherapies
Questions this answers
- › When should flexible survival models be used instead of standard parametric models for immunotherapies?
- › How can external evidence and clinical expertise inform survival model selection?
- › What features of the hazard function suggest the need for flexible modeling?
- › How should long-term survivorship or cure be assessed in model selection?
- › Which plausible models should be included in sensitivity analyses?
- › How can model uncertainty be transparently reported in economic evaluations?
References & sources
Similar by meaning
- Survival extrapolation validation-based case study (Bullement et al.)
- Royston-Parmar flexible parametric spline survival model
- flexsurv
- Survival extrapolation incorporating general-population mortality using excess-hazard and cure models (Sweeting et al. tutorial)
- Mixture Cure Models in Oncology tutorial (Felizzi et al.)
Beta record. Based on the original catalogue summary; primary-source enrichment pending.