Concept Architecture
Concept
Theoretically, a Parametric Survival Model is a statistical survival model that assumes survival times follow a specified probability distribution characterised by a finite set of parameters. It is founded on parametric survival analysis and provides a complete mathematical description of the survival, hazard and cumulative hazard functions. The model exists to estimate survival efficiently and extrapolate outcomes beyond observed follow-up, making it fundamental to health economic evaluation.
Mathematically, a parametric survival model specifies a distribution such as the Exponential, Weibull, Gompertz, Log-Normal, Log-Logistic or Generalised Gamma distribution. The model parameters determine the shape of the survival and hazard functions and are estimated using maximum likelihood estimation. Covariates may be incorporated through accelerated failure time or proportional hazards parameterisations depending on the chosen distribution.
In practice, parametric survival models are fitted to individual patient survival data using specialised statistical software. Alternative distributions are compared using likelihood-based criteria, graphical diagnostics and clinical plausibility before selecting the most appropriate model. In health economics, parametric survival models are routinely used to extrapolate survival beyond clinical trial follow-up for lifetime cost-effectiveness and budget impact analyses.
Purpose
Used to estimate and extrapolate survival using predefined probability distributions, supporting long-term prediction of health outcomes in economic evaluation.
Mathematical Formulae
Primary Formula
S(t) = 1 ? F(t | ?)
where:
S(t) = survival function
F(t | ?) = cumulative distribution function
? = vector of estimated model parameters
Supporting Formulae
h(t) = f(t | ?) / S(t)
H(t) = ?ln(S(t))
L = ?? f(t? | ?)?? S(t? | ?)????
where:
h(t) = hazard function
H(t) = cumulative hazard
�? = event indicator
Related Mathematical Methods
- Maximum Likelihood Estimation
- Parametric Survival Analysis
- Accelerated Failure Time Model
- Proportional Hazards Model
- Weibull Model
- Gompertz Model
- Log-Normal Model
- Generalised Gamma Model
Example
Overall survival data from a phase III oncology trial are fitted using several parametric survival models. A Generalised Gamma model provides the best statistical fit based on AIC and produces clinically plausible long-term survival estimates. The fitted model is subsequently used to estimate lifetime quality-adjusted life-years and costs within a cost-effectiveness model submitted for health technology assessment.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| EXP | =EXP(-((A2/$B$1)^$B$2)) | Calculate survival probabilities for a Weibull model. |
| LN | =-LN(C2) | Calculate cumulative hazard from survival probabilities. |
| Solver | Maximum likelihood optimisation | Estimate survival model parameters. |
| IF | =IF(A2<=Observed_End,C2,D2) | Combine observed and extrapolated survival estimates. |
VBA (Optional)
Automate fitting, comparison and extrapolation of multiple parametric survival models for health economic evaluation.
Sources
- Collett D. Modelling Survival Data in Medical Research.
- Klein JP, Moeschberger ML. Survival Analysis: Techniques for Censored and Truncated Data.
- Royston P, Parmar MKB. Flexible Parametric Proportional-Hazards and Proportional-Odds Models for Censored Survival Data. Statistics in Medicine. 2002.
- NICE. Health Technology Evaluation Manual.
- Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation.
- ISPOR Good Practice Task Force Reports.
Related Concepts (2)
Library
Publications
8
NICE DSU Technical Support Document 14: Survival analysis for economic evaluations alongside clinical trials – extrapolation with patient-level data — Nicholas R. Latimer, TSD 14 ed., 2013 (NICE Decision Support Unit (University of Sheffield))
The reference guidance on survival analysis for economic evaluation: fitting standard parametric models (exponential, Weibull, Gompertz, log-logistic, log-normal) to censored trial data and extrapolating to estimate lifetime survival benefit, with a process guide for model selection and justification.
NICE DSU Technical Support Document 15: Cost-effectiveness modelling using patient-level simulation — Davis, Stevenson, Tappenden & Wailoo, TSD 15 ed., 2014 (NICE Decision Support Unit (University of Sheffield))
Guidance on individual patient-level (microsimulation) cost-effectiveness modelling — when to use it in preference to cohort models, how to structure it, and how to handle the associated computational and uncertainty challenges.
NICE DSU Technical Support Document 19: Partitioned survival analysis as a decision modelling tool — Woods, Sideris, Palmer, Latimer & Soares, TSD 19 ed., 2017 (NICE Decision Support Unit (University of Sheffield))
Guidance on the partitioned survival (area-under-the-curve) modelling approach widely used in oncology cost-effectiveness analysis, contrasting it with state-transition models and setting out its assumptions, strengths and limitations.
NICE DSU Technical Support Document 6: Embedding Evidence Synthesis in Probabilistic Cost-Effectiveness Analysis — Software Choices — Dias, Welton, Sutton & Ades, TSD 6 ed., 2011 (NICE Decision Support Unit (University of Sheffield))
Guidance on the software options and practical steps for embedding a Bayesian evidence synthesis directly within a probabilistic cost-effectiveness model so that parameter uncertainty is propagated consistently.
Cost-Effectiveness Analysis in R Using a Multi-State Modeling Survival Analysis Framework: A Tutorial — Williams, Lewsey, Briggs & Mackay, Vol. 37, No. 4 ed., 2017 (Medical Decision Making)
A tutorial on building cost-effectiveness models in R using a multi-state survival-analysis framework, bridging patient-level survival data and decision modelling — a key reference for survival-based economic models in R.
Journal ArticleView source →Survival Analysis: A Self-Learning Text — David G. Kleinbaum & Mitchel Klein, 3rd Edition ed., 2012 (Springer)
A practical introduction to survival data, censoring, Kaplan-Meier methods, Cox regression, proportional hazards and model interpretation.
BookView source →Statistical Models and Methods for Lifetime Data — Jerald F. Lawless, 2nd Edition ed., 2003 (John Wiley & Sons)
An advanced reference on lifetime and event-time data, including censoring, parametric distributions, hazard functions, regression and model assessment.
BookView source →Modelling Survival Data in Medical Research — David Collett, 3rd Edition ed., 2015 (Chapman & Hall / CRC Press)
A medical-research guide to survival modeling, including censored outcomes, parametric models, Cox regression, diagnostics and practical interpretation.
BookView source →
Media
1
Partitioned Survival Analysis vs Markov Models — Health Economics Explainer — Mtech Access (Hannah Gillies), 2023 (Mtech Access)
An expert explainer video summarising the NICE DSU guidance on partitioned survival analysis versus Markov models — their use in HTA, strengths, limitations and recommendations for cost-effectiveness modelling.
VideoView source →
Tools & Resources
1
survHE — Survival Analysis for Health Economic Evaluation (R package) — Gianluca Baio, R package ed., 2023 (CRAN)
An R package for fitting and comparing parametric survival models for health economic evaluation, including Bayesian estimation, and for extrapolating time-to-event data to inform cost-effectiveness models.
Software (R package)View source →
Frequently Asked Questions (6)
What is a parametric survival model?
A statistical model assuming time-to-event data follow a specific distribution, such as Weibull or log-normal, allowing extrapolation beyond observed data.
Source: Collett 2015
How is the best parametric survival model selected?
Fitting a parametric survival model means estimating the parameters of a chosen distribution so it matches the observed data, usually with several candidate distributions fitted and compared. Selection weighs how well each fits the data, using statistics such as the information criteria, against how plausible its extrapolated tail looks in the light of external evidence and clinical judgement. A form that fits the trial well but projects implausibly is set aside. The choice is not the fit statistic alone. Latimer (2013) sets out this selection process.
Source: Latimer 2013
What are common parametric survival models?
Common parametric survival models include the exponential, with a constant hazard; the Weibull, with a monotonic hazard; the Gompertz, with an exponentially rising hazard; the log-normal and log-logistic, with hazards that rise then fall; the gamma; and the flexible generalised gamma, which nests several of these. Each represents particular hazard shapes, so a range is fitted and compared. The choice depends on which distribution fits the data and extrapolates plausibly, since the models differ in their long-term behaviour.
Source: Latimer 2013
How is a parametric survival model fitted?
A parametric survival model is fitted by choosing a distribution and estimating its parameters from the data, usually by maximum likelihood, which finds the values making the observed data most probable under the distribution while accounting for censoring. The fitted model gives survival and hazard at any time. Candidate distributions are compared using goodness-of-fit measures, visual comparison with the observed survival, and the plausibility of their extrapolation, so that a model both fitting the data and projecting credibly is selected.
Source: Latimer 2013
Why are parametric survival models used for extrapolation?
Parametric survival models are used for extrapolation because their fitted distributions have a known form that can be evaluated beyond the observed follow-up, providing the long-term survival that economic evaluation needs but that trials do not observe. Non-parametric estimates cannot extend beyond the data, so a parametric model is required to project survival and estimate mean survival. The model supplies the unobserved tail of the survival curve, though this extrapolated portion depends on the distribution and is uncertain.
Source: Latimer 2013
What are the limitations of parametric survival models?
Parametric survival models depend on the assumed distribution, so an inappropriate choice biases the fit and especially the extrapolation, which follows the assumed form beyond the data where it cannot be checked, and distributions fitting the data similarly can extrapolate very differently. This makes long-term estimates uncertain and sensitive to the model chosen. These limitations mean several distributions are fitted and compared, background mortality is often imposed, extrapolations are assessed for plausibility, and sensitivity to the choice of model is examined.
Source: Collett 2015
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 22 Oct 2025
Content version: 1.0.0
Canonical Identity
- Term code
- HE-EM-SM-062
Stable URI · Machine-readable · Resolvable · CC BY 4.0