Concept Architecture
Concept
Theoretically, the Generalized F Model is a fully parametric survival model based on the generalised F distribution, one of the most flexible probability distributions used in time-to-event analysis. It encompasses numerous standard survival distributions, including the Weibull, log-logistic and log-normal distributions, as special or limiting cases. In health economics, the generalised F model is used to represent complex hazard functions and to extrapolate long-term survival when simpler parametric models provide an inadequate fit.
Mathematically, the generalised F model specifies survival times as following a generalised F distribution with location, scale and shape parameters. These parameters allow the hazard function to assume a wide variety of forms, including increasing, decreasing, bathtub-shaped and non-monotonic hazards. Model parameters are estimated using maximum likelihood estimation, and model selection is typically based on statistical goodness-of-fit together with clinical plausibility.
In practice, the generalised F model is fitted to individual patient survival data using specialised statistical software. It is commonly evaluated alongside Weibull, Gompertz, log-normal, log-logistic and generalised gamma models when selecting an appropriate survival function for health technology assessment. The fitted model is subsequently used to estimate long-term survival, life expectancy, quality-adjusted life-years and lifetime costs.
Purpose
Used to model complex survival patterns, provide flexible long-term survival extrapolation and improve the estimation of health outcomes and cost-effectiveness when conventional parametric survival models are insufficient.
Mathematical Formulae
Primary Formula
ln(T) ~ GF(?, �, Q, P)
where:
- T = survival time
- ? = location parameter
- � = scale parameter
- Q = first shape parameter
- P = second shape parameter
Supporting Formulae
Survival function:
S(t) = 1 ? F(t)
Hazard function:
h(t) = f(t) / S(t)
Model parameters are estimated using:
?? = arg max L(?)
Related Mathematical Methods
- Maximum likelihood estimation
- Parametric survival modelling
- Generalised gamma model
- Weibull model
- Log-normal model
- Log-logistic model
- Akaike Information Criterion (AIC)
- Bayesian Information Criterion (BIC)
Example
A clinical trial evaluating an immunotherapy demonstrates a complex hazard pattern that is poorly represented by Weibull and Gompertz models. A generalised F model is fitted to the patient-level survival data, producing a substantially better statistical fit and more clinically plausible lifetime survival estimates for calculating quality-adjusted life-years and incremental cost-effectiveness ratios.
Excel Implementation
| Function | Example Formula | Health Economics Application |
|---|---|---|
| LN | =LN(A2) | Calculate log survival times for exploratory analysis. |
| EXP | =EXP(B2) | Transform predicted log survival times back to the original scale. |
| LINEST | =LINEST(Y_range,X_range,TRUE,TRUE) | Perform exploratory regression prior to formal survival modelling. |
| SUMPRODUCT | =SUMPRODUCT(CoefficientRange,CovariateRange) | Calculate the linear predictor for model estimation. |
VBA (Optional)
Automate comparison of multiple parametric survival models, summarise goodness-of-fit statistics and generate long-term survival projections for economic evaluation.
Sources
- Cox C. An analysis of transformed survival data using the generalized F distribution.
- Prentice RL. A generalisation of the proportional hazards model for survival data.
- Royston P, Parmar MKB. Flexible parametric proportional-hazards and proportional-odds models for censored survival data.
- Briggs A, Claxton K, Sculpher M. Decision Modelling for Health Economic Evaluation.
- NICE. Health Technology Evaluation Manual.
Related Concepts (2)
Library
Publications
1
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.
Frequently Asked Questions (6)
What is the generalised F model?
A highly flexible parametric survival model encompassing several other distributions, including the exponential, Weibull, and log-normal, as special cases.
Source: Cox 2008
When is the generalised F model worth its complexity?
The generalised F model is the most flexible of the common parametric survival forms, with enough parameters to take on a very wide range of hazard shapes and to contain simpler distributions as special cases. This flexibility is worth its cost only when the data are rich enough to estimate its several parameters reliably, since with limited follow-up it can be unstable and overfitted. It is most useful as a general form within which simpler, better-behaved models can be compared. Latimer (2013) discusses its role.
Source: Latimer 2013
What distributions does the generalised F model include?
The generalised F model includes many common survival distributions as special cases, such as the exponential, Weibull, log-normal, log-logistic, and generalised gamma, obtained by fixing or restricting its parameters. This nesting means that fitting the generalised F model, or comparing its parameters, can indicate which simpler distribution best represents the data. By encompassing these forms within one flexible family, the generalised F model offers a comprehensive parametric approach from which the appropriate simpler model can emerge.
Source: Cox 2008
Why use the generalised F model?
The generalised F model is used for its flexibility in fitting a wide range of hazard shapes and for nesting many common distributions, so that the most suitable simpler model can be identified from the data rather than assumed. This is valuable in survival extrapolation, where the choice of distribution strongly affects long-term estimates and no single simple form may be clearly best. By providing a general framework, the model allows a systematic comparison of forms and can capture complex hazards, aiding the selection of an appropriate survival model.
Source: Latimer 2013
How flexible is the generalised F model?
The generalised F model is among the most flexible standard parametric survival models, since its several parameters allow it to represent a wide variety of hazard shapes, including those of the many distributions it nests. This flexibility lets it fit complex survival patterns that simpler distributions cannot. However, greater flexibility comes with more parameters to estimate, which requires sufficient data and can risk overfitting, so the model's flexibility is balanced against the reliability of estimation and the plausibility of extrapolation.
Source: Cox 2008
What are the limitations of the generalised F model?
The generalised F model's flexibility requires estimating several parameters, which demands sufficient data and can lead to unstable estimates or overfitting when data are limited, and its complexity makes it harder to fit and interpret than simpler distributions. As with any parametric model, its extrapolation beyond the data depends on the assumed form and remains uncertain. These limitations mean the generalised F model is used where the data support its estimation, often to identify a suitable simpler nested distribution rather than always as the final model.
Source: Latimer 2013
Trust Record
Verified by Dr Darrin Baines
British health economist
Professional identity: darrinbaines.org
Verification date: 20 Oct 2025
Content version: 1.0.0
Canonical Identity
- Persistent URI
- https://healtheconomics.wiki/concept/generalized-f-model
- Term code
- HE-EM-SM-028
Stable URI · Machine-readable · Resolvable · CC BY 4.0