- Accelerated failure time model
In the statistical area of
survival analysis, an accelerated failure time model (AFT model) is a parametric model that provides an alternative to the commonly-used proportional hazards models. Whereas a proportional hazards model assumes that the effect of a covariateis to multiply the hazard by some constant, an AFT model assumes that the effect of a covariate is to multiply the predicted event time by some constant. AFT models can be therefore be framed as linear models for the logarithm of the survival time.
Comparison with proportional hazard models
The biggest difference is that AFT models are always fully parametric, i.e. a
probability distributionmust be specified, as there is no known equivalent of Cox's semi-parametric proportional hazards model. The choice of origin from which to measure time at risk is important in all parametric survival models.
Unlike proportional hazards models, the regression parameter estimates from AFT models are robust to the presence of unmeasured
confounders. They are also less affected by the choice of probability distribution. [Citation | journal=Statistics in Medicine| year=2004 |volume=23 |pages=3177–3192 |doi=10.1002/sim.1876| title=Parametric accelerated failure time models with random effects and an application to kidney transplant survival| first1=Philippe |last1=Lambert |first2= Dave |last2=Collett| first3=Alan| last3= Kimber |first4=Rachel |last4=Johnson]
The results of AFT models are easily interpreted. [Citation| title=On the use of the accelerated failure time model as an alternative to the proportional hazards model in the treatment of time to event data: A case study in influenza | journal=Drug Information Journal | year= 2002 | last1=Kay |first1= Richard| last2= Kinnersley|first2= Nelson| volume=36| pages=571–579| url=http://findarticles.com/p/articles/mi_qa3899/is_200207/ai_n9139743] For example, the results of a
clinical trialwith mortality as the endpoint could be interpreted as a certain percentage increase in future life expectancyon the new treatment compared to the control. So a patient could be informed that he would be expected to live (say) 15% longer if he took the new treatment. Hazard ratios can prove harder to explain in layman's terms.
probability distributions can be used in AFT models than parametric proportional hazard models, including distributions that have unimodal hazard functions.
Distributions used in AFT models
To be used in an AFT model, a distribution must have a parameterisation that includes a
scale parameter. The logarithm of the scale parameter is then modelled as a linear function of the covariates.
log-logistic distributionprovides the most commonly-used AFT model. Unlike the Weibull distribution, it can exhibit a non- monotonichazard function which increases at early times and decreases at later times. It is similar in shape to the log-normal distributionbut its cumulative distribution functionhas a simple closed form, which becomes important computationally when fitting data with censoring.
Weibull distribution(including the exponential distributionas a special case) can be parameterised as either a proportional hazards model or an AFT model, and is the only family of distributions to have this property. The results of fitting a Weibull model can therefore be interpreted in either framework.
Other distributions suitable for AFT models include the log-normal, gamma and
inverse Gaussian distributions, although they are less popular than the log-logistic, partly as their cumulative distribution functions do not have a closed form.
*Citation | title =Modelling Survival Data in Medical Research|first=D. |last=Collett | year=2003 | edition=2nd | publisher=CRC press| isbn=1584883251
*Citation| journal=British Journal of Cancer |year=2003 |number=89 |pages=431-436 | doi=10.1038/sj.bjc.6601119 |title=Survival Analysis Part II: Multivariate data analysis - an introduction to concepts and methods |first1=MJ| last1=Bradburn| first2= TG |last2=Clark |first3=SB | last3=Love | first4=DG |last4=Altman
*Citation |title=Fundamentals of Survival Data| first=Philip |last=Hougaard| journal=Biometrics| volume=55 |number=1| year=1999| pages= 13-22| url=http://links.jstor.org/sici?sici=0006-341X%28199903%2955%3A1%3C13%3AFOSD%3E2.0.CO%3B2-5
*Citation | title=Analysis of Survival Data | first1= David Roxbee | last1=Cox | author1-link=David Cox (statistician) | first2= D.|last2= Oakes | publisher=CRC Press | year=1984 | isbn= 041224490X
*Citation| title=Analysing Survival Data from Clinical Trials and Observational Studies| first1=Ettore| last1= Marubini| first2= Maria Grazia| last2=Valsecchi | year=1995| publisher=Wiley |isbn=0470093412
Wikimedia Foundation. 2010.