Frequent Links
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 covariate is to multiply the hazard by some constant, an AFT model assumes that the effect of a covariate is to accelerate or decelerate the life course of a disease by some constant. This is especially appealing in a technical context where the 'disease' is a result of some mechanical process with a known sequence of intermediary stages.
Contents
Model specification
In full generality, the accelerated failure time model can be specified as^{[1]}
 <math>
\lambda(t\theta)=\theta\lambda_0(\theta t) </math> where <math>\theta</math> denotes the joint effect of covariates, typically <math>\theta=\exp([\beta_1X_1 + \cdots + \beta_pX_p])</math>. (Specifying the regression coefficients with a negative sign implies that high values of the covariates increase the survival time, but this is merely a sign convention; without a negative sign, they increase the hazard.)
This is satisfied, if the probability density function of the event is taken to be <math>f(t\theta)=\theta f_0(\theta t)</math>, from which is follows for the survival function that <math>S(t\theta)=S_0(\theta t)</math>. From this it is easy to see that the moderated life time <math>T</math> is distributed such that <math>T\theta</math> and the unmoderated life time <math>T_0</math> have the same distribution. Consequently, <math>log(T)</math> can be written as
 <math>
log(T)=log(\theta)+log(T\theta):=log(\theta)+\epsilon </math> where the last term is distributed as <math>log(T_0)</math>, i.e. independently of <math>\theta</math>. This reduces the accelerated failure time model into regression analysis (typically a linear model) where <math>log(\theta)</math> represents the fixed effects, and <math>\epsilon</math> represents the noise. Different distributional forms of <math>\epsilon</math> imply different distributional forms of <math>T_0</math>, i.e. different baseline distributions of the survival time. It is typical of survivalanalytic contexts, that many of the observations are censored, i.e. we only know that <math>T_i>t_i</math>, not <math>T_i=t_i</math>. In fact, the former case represents survival, while the later case represents an event/death/censoring during the followup. These rightcensored observations can pose technical challenges for estimating the model, if the distribution of <math>T_0</math> is unusual.
The interpretation of <math>\theta</math> in accelerated failure time models is straight forward: E.g. <math>\theta=2</math> means that everything in the relevant life history of an individual happens twice as fast. For example, if the model concerns the development of a tumor, it means that all of the prestages progress twice as fast as for the unexposed individual, implying that the expected time until a clinical disease is 0.5 of the baseline time. However, this does not mean that the hazard function <math>\lambda(t\theta)</math> is always twice as high  that would be the proportional hazards model.
Statistical issues
Unlike proportional hazards models, in which Cox's semiparametric proportional hazards model is more widely used than parametric models, AFT models are predominantly fully parametric i.e. a probability distribution is specified for <math>log(T_0)</math>. (Buckley and James^{[2]} proposed a semiparametric AFT but its use is relatively uncommon in applied research; in a 1992 paper, Wei^{[3]} pointed out that the Buckley–James model has no theoretical justification and lacks robustness, and reviewed alternatives.) This can be a problem, if a degree of realistic detail is required for modelling the distribution of a baseline lifetime. Hence, technical developments in this direction would be highly desirable.
Unlike proportional hazards models, the regression parameter estimates from AFT models are robust to omitted covariates. They are also less affected by the choice of probability distribution.^{[4]}^{[5]}
The results of AFT models are easily interpreted.^{[6]} For example, the results of a clinical trial with mortality as the endpoint could be interpreted as a certain percentage increase in future life expectancy on 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.
Distributions used in AFT models
The loglogistic distribution provides the most commonly used AFT model. Unlike the Weibull distribution, it can exhibit a nonmonotonic hazard function which increases at early times and decreases at later times. It is similar in shape to the lognormal distribution but its cumulative distribution function has a simple closed form, which becomes important computationally when fitting data with censoring. For the censored observations one needs the survival function, which is the complement of the cumulative distribution function, i.e. one needs to be able to evaluate <math>S(t\theta)=1F(t\theta)</math>.
The Weibull distribution (including the exponential distribution as 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. However, the biological applicability of this model may be limited by the fact that the hazard function is monotonic, i.e. either decreasing or increasing.
Other distributions suitable for AFT models include the lognormal, gamma and inverse Gaussian distributions, although they are less popular than the loglogistic, partly as their cumulative distribution functions do not have a closed form. Finally, the generalized gamma distribution is a threeparameter distribution that includes the Weibull, lognormal and gamma distributions as special cases.
References
 ^ Kalbfleisch & Prentice (2002). The Statistical Analysis of Failure Time Data (2nd ed.). Hoboken, NJ: Wiley Series in Probability and Statistics.
 ^ Buckley, Jonathan; James, Ian (1979), "Linear regression with censored data", Biometrika 66 (3): 429–436, JSTOR 2335161, doi:10.1093/biomet/66.3.429
 ^ Wei, L. J. (1992). "The accelerated failure time model: A useful alternative to the cox regression model in survival analysis". Statistics in Medicine 11 (14–15): 1871–1879. PMID 1480879. doi:10.1002/sim.4780111409.
 ^ Lambert, Philippe; Collett, Dave; Kimber, Alan; Johnson, Rachel (2004), "Parametric accelerated failure time models with random effects and an application to kidney transplant survival", Statistics in Medicine 23 (20): 3177–3192, PMID 15449337, doi:10.1002/sim.1876
 ^ Keiding, N.; Andersen, P. K.; Klein, J. P. (1997). "The Role of Frailty Models and Accelerated Failure Time Models in Describing Heterogeneity Due to Omitted Covariates". Statistics in Medicine 16 (1–3): 215–224. PMID 9004393. doi:10.1002/(SICI)10970258(19970130)16:2<215::AIDSIM481>3.0.CO;2J.
 ^ Kay, Richard; Kinnersley, Nelson (2002), "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", Drug Information Journal 36 (3): 571–579, doi:10.1177/009286150203600312
Further reading
 Bradburn, MJ; Clark, TG; Love, SB; Altman, DG (2003), "Survival Analysis Part II: Multivariate data analysis  an introduction to concepts and methods", British Journal of Cancer 89 (89): 431–436, PMC 2394368, PMID 12888808, doi:10.1038/sj.bjc.6601119
 Hougaard, Philip (1999), "Fundamentals of Survival Data", Biometrics 55 (1): 13–22, PMID 11318147, doi:10.1111/j.0006341X.1999.00013.x
 Collett, D. (2003), Modelling Survival Data in Medical Research (2nd ed.), CRC press, ISBN 1584883251
 Cox, David Roxbee; Oakes, D. (1984), Analysis of Survival Data, CRC Press, ISBN 041224490X
 Marubini, Ettore; Valsecchi, Maria Grazia (1995), Analysing Survival Data from Clinical Trials and Observational Studies, Wiley, ISBN 0470093412
 Martinussen, Torben; Scheike, Thomas (2006), Dynamic Regression Models for Survival Data, Springer, ISBN 0387202749
 Bagdonavicius, Vilijandas; Nikulin, Mikhail (2002), Accelerated Life Models. Modeling and Statistical Analysis, Chapman&Hall/CRC, ISBN 1584881860
