Weibull accelerated failure time model
A Weibull accelerated failure time model estimates how covariates stretch or shorten event times under a specified survival distribution.
Fit
When to use it, and when not to
Use it when
- You want a time ratio and a Weibull distribution is plausible for the event times.
- You can assess the fitted curve against observed follow-up and check the assumed shape.
Not when
- Hazards rise and then fall, or effects change the shape of the distribution. Consider a more flexible survival model. Read: Choosing a survival measure beyond the hazard ratio
- You want a time contrast without choosing a survival distribution. Consider restricted mean survival time through a supported horizon. Read: Choosing a survival measure beyond the hazard ratio
Before you start
What you need
- Event and censoring timesDefine the event and time origin. Keep right-censoring indicators.
- Covariates with clear unitsSpecify reference values and the change each time ratio compares.
- Software parameterizationCheck whether output describes log time or log hazard. Translate signs and scale parameters before interpreting.
- Loss to follow-up is unrelated to future events among people with the same recorded factors.How to check: Record why follow-up ended. Assess how results change if those lost have a worse outlook.
- Different people supply independent observations.How to check: For patients grouped within clinics or families, use inference that accounts for those groups.
- Missing covariates are handled under a stated assumption.How to check: Under missing at random, impute using observed predictors of missingness and outcome. If missingness is not at random, use delta or tipping-point sensitivity analyses.
- The Weibull shape is common across covariate profiles.How to check: Compare fitted and observed curves and inspect residuals. A Weibull hazard only rises, falls or stays constant.
- Covariates multiply the whole time scale.How to check: The same time ratio applies to all survival quantiles, not just the median.
Plain language
How it works
An accelerated failure time model works on log event time. Exponentiating a coefficient gives a time ratio. A ratio above one stretches the time scale, so corresponding survival quantiles occur later. A Weibull model also specifies a shape: above one gives a rising hazard, below one a falling hazard, and one a constant hazard.
This article teaches the Weibull accelerated failure time model, not a choice among survival distributions. For a common shape, its hazard ratio is the time ratio raised to minus that shape. Li and colleagues estimate both ratios in a primary application. Askari Tajabadi and colleagues apply a Weibull time model to prostate cancer. Wei explains the broader time-model framework; extrapolation remains an assumption to justify.
Worked example
A worked example
- Setting
- Assume a Weibull model with reference scale 10 months and shape 2.
- Question
- What does a positive log time coefficient mean?
- Approach
- The assumed fitted log time coefficient is 0.405465.
- Result
- The time ratio is 1.50, so the new scale is 15 months. Median survival changes from 8.33 to 12.49 months.
- What it means
- Under this common-shape Weibull model, the hazard ratio is 0.44. A longer time scale corresponds to a lower hazard.
In practice
Recent studies PubMed search · not vetted
Watch out
Common mistakes
The one to remember
The Weibull accelerated failure time model assumes a common shape and a monotone hazard. A time ratio is not a hazard ratio. A Kaplan-Meier curve that censors a competing death or another event overstates cumulative incidence when read as one minus survival.
- Applying the time coefficient sign as if it were a Cox coefficient.Instead: A positive log time coefficient lengthens survival. A positive log hazard coefficient increases the event rate.
- Assuming every accelerated failure time model has proportional hazards.Instead: The common-shape Weibull model has both representations. Other time distributions need not have that property.
Inside Mallard
How Mallard uses this
Learn search lists Weibull accelerated failure time model. The method registry links this method to its single-method article and code.
weibull-accelerated-failure-time).analysis.weibull_aft, the id a plan cue records.Related Learn pages
Related designs and methods
Evidence
Sources & further reading
- Weibull parametric model for survival analysis in women with endometrial cancer using clinical and T2-weighted MRI radiomic features. (opens in a new tab) · BMC Med Res Methodol · doi 10.1186/s12874-024-02234-1
- The Evaluation of Survival Rate in Patients with Prostate Cancer by Bayesian Weibull Parametric Accelerated Failure-Time Model. (opens in a new tab) · Iran J Public Health · doi 10.18502/ijph.v51i9.10566
- The accelerated failure time model: a useful alternative to the Cox regression model in survival analysis. (opens in a new tab) · Stat Med · doi 10.1002/sim.4780111409