Model Estimation and Analysis: Survival Models

Nonparametric: Life table (Kaplan-Meier)

  • Survival rates
  • Hazard rates
  • Standard errors for the hazard function
  • Plot survival and hazard
  • Median and mean survival time
  • Variable number of intervals
  • Intervals and individual data calculations
  • Size of the risk set at the beginning of the period
  • Stratification
    • Log-rank test for homogeneity
    • Generalized Wilcoxon tests for homogeneity

Semiparametric (Cox proportional hazard)

  • Baseline hazard
  • Covariates
  • Stratification
  • Plot survival, hazard and integrated hazard functions
  • Keeping survival and integrated hazard values
  • Time dependent covariates
  • Ordered extreme value model
  • Ties handled explicitly

Parametric models

  • Distributions
    • Weibull
    • Exponential
    • Lognormal
    • Loglogistic
    • Generalized log-gamma
    • Generalized F
    • Gompertz
    • Inverse Gaussian
    • Weibull/exponential with gamma heterogeneity
    • Split population survival models
    • Latent normally distributed heterogeneity (frailty)
  • Censoring - left and right
  • Truncation
  • Keep survival probabilities and hazard rates
  • Time dependent covariates
  • Panel data models
    • Fixed effects
    • Random effects
    • Random parameters
    • Latent class
    • Time varying covariates
  • Robust covariance matrix (cluster)
  • Restrictions and tests: Wald, LM, LR
  • Stratification
  • Weights
  • Mean, median and percentiles of survival distribution (with limits)
  • Predicted survival, hazard, integrated hazard (generalized residuals)
  • Left and right truncation
  • Heterogeneity (heteroscedasticity) in the scale parameter