simIC

The simIC package provides tools for simulating and analyzing interval-censored survival data, including left-, right-, and uncensored observations, using a variety of parametric distributions. It is useful for teaching, model development, and method evaluation in survival analysis.

✨ Features

• Supports commonly used parametric distributions:

  • Weibull
  • Exponential
  • Log-Normal
  • Logistic
  • Normal
  • Log-Logistic
  • Gamma
  • Gompertz
  • EMV (Extreme Minimum Value / Gumbel)

• Simulates survival data with interval, left, right, and uncensored observations using user-defined visit schedules (start_time, end_time) and an optional tolerance (uncensored_tol) for detecting exact event times.

• Provides two estimation functions:

  • mle_int()

    • Performs direct maximum likelihood estimation
    • Automatically detects and handles:
      • Interval-censored: contribution from F(Ri) - F(Li)
      • Left-censored: contribution from F(Ri)
      • Right-censored: contribution from 1 - F(Li)
      • Uncensored: contribution from f(ti)

  • mle_imp()

    • Uses imputation-based likelihood for interval- and uncensored data
    • For left- and right-censored, uses the proper likelihood contributions without imputation
    • Specifically:
      • Interval-censored: event times imputed from (Li, Ri) using midpoint, random, medians, or survival-based methods
      • Left-censored: contribution from F(Ri)
      • Right-censored: contribution from 1 - F(Li)
      • Uncensored: contribution from f(ti)