Quote from Eric Grant on 09/12/2023, 12:02 pm

• Provides a simple method for defining and fitting rate and risk regression models that go beyond the log linear models routinely available in most other packages. These include:
  • Excess relative rate and excess odds models for analyses dose response shape and effect modification
  • Excess risk/rate (risk/rate difference) models.
  • Additive and multiplicative joint effect models
• Model specification is simple but powerful. It is easy to:
  • add or remove covariates from a model
  • fix the values of individual parameters at specific value
  • specify boundary constraints for individual parameters,
• Powerful tools for the creation of detailed rate (person-year) tables
  • Multiple time scales (e.g., attained age, calendar time, smoking status and duration)
  • Time-dependent category variables (e.g., cumulative dose, pre-/post-menopause, number of pregnancies)
  • Multiple summary variables (e.g., multiple causes of death, current age/year/dose, pack years)
• Either fully parametric or stratified rate (Poisson regression) or risk models (unconditional logistic regression)