Piecewise exponential additive mixed models (PAMMs) provide a flexible framework for analyzing censored and truncated time-to-event data, bridging classical hazard-based modeling with modern regression techniques. They enable the estimation of complex covariate effects, including non-linear and time-varying (cumulative) effects, and naturally incorporate time-varying covariates. Moreover, PAMMs are applicable across a wide range of survival settings, including non-proportional hazards, recurrent events, competing risks, and multi-state analyses. This article introduces the pammtools package, which facilitates data transformation, estimation, and interpretation for PAMMs within a unified workflow. The package provides a comprehensive, user-friendly, and extensible interface covering the full modeling pipeline, from data transformation to estimation and visualization. In addition, simulation-based inference allows the calculation of confidence intervals for arbitrary quantities of interest such as covariate dependent hazards, survival probabilities, restricted mean survival times and transition probabilities.

翻译：暂无翻译