Abstract:Background Traditionally, meta‐analysis of time‐to‐event outcomes reports a single pooled hazard ratio assuming proportional hazards (PH). For health technology assessment evaluations, hazard ratios are frequently extrapolated across a lifetime horizon. However, when treatment effects vary over time, an assumption of PH is not always valid. The Royston‐Parmar (RP), piecewise exponential (PE), and fractional polynomial (FP) models can accommodate non‐PH and provide plausible extrapolations of survival curves beyond observed data. Methods Simulation study to assess and compare the performance of RP, PE, and FP models in a Bayesian framework estimating restricted mean survival time difference (RMSTD) at 50 years from a pairwise meta‐analysis with evidence of non‐PH. Individual patient data were generated from a mixture Weibull distribution. Twelve scenarios were considered varying the amount of follow‐up data, number of trials in a meta‐analysis, non‐PH interaction coefficient, and prior distributions. Performance was assessed through bias and mean squared error. Models were applied to a metastatic breast cancer example. Results FP models performed best when the non‐PH interaction coefficient was 0.2. RP models performed best in scenarios with complete follow‐up data. PE models performed well on average across all scenarios. In the metastatic breast cancer example, RMSTD at 50‐years ranged from −14.6 to 8.48 months. Conclusions Synthesis of time‐to‐event outcomes and estimation of RMSTD in the presence of non‐PH can be challenging and computationally intensive. Different approaches make different assumptions regarding extrapolation and sensitivity analyses varying key assumptions are essential to check the robustness of conclusions to different assumptions for the underlying survival function.

A parametric additive hazard model for time-to-event analysis

Semiparametric Analysis of the Additive Risk Model

A non-parametric proportional risk model to assess a treatment effect in time-to-event data

A nonparametric proportional risk model to assess a treatment effect in time‐to‐event data

Additive Hazards Regression Models for Survival Data

Semiparametric estimation for the functional additive hazards model

A generalized additive model approach to time-to-event analysis

Semiparametric Analysis of General Additive-Multiplicative Hazard Models for Counting Processes

Analysis of time-to-event for observational studies: Guidance to the use of intensity models

Flexible Bayesian survival modeling with semiparametric time-dependent and shape-restricted covariate effects

Marginal Proportional Hazards Models for Multiple Event-Time Data

Shape restricted additive hazards models: Monotone, unimodal, and U‐shape hazard functions

Methods for non-proportional hazards in clinical trials: A systematic review

Semiparametric Additive Modeling of the Restricted Mean Survival Time

Bayesian pairwise meta‐analysis of time‐to‐event outcomes in the presence of non‐proportional hazards: A simulation study of flexible parametric, piecewise exponential and fractional polynomial models

A proportional hazards model for interval-censored data subject to instantaneous failures

The hazard of using the Poisson model to cope with immortal time bias in the case of time-varying hazard

Nonlinear multiparametric modeling of life-table data with adaptive distributions: time evolution of hazard ratios

Bayesian inference in time-varying additive hazards models with applications to disease mapping

A semiparametric model for the cause-specific hazard under risk proportionality

Empirical Likelihood Method for General Additive-Multiplicative Hazard Models