Audience: Of interest to HTA producers.
What you’ll learn: This is an introduction to survival analysis methods for model-based economic evaluations through practical examples in R. The advantages and disadvantages of each method will be discussed.
- Dr. Petros Pechlivanoglou, The Hospital for Sick Children (SickKids)
Note: Attendance is limited to 30 participants.
Abstract: Economic evaluations often rely on input from time-to-event or survival data. The unique characteristics of such data (i.e., right skewness, non-negative values, competing risks, censoring) require the use of more advanced statistical modelling techniques. Consequently, incorporating input from such statistical models into a model-based economic evaluation can be challenging. This course will teach participants how to appropriately integrate survival analysis data in decision models using R. We will provide an overview of the available methods (e.g., partitioned survival analysis, multi-state models, Markov/semi-Markov models, mixture cure models) and the advantages and disadvantages of each. The implementation of these methods in R will be outlined. Course participants will be asked to complete simple survival analysis exercises in the context of decision modelling to familiarize themselves with the main concepts. Finally, more advanced examples will be presented using different types of data (patient-level, life tables, digitized data from published curves). Participants will also discuss propagating parameter uncertainty from a parametric survival model to the outcomes of the decision model. By the end of the course the participants will be able to:
- understand the advantages and limitations of the different survival analysis methods when used in an economic evaluation context
- fit different parametric forms to survival data using R
- fit competing risks model and multi-state models in R
- integrate the results of different survival analysis models in a decision-modelling framework
- conduct probabilistic sensitivity analysis using survival models in R.
All R code used in the short course will be provided to participants for future use.