Partitioned survival analysis TSD
A technical support document about partitioned survival analysis produced by the Nice DSU.
TSD 13
TSD 13 one-page summary (PDF, 2.7MB)
TSD 15
TSD 15: Cost-effectiveness modelling using patient-level simulation (PDF, 494KB)
TSD 15 one-page summary (PDF, 124KB)
Appendices
- Appendix A: Visual basic code for discrete event simulation example model (PDF, 140KB)
- Appendix B: R code for DES example model (PDF, 168KB)
- Appendix C: Simul8 documentation (PDF, 155KB)
- Code for Excel (.zip, 2.4MB)
- Code for R (.zip, 6KB)
- Code for Simul8 (.zip, 31KB)
- Code for TreeAge (.zip, 18KB)
- Code for DICE version (.zip, 335KB)
All the documents above are also available in a zip file (.zip, 3.3MB).
Related publications
Moller J, Davis S, Stevenson M, Cairo J. (2017). Validation of a DICE simulation against a discrete event simulation implemented entirely in code. Pharmacoeconomics
TSD 19
TSD 19: Partitioned survival analysis as a decision modelling tool (PDF, 529KB)
About
- TSD 13
A guide to identifying, selecting and using evidence to inform the model structure and its parameters, including recommendations for practice and reporting.
- TSD 15
Economic models can use either a patient- or cohort-level modelling approach to estimate the expected costs and outcomes across a particular population. In a patient-level simulation, the outcomes are modelled for individual patients and then the average is taken across a sufficiently large sample of patients. In a cohort-level model the outcomes are estimated for the cohort as a whole, without considering the outcomes for individual patients within that cohort.
The aims of this technical support document (TSD) on patient-level simulation are to:
Increase awareness of patient-level modelling approaches and highlight the key differences between patient-level and cohort-level modelling approaches. Provide guidance on situations in which patient-level modelling approaches may be preferable to cohort-level modelling approaches.
The focus of this document is on the application of patient-level simulation to systems where patients can be assumed to be independent as systems incorporating patient interaction are not a common feature with Technology Appraisal.
- Provide a description of how to implement a patient-level simulation using either a patient-level state-transition or discrete event simulation (DES) framework in a variety of software environments (Microsoft Excel®, R, TreeAge Pro®).
- Provide example models illustrating how to implement a simple cost-effectiveness model using a patient-level approach. The example models cover both DES (R, Visual Basic module within Microsoft Excel®, TreeAge Pro®) and patient-level state-transition (Microsoft Excel® and TreeAge Pro®) frameworks.
- Provide guidance on good practice when developing and reporting a patient-level simulation to inform a NICE TA.
- TSD19
Cost-effectiveness analyses informing NICE appraisals use a wide range of modelling approaches. Most of these have been subject to detailed discussion within the economic evaluation literature. This has not been the case for one approach, partitioned survival analysis, which is now used in a significant proportion of appraisals.
The objective of this technical support document (TSD) is to describe and critique this approach as a decision modelling tool in order to assist different stakeholders in determining its appropriateness as a modelling approach and basis for informing policy decisions.