EXCEL WITH YOUR ECONOMIC MODELS USING R
Author(s)
Discussion Leaders: Devin Incerti, PhD, Senior Research Economist, Precision Health Economics, Oakland, CA, USA Joseph Frank Levy, PhD, Assistant Scientist, Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA; Jeroen P Jansen, PhD, Lead Scientific Advisor - Open Source Value Project, Innovation and Value Initiative, Los Angeles, CA, USA
Presentation Documents
PURPOSE: Historically, economic models for health technology assessment (HTA) have been developed with specialized commercial software (such as TreeAge) or more commonly with spreadsheet software (almost always Microsoft Excel). Recently, there has been criticism that this software limits (i) clinical realism, (ii) quantification of decision uncertainty, (iii) transparency and reproducibility, (iv) reusability and adaptability, and (v) modern software engineering methods. The aim of the workshop is to demonstrate the advantages of modern programming languages and software techniques for health economic modeling with a focus on the use of R. Participants will learn how to perform a completely integrated cost-effectiveness analysis in a single software environment while utilizing tools from modern software engineering.
DESCRIPTION: The workshop will begin with an overview of commonly used programming languages such as R, Python, and Julia and a summary of packages relevant to HTA. Development of a model using best practices in software engineering will be compared to development of a model using spreadsheet software. The discussion will then turn toward developing completely integrated economic models that combine parameter estimation, simulation, and decision analysis in the same software environment. Appropriate statistical techniques for different types of models (e.g., decision trees, partitioned survival models, state-transition models, compartmental models) and available data (e.g., single clinical trial, evidence synthesis) will be reviewed. The workshop will conclude with an example cost-effectiveness analysis using a state-transition model. A novel multi-state network meta-analysis approach utilizing flexible parametric survival models (e.g., fractional polynomials) will be used for parameterization. Computationally efficient methods will be used to simulate model outcomes and represent decision uncertainty from a probabilistic sensitivity analysis. We hope our presentations and live demonstrations will encourage a lively discussion with audience members, who will be encouraged to ask questions and share their experiences.
Conference/Value in Health Info
Code
W15