Metamodeling for Simulation-Based Optimization of Strategies in Healthcare
Description and Objectives:
It can be computationally challenging to perform analyses requiring many simulation model evaluations. For example, studies applying optimization algorithms to optimize healthcare strategies may require tens of thousands to millions of model runs. Despite advancements in computing power, the computational burden of such analyses quickly makes them infeasible to perform within acceptable time frames, in particular for patient-level simulation models. Consequently, often only a limited set of strategies are evaluated given that strategy optimization is infeasible. By substituting original simulation models by metamodels, the computational burden of running optimization algorithms can be substantially reduced. Metamodels can be seen as functions approximating the outcomes of original simulation models almost instantaneously. Metamodels are frequently applied in many research fields and recent research has demonstrated their value within health economics. To introduce modelers to the potential and application of metamodeling methods, this course will cover all the basics of using metamodels to negate runtime issues with simulation models. A single case study will be used throughout several hands-on exercises, in which metamodels will be developed to enable otherwise infeasible optimization of a population screening program using a computationally demanding health economic simulation model. R is used for all exercises, and several ready-to-use scripts are provided. Therefore, participants will be equipped with all the knowledge and tools required to develop metamodels for their own projects.
After the completion of this course, participants will be able to:
Explain the concept of metamodeling and when its use can support computationally challenging model-based analyses.
Explain the steps and design choices necessary for developing metamodels.
Distinguish between alternative metamodeling techniques and between alternative design of experiments, and select an appropriate technique and design of experiments based on specific study characteristics.
Perform a simple metamodeling study in R.
Become a Member
SMDM members contribute critically to health policy research in the areas of evidence-based medicine, cost effective health care, patient decision making and public health. SMDM helps you to be more than just a face in the crowd. The connections you make through SMDM can help you build a network of long-term contacts to help you throughout your career.
MDM offers rigorous and systematic approaches to decision making that are designed to improve the health and clinical care of individuals and to assist with health policy development.