OBJECTIVES: Occasionally, meta-models have successfully been used to negate runtime issues associated with computational demanding analyses of health economic models. However, challenges remain regarding the selection of appropriate meta-modeling methods for use in health economics due to a lack of comprehensive guidance, detaining modelers from utilizing these methods. This study aimed to address these challenges by identifying meta-modeling methods applicable to health economics, including directions for their selection and application.

METHODS: Meta-modeling methods were identified from general (i.e., non-health economic specific) meta-modeling literature and clinical prediction modeling literature, since the process of meta-modeling shares many aspects with that of clinical prediction modeling. Additionally, results from a systematic literature search on meta-modeling applications in health economics were used. Methods were considered applicable to health economics, if they are able to account for mixed (i.e., continuous and discrete) input parameters and continuous outcomes (i.e., response variables).

RESULTS: A six-step process, existing of five mandatory steps and one optional step, has been identified for applying meta-modeling methods in health economics: 1) identification of suitable techniques, 2) data simulation, 3) meta-model fitting, 4) performance assessment, 5) analysis execution, and 6) results verification. For each step, applicable methods are selected, including their characteristics, directions for use, key references, and available R packages. For example, identified meta-modeling techniques were linear regression, symbolic regression, response surface methodology, multivariate adaptive regression splines, Gaussian processes, and neural networks.

CONCLUSIONS: Challenges regarding the selection of meta-modeling methods for use in health economics have been addressed as a first guide towards using meta-models to negate runtime issues with the analysis of health economic models. Identified methods were structured in a six-step application process, which includes practical guidance for method selection and application. This guidance may boost meta-modeling applications in health economics, enabling increase use of state-of-the-art analyses with complex simulation models.