Artificial neural network metamodel for sensitivity analysis in a total hip replacement health economic model

ABSTRACT

Objectives: Metamodels have been used to approximate complex simulations and have many applications with sensitivity analysis, optimization, etc. However, their use in health economics is very limited. Application of artificial neural network (ANN) with a health economic model has never been investigated. The study intends to introduce ANN as a metamodeling method to conduct sensitivity analysis in a total hip replacement decision analytical model and compare its performance with two other counterparts.

Methods: First, a nonlinear factor screening method was adopted to screen out unimportant factors from the simulation. Second, an ANN was developed using the important variables to approximate the simulation. Performance of the ANN metamodel was then compared with its Gaussian Process (GP) and multiple linear regression (MLR) counterparts.

Results: Out of 31, the factor screening method identified 12 important variables from the simulation. ANN metamodels showed best predictive capabilities in terms of performance measures (mean squared error of prediction, MSEP and mean absolute percentage deviation, MAPD) used for predicting both costs and quality-adjusted life years (QALYs) for two prostheses.

Conclusion: The study provides a methodological development in sensitivity analysis and demonstrates that an ANN metamodel is a potential approximation method for computationally expensive health economic simulations.

KEYWORDS:

• Metamodel
• neural network
• sensitivity analysis
• health economic model

Article Highlights

• Although metamodels have been used in approximating simulations, their use in health economics, especially in relation to a decision analytical model is very scarce.

• Most of the health economic models are nonlinear; hence, the underlying metamodel to approximate the original simulation should account for this potentially complex non-linear behavior of the response surface.

• Artificial neural network, which does not consider any input-output functional relationships of model parameters, can be an efficient metamodeling technique for approximating non-linear health economic model.

• Further research should focus on how artificial neural networks can be applied to computationally expensive individual patient simulations for the PSA and potentially VOI analysis.

Declaration of interest

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer Disclosures

Peer reviewers on this manuscript have no relevant financial or other relationships to disclose.

Author contributions

The study was designed, planned and developed by MFA with the help from AB. AB has provided his Total Hip Replacement health economic spreadsheet model. The initial manuscript was drafted by MFA, which was then revised and improved by incorporating AB’s comments and suggestions. Both authors have approved the final version and agreed to be accountable for all aspects of the work. The overall guarantor of the study is MFA.

Additional information

Funding

This paper was not funded.