Validating Health Economic Models With the Probabilistic Analysis Check dashBOARD

Area of Science:

• Health Economics
• Computational Modeling
• Health Policy

Background:

• Health economic (HE) models are often opaque "black boxes" lacking transparency and hindering independent scrutiny.
• Systematic reporting of validation status for HE models is insufficient, impeding reliable health policy decisions.

Purpose of the Study:

• To develop and test a novel dashboard for systematically exploring and validating HE models.
• To enhance transparency and feasibility of validation for HE models.

Main Methods:

• Developed the Probabilistic Analysis Check dashBOARD (PACBOARD) incorporating literature review, health economist, and data scientist insights.
• Implemented functionalities for parameter/outcome validation, metamodeling for parameter-outcome relationships, and outcome prediction.
• Tested PACBOARD using mock HE models with introduced errors (e.g., negative costs, invalid utility values).

Main Results:

• PACBOARD successfully and automatically identified all introduced errors in the mock HE models.
• Metamodel predictions of incremental net monetary benefit closely matched original model outcomes, demonstrating accuracy.

Conclusions:

• PACBOARD offers a unique solution to improve the transparency and validation of HE models.
• The dashboard facilitates exploration of HE model workings through metamodeling, enhancing trust and usability in health policy.