A prototype software framework for transferable computational health economic models and its early application in youth mental health

TL;DR

This paper presents a software framework in R for creating transparent, reusable, and updatable health economic models, demonstrated through youth mental health applications in Australia, aiming for transferability across jurisdictions.

Contribution

It introduces a modular, user-friendly software framework for developing transferable computational health economic models, with initial implementation and application in youth mental health.

Findings

Framework modules meet transparency and reusability criteria

Successfully developed five CHEM modules for youth mental health

Framework shows potential for supporting transferable health economic models

Abstract

We are developing an economic model to explore multiple topics in Australian youth mental health policy. We want that model to be readily transferable to other jurisdictions. We developed a software framework for authoring transparent, reusable and updatable Computational Health Economic Models (CHEMs) (the software files that implement health economic models). We specified framework user requirements of a template CHEM module that facilitates modular model implementations, a simple programming syntax and tools for authoring new CHEM modules, supplying CHEMs with data, reporting reproducible CHEM analyses, searching for CHEM modules and maintaining a CHEM project website. We implemented the framework as six development version code libraries in the programming language R that integrate with online services for software development and research data archiving. We used the framework to author five development version R libraries of CHEM modules focused on utility mapping in youth mental health. These modules provide tools for variable validation, dataset description, multi-attribute instrument scoring, construction of mapping models, reporting of mapping studies and making out of sample predictions. We assessed these CHEM module libraries as mostly meeting transparency, reusability and updatability criteria that we have previously developed, but requiring more detailed documentation and unit testing of individual modules. Our software framework has potential value as a prototype for future tools to support the development of transferable CHEMs.