Exploring the role of replication in health economic decision models

Mcmanus, Emma (2023) Exploring the role of replication in health economic decision models. Doctoral thesis, University of East Anglia.

Preview

PDF Download (2MB) | Preview

Abstract

Several scientific disciplines have announced a “reproducibility crisis”, initiated by numerous high-profile studies being found to be unreproducible. Although health economic research has not been subject to such discussions, decision models are often termed ‘black boxes’ and there have been calls for heightened transparency in their reporting.

This thesis explores the role and value of replication within health economic decision modelling, specifically, how replicability is defined, what it means for a model to be replicable and the challenges facing modellers in incorporating replicability. This was achieved in a series of interlinked works. First, I identified studies defining replication success across all scientific disciplines. Whilst many studies discussed replicability, few defined replication success, none of which were found within health economics. Definitions ranged from subjective assessment to obtaining identical results. From these, definitions with varying specificity applicable to decision models were proposed.

Next, to examine factors influencing replication and assess the viability of the proposed definitions, five published models were replicated. This identified barriers and facilitators to replication and found that common reporting checklists were poor indicators of model replicability.

Finally, a decision model was developed with replicability in mind, to assess the feasibility of implementing the replication facilitators identified and overcoming the barriers. This highlighted the considerable time required to develop accessible models using open-source methods. These time requirements conflicted with the funded research project’s timeline, suggesting that in order to build replication into research, specific researcher time must be funded for replicability.

Overall, this thesis has shown that there is currently no consensus about how to define replication success for health economic models. Despite this, the importance of replication has been demonstrated as has the need for further work when reporting models to facilitate replication. To enable this, reforms to research infrastructure have been proposed.

Item Type:	Thesis (Doctoral)
Faculty \ School:	Faculty of Medicine and Health Sciences > Norwich Medical School
Depositing User:	Nicola Veasy
Date Deposited:	27 Mar 2024 14:09
Last Modified:	27 Mar 2024 14:09
URI:	https://ueaeprints.uea.ac.uk/id/eprint/94791
DOI:

Downloads

Actions (login required)

View Item