# The neglected model validation of antimicrobial resistance transmission models – a systematic review

## TL;DR

This paper reviews antimicrobial resistance transmission models and finds significant gaps in validation and documentation, which hinder their usefulness for public health strategies.

## Contribution

The study expands the scope of antimicrobial resistance modeling to include new interventions and identifies persistent gaps in model validation and documentation.

## Key findings

- Most studies focused on Mycobacterium tuberculosis and Staphylococcus aureus resistance transmission.
- Population-based compartmental models were most commonly used, but lacked adequate validation and documentation.
- There is a need for better model validation and updated practices to improve policymaking effectiveness.

## Abstract

In the fight against antimicrobial resistance, mathematical transmission models have been shown as a valuable tool to guide intervention strategies in public health.

This review investigates the persistence of modelling gaps identified in earlier studies. It expands the scope to include a broader range of control measures, such as monoclonal antibodies, and examines the impact of secondary infections.

This review was conducted according to the PRISMA guidelines. Gaps in model focus areas, dynamics, and reporting were identified and described. The TRACE paradigm was applied to selected models to discuss model development and documentation to guide future modelling efforts.

We identified 170 transmission studies from 2010 to May 2022; Mycobacterium tuberculosis (n = 39) and Staphylococcus aureus (n = 27) resistance transmission were most commonly modelled, focusing on multi-drug and methicillin resistance, respectively. Forty-one studies examined multiple interventions, predominantly drug therapy and vaccination, showing an increasing trend. Most studies were population-based compartmental models (n = 112). The TRACE framework was applied to 39 studies, showing a general lack of description of test and verification of modelling software and comparison of model outputs with external data.

Despite efforts to model antimicrobial resistance and prevention strategies, significant gaps in scope, geographical coverage, drug-pathogen combinations, and viral-bacterial dynamics persist, along with inadequate documentation, hindering model updates and consistent outcomes for policymakers. This review highlights the need for robust modelling practices to enable model refinement as new data becomes available. Particularly, new data for validating modelling outcomes should be a focal point in future modelling research.

The online version contains supplementary material available at 10.1186/s13756-025-01574-x.