Beyond QMRA: Modelling microbial health risk as a complex system using Bayesian networks.

Area of Science:

• Environmental microbiology
• Public health
• Computational toxicology

Background:

• Quantitative microbial risk assessment (QMRA) is standard for evaluating health risks from microorganisms.
• Current QMRA methods face limitations due to data availability and inability to incorporate diverse risk factors.
• Systems models, particularly Bayesian networks (BNs), are emerging as effective complementary approaches.

Purpose of the Study:

• To comparatively evaluate current QMRA methods and BN models for microbial risk assessment.
• To conduct a scoping review of recent literature utilizing BNs in QMRA.
• To discuss the advantages and disadvantages of systems approaches in this field.

Main Methods:

• A literature search identified 15 peer-reviewed articles using BNs for foodborne and waterborne pathogen QMRAs.
• Analysis focused on the application, uses, and benefits of BNs within QMRA studies.
• Comparative evaluation of BN capabilities versus traditional QMRA methods.

Main Results:

• BN applications in QMRA were diverse, including prediction, scenario assessment, and risk minimization.
• BNs demonstrated utility in reducing uncertainty and separating uncertainty from variability.
• Most reviewed studies focused on specific exposure pathway segments, highlighting broader potential.

Conclusions:

• BNs enhance QMRA through transparency, handling of poor-quality data, and support for causal reasoning.
• The systems modeling approach offers significant untapped potential for understanding complex microbial exposure-health relationships.
• BNs provide a flexible framework for advancing microbial risk assessment methodologies.