A Novel Transmission Model-based Method to Evaluate Population Attribution Methods for Norovirus and Rotavirus...

Area of Science:

• Epidemiology
• Infectious Disease Modeling

Background:

• Estimating the burden of diarrheal diseases attributed to specific enteric pathogens relies on calculating the population attributable fraction (PAF).
• Conventional methods for PAF estimation include "detection-as-etiology" (pathogen prevalence in cases) and "odds-ratio" (pathogen detection vs. diarrhea), with a risk ratio method also used.

Purpose of the Study:

• To compare conventional PAF estimation methods against a novel model-based PAF derived from an enteric infection transmission model.
• To quantify the deviations of conventional PAF methods from a mechanistic model-based reference.

Main Methods:

• A transmission model was developed and fitted to site-specific qPCR data for norovirus and rotavirus from the eight-country MAL-ED cohort study.
• Model-based PAFs were calculated using equilibrium states, and deviations were quantified as crude differences from conventional PAF estimates (detection-as-etiology, odds-ratio, risk-ratio).

Main Results:

• The "detection-as-etiology" method consistently overestimated the population attributable fraction (PAF) for both norovirus and rotavirus.
• In contrast, the "odds-ratio" and "risk-ratio" methods showed small deviations from the model-based PAF, ranging from -5% to +3% across all sites.

Conclusions:

• A mechanistic transmission model offers an independent approach for quantifying pathogen-specific enteric disease burden and evaluating conventional PAF methods.
• The "detection-as-etiology" method demonstrates significant overestimation, while the "odds-ratio" and "risk-ratio" methods are validated as feasible and accurate measures for estimating enteric disease burden.