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

BACKGROUND: The estimate of diarrhea burden attributed to a specific enteric pathogen-the population attributable fraction (PAF)-depends on the specific calculation method. Two conventional methods are commonly used to estimate the PAF for enteric infections: the "detection-as-etiology" method, which defines the PAF as the pathogen prevalence in diarrheal cases; and the "odds-ratio" method, which expresses the PAF as a function of the odds ratio between pathogen detection and diarrhea. A third, less frequently used method uses the risk ratio to quantify the strength of infection. METHODS: We compared each conventional PAF (detection-as-etiology, odds-ratio, or risk-ratio PAF) to a model-based PAF derived from a transmission model of enteric infection. We quantified deviation as the crude difference from this model-based PAF. We fitted the transmission model to site-specific qPCR data for norovirus and rotavirus detection from an eight-country birth cohort studying enteric infections (MAL-ED) and used the equilibrium states to calculate the model-based PAF. RESULTS: For both pathogens, the odds-ratio and risk-ratio deviations were small at all sites (ranging from -5% to +3%), whereas the detection-as-etiology method consistently overestimated the PAF and its deviation was the largest of the conventional methods. CONCLUSIONS: Our mechanistic model provides an independent alternative to conventional methods, quantifying pathogens-specific enteric burden and the deviations in those methods. Our model suggests the detection-as-etiology PAF estimates consistently deviated from our model-based reference, and validates the odds-ratio and risk-ratio methods as feasible, low-deviation measures for quantifying enteric burden.