A Flexible Framework for Local-Level Estimation of the Effective Reproductive Number in Geographic Regions with...

Area of Science:

• Epidemiology
• Biostatistics
• Spatial modeling

Background:

• Accurate local-level estimation of the effective reproductive number (R_t) is crucial for public health preparedness and resource allocation.
• Geographically granular R_t estimation faces challenges due to sparse or missing infectious disease outcome data in certain regions.

Purpose of the Study:

• To develop and validate a flexible statistical framework for small-area estimation of R_t.
• To integrate existing R_t estimation methods with spatial modeling to predict R_t in data-scarce regions.

Main Methods:

• A two-step approach combining established R_t estimation procedures (EpiEstim, EpiFilter, EpiNow2) with a covariate-adjusted Bayesian Integrated Nested Laplace Approximation (INLA) spatial model.
• The framework allows for the incorporation of any R_t estimation procedure for regions with limited or absent data.

Main Results:

• The proposed method demonstrated high predictive accuracy for R_t in regions with missing data, validated through external validation and a simulation study.
• EpiNow2, when used in the first step, showed the highest accuracy in predicting R_t in data-deficient areas.
• Median county-level percentage agreement (PA) reached 90.9% and 92.5% for Wave 1 and Wave 2, respectively; zip code-level PA reached 95.2% and 96.5%.

Conclusions:

• The developed methodology provides a robust tool for small-area estimation of the effective reproductive number.
• The flexible framework ensures high prediction accuracy even with coarse or missing data, supporting targeted public health interventions.