Multilevel Regression and Poststratification Interface: Application to Track Community-level COVID-19 Viral Transmission

2024-05-09

AI Summary

During the COVID-19 pandemic, the absence of randomized community testing hindered accurate assessment of transmission dynamics. To address this, we propose a community-level transmission estimation method leveraging routine asymptomatic SARS-CoV-2 test data from elective surgical patients—treating these clinical tests as synthetic random samples. We develop a lightweight, public health–oriented interface integrating multilevel regression with post-stratification (MRP), incorporating selection bias correction and small-area modeling to enable stable, real-time, fine-grained tracking of transmission dynamics. Applied to Michigan, our approach demonstrates significantly improved estimation stability compared to conventional weighting methods. The methodology is implemented in an open-source R/Python toolkit, supporting both public deployment and reproducible analysis.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Security and Privacy: Data transparency and provenance
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics

Abstract

In the absence of comprehensive or random testing throughout the COVID-19 pandemic, we have developed a proxy method for synthetic random sampling to estimate the actual viral incidence in the community, based on viral RNA testing of asymptomatic patients who present for elective procedures within a hospital system. The approach collects routine testing data on SARS-CoV-2 exposure among outpatients and performs statistical adjustments of sample representation using multilevel regression and poststratification (MRP). MRP adjusts for selection bias and yields stable small area estimates. We have developed an open-source, user-friendly MRP interface for public implementation of the statistical workflow. We illustrate the MRP interface with an application to track community-level COVID-19 viral transmission in the state of Michigan.

Problem

• Estimates community-level COVID-19 viral transmission
• Uses synthetic random sampling from asymptomatic patients
• Adjusts sample bias via multilevel regression and poststratification

Innovation

• Open-source MRP interface for public use
• Proxy method for synthetic random sampling
• Multilevel regression adjusts sample representation