Title: Incorporating Sequencing Data to Forecast Future Pandemics: Quantifying Reporting Delay Bias for SARS-CoV-2 Proportions
 
Abstract: The COVID-19 pandemic revealed current forecasting limitations at the times when forecasts are most needed. One of the main limitations involves the inability to adequately forecast cases at infection points and transitions between variant waves. Incorporating variant and strain information may be the missing piece to improve case forecasting. In this talk, we present LANL-led efforts to include sequencing data to forecast future pandemics. We further delve into a specific aspect of data quality and reliability – the statistical bias of SARS-CoV-2 proportions that arises from reporting delays. Accurately estimating variant proportions is crucial for predicting the emergence of the next variant of concern, and this information can be leveraged to improve real-time disease monitoring. We use effect size measures (Cramer’s V) to explore the impact of reporting delay bias on SARS-CoV-2 variant proportion estimates across space and time. Evidence of reporting delay bias was highly variable across countries and variant waves. We further discuss how we will apply this work to improve variant nowcasting, disease forecasting, and other infectious diseases.

Zoom Information:

https://unm.zoom.us/j/98912959181

Meeting ID: 989 1295 9181
Passcode: 811422