A Bayesian estimator of the real-time case fatality rate in emerging pandemic to inform public health policy-making.

backgroundThe case fatality rate (CFR) is a crucial metric for monitoring ongoing epidemics. However, existing CFR estimators often fail to account for the time lag between symptom report and death, and cannot make real-time predictions. This limits their effectiveness in providing accurate and timely policy guidance during emerging epidemics, which motivates us to develop a more robust and accurate estimator.

methodsWe present a novel Bayesian real-time adjusted CFR (BrtaCFR) estimator that operates in real-time and requires only basic epidemiological count data. The estimator is based on the Poisson model within a Bayesian framework, incorporating prior knowledge of fatality rates and a fused LASSO component for stability of posterior real-time estimation.

resultsSimulation studies showed that the BrtaCFR estimator accurately captured various patterns of true fatality rates, outperforming traditional estimators. The BrtaCFR estimator demonstrated high sensitivity to changes in disease severity over time and remained robust across different hyperparameter settings. When applied to the Japan and Germany COVID-19 datasets, the estimator effectively captured the surveillance signals of both infection surges and implemented public health policies on the fatality rate across different pandemic waves.

conclusionThe proposed BrtaCFR estimator offers a more accurate and responsive tool for assessing disease severity in real-time during emerging epidemics. Accounting for reporting delays and incorporating prior knowledge of mortality rates, it provides a more reliable basis for public health decision-making. This approach could significantly enhance our ability to monitor and respond to evolving epidemic situations, potentially improving the effectiveness of public health interventions and resource allocation during future outbreaks. CLINICAL TRIAL: Not applicable.