AbstractInferring the transmission potential of an infectious disease during the low-incidence period following an epidemic wave is crucial for preparedness. In this period, necessarily scarce data hamper existing inference methods, blurring early-warning signals essential for discriminating between the likelihoods of resurgence versus elimination. Advanced insight into whether a region of interest will face elevating caseloads (requiring swift community-wide interventions) or achieve local elimination (allowing interventions to be relaxed or refocussed on controlling the importation of infections), can be the difference between decisive and ineffective policy. We propose a novel early-warning framework that formally maximises information extracted from low-incidence data to robustly infer the chances of sustained local transmission or elimination in real time, at any desired scale of investigation. Applying this framework, we decipher previously hidden disease-transmission signals from the prolonged low-incidence COVID-19 data of New Zealand, Hong Kong and Victoria state, Australia. We uncover how timely interventions averted dangerous, resurgent waves of COVID-19 and support official declarations of elimination. Across these locations, we obtain strong evidence for the effectiveness of rapid and adaptive COVID-19 responses.
Prospects for detecting early warning signals in discrete event sequence data: Application to epidemiological incidence data
