China's post-zero-COVID Omicron wave: A Bayesian analysis.

Following the late 2022 transition from its "dynamic zero-COVID" policy, China experienced a major nationwide severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Omicron wave. To characterize the wave's transmission dynamics, we used a Bayesian framework to fit a deterministic transmission model to two key data streams: reported COVID-19 daily case counts preceding the policy shift (up to November 11, 2022) and weekly virological and influenza-like illness (ILI) surveillance data afterward (through February 12, 2023). We estimated a nationwide cumulative infection attack rate reaching 87.8% (95% CrI: 75.9 to 93.3%) by mid-February 2023. Notably, 84.1% of the population became infected within just 1 mo following the full policy relaxation on December 7. The estimated time-varying effective reproduction number peaked at 5.69 (95% CrI: 4.56 to 6.85) on December 8, 2022. Although transmission intensity increased during the Spring Festival travel rush (Chunyun), widespread population immunity prevented a subsequent wave. Prior to the Chunyun period, distinct relationships emerged: Estimated transmission rates showed a significant positive correlation with long-term population behavioral response coefficient (reflecting cumulative infections; Pearson correlation: ρ = 0.92, P < 0.001), while mobility patterns correlated positively with short-term behavioral response coefficient (reflecting current infection prevalence; Pearson correlation: ρ = 0.87, P < 0.001). These dynamic behavioral associations, which we further validated against empirical data on keyword search and media coverage, then weakened during the Chunyun period. In summary, this analysis quantifies Omicron's transmission potential and highlights the importance of incorporating time-varying behavioral factors into epidemic models to accurately describe transmission dynamics, especially during periods of abrupt policy change.