Background: Before major non-pharmaceutical interventions were implemented, seasonal influenza incidence in Hong Kong showed a rapid and unexpected reduction immediately following the early spread of COVID-19 in mainland China in January 2020. This decline was presumably
associated with precautionary behavioural changes (e.g., wearing face masks and avoiding crowded places). Knowing their effectiveness on the transmissibility of seasonal influenza could inform future influenza prevention strategies.

Methods: We estimated the effective reproduction number (Rt) of seasonal influenza in 2019/20 winter using a Time-Series Susceptible-Infectious-Recovered (TS-SIR) model with a Bayesian inference by integrated nested Laplace approximation (INLA). After taking account of changes in under-reporting and herd immunity, the individual effects of the behavioural changes were calculated.

Findings: The model-estimated mean Rt reduced from 1.29 (95%CI, 1.27-1.32) to 0.73 (95%CI, 0.73-0.74) after the COVID-19 community spread began. Wearing face masks protected 17.4% (95%CI, 16.3%-18.3%) from infections, about half of the effect of avoiding crowded places (44.1%,
95%CI, 43.5%-44.7%). Within the current model, if more than 85% of people had adopted both behaviours, the initial Rt could have been less than 1.

Conclusion: Our model results indicate that wearing face masks and avoiding crowded places could have potentially significant suppressive impacts on influenza.