Skip to main content

Table 2 Multivariable analysis for the potential impact of NPIs on weekly ILI counts in Beijing and the Hong Kong SAR, 2020–2021

From: Assessing the impact of COVID-19 interventions on influenza-like illness in Beijing and Hong Kong: an observational and modeling study

Interventions

Beijing

Hong Kong SAR

For individual NPI indicators

 + C1: Closings of schools

− 0.046 (− 0.111, 0.020)

 + C2: Closings of workplaces

0.092 (0.027, 0.158)

 + C3: Cancelling public events

− 0.159 (− 0.220, − 0.099)*

− 0.023 (− 0.081, 0.036)

 + C4: Limits on gatherings

0.041 (0.099, 0.006)*

− 0.025 (− 0.065, 0.015)

 + C5: Closing of public transport

 + C6: Staying in place or at home

− 0.046 (− 0.099, 0.006)

0.140 (− 0.005, 0.285)

 + C7: Restrictions on internal travel

− 0.127 (− 0.203, − 0.051)*

− 0.002 (− 0.220, 0.216)

 + C8: Restrictions on international travel

0.030 (− 0.020, 0.080)

− 0.132 (− 0.187, − 0.077)*

 + H6: Mask wearing outside the home

− 0.101 (− 0.207, 0.004)

0.002 (− 0.081, 0.086)

For combined NPI indicators

 + C12: Closings of schools and workplaces

0.036 (− 0.024, 0.097)

 + C34: Cancelling public events or gatherings

− 0.041 (− 0.095, 0.014)

− 0.009 (0.071, 0.052)

 + C567: Restrictions on internal movement

− 0.068 (− 0.106, − 0.031)*

0.297 (0.095, 0.499)

 + C8: Restrictions on international travel

− 0.008 (− 0.061, 0.045)

− 0.115 (− 0.170 − 0.061)*

 + H6: Mask wearing outside the home

− 0.068 (− 0.183, 0.047)

0.001 (− 0.075, 0.076)

  1. The numbers presented here are the coefficient (95% CI) of the GAM models (Eq. 13), to reflect the potential impact of synthetical COVID-19 interventions on weekly ILI counts. A positive coefficient represents a potential effect of increasing the weekly ILI counts. A negative coefficient represents a potential effect of reducing the weekly ILI counts. Besides the above NPIs indicators, non-parametric terms of s(time), s(mean temperature change), s(relative humidity change) and population density change were also incorporated into the GAM models. NPI non-pharmaceutical intervention, ILI influenza-like illness, Hong Kong SAR Hong Kong Special Administrative Region, GAM generalized additive model, CIs Confidence intervals, COVID-19 Coronavirus disease 2019. –, not incorporated into the model due to the multicollinearity. *P < 0.05