Skip to main content

Impact assessment of non-pharmaceutical interventions against COVID-19 and influenza in Hong Kong: an observational study

Our take —

This study convincingly shows that a large, diverse package of surveillance and containment measures effectively suppressed transmission of both COVID-19 and influenza in Hong Kong. However, it is not possible to disaggregate effects of any particular intervention, which limits the usefulness of the results. Surveys show increases in mask-wearing and social distancing in the beginning of the epidemic, but these surveys had low response rates. It is unlikely that results are applicable to countries in a more advanced stage of the epidemic, or those in which such severe suppression measures are impractical.

Study design


Study population and setting

From January 17, 2020 to March 31, 2020 in Hong Kong, an extremely broad array of interventions was introduced to slow transmission of SARS-CoV-2, including aggressive surveillance, mandatory quarantines for close contacts of cases, severe travel restrictions, and bans on gatherings larger than 4 people. This study estimated combined effects of these interventions and behavioral changes on COVID-19 and influenza cases. COVID-19 and influenza activity was assessed via (i) daily lab-confirmed COVID-19 cases from 60 outpatient clinics; (ii) weekly sentinel surveillance data from outpatient clinics on “influenza-like illness”; and (iii) daily influenza hospitalizations in children from two hospitals. Behavioral changes were measured via three cross-sectional telephone surveys of adults (January 20-23, n=1008; February 11-14, n=1000; March 10-13, n=1005).

Summary of Main Findings

After the Chinese New Year in late January, more respondents reported wearing masks outside (75% in January, 98% in February, and 99% in March) and avoiding large gatherings (61%, 90%, and 85%). Results suggest that containment measures, social distancing, and changes in behavior contained the spread of SARS-CoV-2 in Hong Kong through March 31, 2020 by keeping the effective reproduction number at or below 1. Influenza transmissibility and hospitalization rates were estimated to drop by 44% and 33%, respectively, from the time preceding school closures in January to the first two weeks of closures.

Study Strengths

The first telephone survey was conducted in January, and is one of the few to capture self-reported behavior so early in the epidemic. Inclusion of influenza and influenza-like illness circumvents issues that may be specific to COVID-19 detection.


The simultaneous implementation of multiple containment measures and changes in behavior prohibited identification of the effects of specific policies and behaviors. Reductions in the transmission of influenza-like illness were used as a proxy for COVID-19, but this approach is only valid if the two are similar in terms of how they spread and how efficiently they are spread. Phone surveys used to measure changes in behavior had poor response rates (22%). Clinic closures over the Chinese New Year period, coupled with heavy self-reported avoidance of health clinics due to COVID-19, could have led to the reduction in influenza-like illnesses reported by the sentinel surveillance system.

Value added

This study is one of the first to analyze the impact of a package of surveillance, containment measures, and behavioral changes on the transmission of SARS-CoV-2.

This review was posted on: 27 April 2020