key: cord-0802180-63xc921p
authors: Liao, Qiuyan; Dong, Meihong; Yuan, Jiehu; Fielding, Richard; Cowling, Benjamin J.; Wong, Irene Oi Ling; Lam, Wendy Wing Tak
title: Assessing Community Vulnerability over 3 Waves of COVID-19 Pandemic, Hong Kong, China
date: 2021-07-03
journal: Emerg Infect Dis
DOI: 10.3201/eid2707.204076
sha: 19c2d8ea4d9ce8a66431c680f70468ed9e51c601
doc_id: 802180
cord_uid: 63xc921p

We constructed a coronavirus disease community vulnerability index using micro district-level socioeconomic and demographic data and analyzed its correlations with case counts across the 3 pandemic waves in Hong Kong, China. We found that districts with greater vulnerability reported more cases in the third wave when widespread community outbreaks occurred.

CVI. We calculated the Pearson correlations of indicator, domain, and overall CVI with COVID-19 case counts across districts and pandemic waves. We analyzed the differences in temporal trends of accumulated COVID-19 counts by districts of different vulnerability categories using Poisson regression models and plotted the results. We included 3,847 cases reported during January 23-August 31, 2020, for which a residence was locatable.

We plotted the spatial distribution of overall CVI and case counts ( Figure 1 ) and domain CVI by districts (Appendix Figure 2 ). The 4 districts with very high vulnerability districts reported 1,333 CO-VID-19 cases, accounting for 34.6% (95% CI 33.1%-36.2%) of the total cases; the 4 districts with very low vulnerability reported 491 COVID-19 cases, 12.7% (95% CI 11.7%-13.9%) of the total. Of the 81 COVID-19-attributed deaths with recorded residence, 45.7% (95% CI 34.6%-57.1%) were reported in the 4 districts with very high vulnerability and 34.6% (95% CI 24.3%-46.0%) were reported in the 3 districts with high vulnerability. Only 2 COVID-19attributed deaths were reported from the 7 very low or low-vulnerability districts.

By pandemic wave, the correlation between overall CVI and case counts was not significant for wave 1, negative in wave 2, and positive in wave 3, and, consequently, positive overall (Table 2 ). In wave 2, the case counts correlated negatively with most indicator and domain CVI but correlated positively with distribution of entertainment venues and non-Chinese ethnicities. In wave 3, the case counts correlated positively with most indicators and all domain CVIs but correlated negatively with non-Chinese ethnicities. Overall, community profile variables that correlated positively with case counts over the 3 waves included poverty, income, educational level, singleparent households, area of accommodation, hospital beds, population density, obesity, and working outside area of residency. We plotted the temporal changes in cumulative cases by vulnerability categories (Figure 2 ). The Poisson regression model revealed an overall significant effect of vulnerability levels on cumulative cases (moderate vulnerability, β = 0.17, p<0.001; high/very high vulnerability, β = 0.31, p<0.001). By pandemic wave, districts of high/very high vulnerability reported fewer cases in the first 2 waves (β = −0.45, p<0.001) but significantly more cases in wave 3 (β = 0.68, p<0.001).

Adding to existing literature (2) (3) (4) (5) , our study indicates that community vulnerability is dynamic, changing with the evolution of the pandemic. In waves 1 and 2, COVID-19 cases were mainly imported cases, including those infecting students and domestic helpers returning from overseas, as well as business travelers (12) . These cases were found mainly within more socially privileged families and thereby an inverse association between socioeconomic status and case counts was seen. In Hong Kong, 55% of persons with non-Chinese ethnicities are domestic helpers for more socially privileged families and another 25% are executives or professionals (13) who have greater work-from-home flexibility (12) . Subsequently, after tightening measures for inbound travelers and because there were more work-from-home arrangements in wave 3, the positive correlation between non-Chinese ethnicities and vulnerability to COVID-19 infection in waves 1 and 2 shifted to be negative. Entertainment venues constituted a primary exposure setting that spread COVID-19 in waves 1 and 2 (14) but ceased to be a major contributor to community vulnerability in wave 3 after these venues were closed. In wave 3, socioeconomic deprivation, poor housing, and dense household composition, as well as epidemiologic factors that facilitate viral transmission, became more key contributors to community vulnerability to COVID-19 infection. By April 2021, Hong Kong had experienced another pandemic wave (wave 4), characterized, again, by cases mainly in younger persons with higher socioeconomic status, linking to the largest local cluster (dancing/singing studio cluster) and the second largest cluster (fitness center outbreak) (10) . Updated analyses found that socioeconomic deprivation and poor housing were no longer major contributors, whereas entertainment venues again became strong contributors to community vulnerability in wave 4 (Appendix Table) .

Overall, our study indicates that a COVID-19 CVI can be applied to district-level data within a city to help city-level policy makers in resource allocation planning, but these measures should be viewed as dynamic at different pandemic stages. For instance, infection control and prevention measures should be intensified, perhaps by more strict or substantial social distancing in community settings with entertainment venues where persons may remove their face masks to exercise, dance, or eat and drink when community incidence is lower to minimize pandemic resurgence, whereas more material resources can be allocated to support social distancing measures among more socially disadvantaged communities when widespread community outbreaks occur. Our analysis focused on the correlation of CVI with COVID-19 case counts rather than infection risk (i.e., incidence) or severity (e.g., Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 27, No. 7, July 2021 fatalities) because of the relatively small number of cases and COVID-19 mortality in Hong Kong. However, because symptoms are generally mild in most cases, the magnitude of the pandemic impact is a key determinant for resource allocation.

Redefining vulnerability in the era of COVID-19

A vulnerability index for the management of and response to the COVID-19 epidemic in India: an ecological study

County-level association of social vulnerability with COVID-19 cases and deaths in the USA

The impact of social vulnerability on COVID-19 in the U.S.: an analysis of spatially varying relationships

Social vulnerability and racial inequality in COVID-19 deaths in Chicago

Surgo Foundation. The U.S. COVID-19 community vulnerability index (CCVI)

Development of a vulnerability index for diagnosis with the novel coronavirus, COVID-19

Statistics by subject: Hong Kong in figures

2016 population by-census

Hong Kong Home Affairs Department. Hong Kong: the facts-district administration

Latest situation of cases of COVID-19 (as of 10

Changing disparities in coronavirus disease 2019 (COVID-19) burden in the ethnically homogeneous population of Hong Kong through pandemic waves: an observational study

Population by-census thematic report: ethnic minorities

Settings of virus exposure and their implications in the propagation of transmission networks in a COVID-19 outbreak

Dr. Liao is an assistant professor of behavioral sciences and public health at the University of Hong Kong. Her research interests include public risk perception and risk communication in the context of communicable and noncommunicable diseases.