Skip to main content

Clusters of Coronavirus Disease in Communities, Japan, January-April 2020

Our take —

From the 3,184 COVID-19 cases reported in Japan from January 15 to April 4 2020, this study found that there were 61 infection clusters of 5 or more cases outside of the household, and that most of them occurred in healthcare settings or crowded areas (e.g. restaurants). Probable index cases were identified for 22/61 (36%) of the clusters based on reported symptoms during the time of the congregation; for the other 64% the index case was unidentified, suggesting that transmission may have occurred during the pre/asymptomatic period. Among these probable primary cases, half were between the ages of 20-39 years. These findings further support the risk of high transmission in crowded areas and healthcare settings, including from adults not experiencing symptoms at the time; however, there are concerns of recall bias and potential misclassification as epidemiologic links were provided through self-reports.

Study design


Study population and setting

Data are from 3,184 laboratory confirmed cases (through polymerase chain reaction) of COVID-19 in Japan (including 309 cases imported from outside Japan) reported to the Japan Ministry of Health, Labor, and Welfare between January 15 and April 4, 2020. Investigators obtained data on the source of infection, contact history, and travel from case interviews conducted by local health officials for each case. A cluster of cases was defined as at least 5 cases or more that share the same primary exposure from a common event or place (aside from secondary household transmission). They also excluded cases that acquired infection from a secondary case.

Summary of Main Findings

After excluding imported cases, investigators found that 61 % of the cases (1760 cases) contributed to 61 clusters. These clusters included 18 from healthcare facilities, 10 from other care facilities, 10 from restaurants or bars, 8 from workplaces, 7 from music-related events, 5 from gyms, 2 from ceremonies, and 1 from an airplane. The largest cluster was over 100 cases in a hospital setting. Of the 61 clusters, investigators identified 22 probable primary cases who contributed to the clusters as they either had symptoms or prior epidemiologic links before contacting other cases. Thirteen of these cases were male while 9 were female. These 22 cases were substantially of younger ages, including half from those aged 20-29 years (27%) or 30-39 years (23%). Asymptomatic and pre-symptomatic transmission occurred across all age groups.

Study Strengths

This study draws from a large sample size (over 3,000 cases), which allowed investigators to see a high number of clusters.


The epidemiology links are all based on self-report and thus there are concerns of recall bias as symptomatic cases may have more accurate descriptions concerning timing and nature of transmission compared to asymptomatic or pre-symptomatic cases.

Value added

This study provides a description of the types of settings COVID-19 clusters can occur as well as shows the proportion of cases that can contribute to non-household related clusters.

This review was posted on: 26 July 2020