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

We analyzed 3,184 cases of coronavirus disease in Japan and identified 61 case-clusters in healthcare and other care facilities, restaurants and bars, workplaces, and music events. We also identified 22 probable primary case-patients for the clusters; most were 20–39 years of age and presymptomatic or asymptomatic at virus transmission.


DISPATCHES
We analyzed 3,184 cases of coronavirus disease in Japan and identified 61 case-clusters in healthcare and other care facilities, restaurants and bars, workplaces, and music events. We also identified 22 probable primary case-patients for the clusters; most were 20-39 years of age and presymptomatic or asymptomatic at virus transmission.
functions; and 1 (2%) transportation-related incident in an airplane. Most (39/61; 64%) clusters involved 5-10 cases ( Figure 1, panel B). The largest cluster involved >100 cases in a hospital, including nosocomial infections and staff infections. The largest non-healthcare-related cluster we observed was among >30 persons who attended a live music concert, including performers, audience members, and event staff. Healthcare and care facilities accounted for >50% of clusters at epidemiologic weeks 11 and 14 ( Figure 1, panel C). We identified 22 probable primary case-patients who had symptoms develop before they had contact with other case-patients in a cluster or who had prior epidemiologic links before contact with a cluster. We did not identify probable primary cases for nosocomial clusters. We believe these 22 case-patients contributed to the incidence of clusters. Demographic data show that 9 (41%) probable primary case-patients were female and 13 (59%) were male. The most frequently observed age groups among probable primary cases were 20-29 years (n = 6; 27%) and 30-39 years (n = 5, 23%) ( Figure 2, panel A). For 16 clusters, we determined the date of transmission from probable primary case-patients to other case-patients in a cluster and found 41% (9/22) of probable primary casepatients were presymptomatic or asymptomatic at the time of transmission; only 1 had a cough at the time of transmission ( Figure 2, panel B). Of the 22 probable primary case-patients, 45% (10/22) had cough at the time of diagnosis. Of the 16 probable primary case-patients with the determined date of transmission, transmission occurred one day before illness onset for 5 (31%) case-patients and on the same day of illness onset for 4 (25%) case-patients ( Figure 2, panel C). All age groups demonstrated presymptomatic or asymptomatic transmission.

Conclusions
We investigated clusters of COVID-19 cases and probable primary cases in Japan during January 15-April 4, 2020. We found that healthcare facilities, such as hospitals, and care facilities, such as nursing homes, were the primary sources of clusters, some of which had >100 cases. Japan experienced 2 waves of imported COV-ID-19 cases, after which local transmission occurred and the epidemic grew (8). Of note, clusters of COVID-19 cases at healthcare and care facilities predominated at epidemiologic weeks 11 (March 9-15) and 14 (March 30-April 4), which corresponds to ≈3 weeks after the 2 waves of imported cases (Figure 1, panel C). Healthcare and care facilities might be located at the end of the local transmission chain because clusters in those facilities only became evident several weeks after community transmission persisted.
We noted many COVID-19 clusters were associated with heavy breathing in close proximity, such as singing at karaoke parties, cheering at clubs, having conversations in bars, and exercising in gymnasiums. Other studies have noted such activities can facilitate clusters of infection (9,10). Japan's Prime Minister's Office and the Ministry of Health, Labour and Welfare announced 3 situations that could increase the risk for COVID-19 cases and advised the population to avoid the "Three Cs": closed spaces with poor ventilation, crowded places, and close-contact settings (11).
Among the probable primary COVID-19 cases we identified from non-nosocomial clusters, half (11/22) were 20-39 years of age, which is younger than the age distribution of all COVID-19 cases in Japan (Figure 2, panel A). We do not know whether social, biological, or both factors play a role in the difference in transmission patterns between the younger and older persons. We also noted probable primary COVID-19 case-patients appear to transmit the virus and generate clusters even in the absence of apparent respiratory symptoms, such as cough.
Our study has some limitations. The epidemiologic investigation mostly relied on voluntary cooperation. Because some case-patients could not disclose contact history, epidemiologic links and clusters of cases might have missed. Recall bias is likely because Japan did not use digital devices for contact tracing and information was obtained only through interviews. In addition, we could not calculate a secondary attack rate from probable primary cases because data were unavailable for denominator, such as the number of persons present in the places where clusters of cases were detected.
Active case finding and investigation are key to establishing links to other cases or transmission events. Detecting clusters of cases can lead to effective quarantine of close contacts and to the identification of risk factors for the formation of such clusters (12). Our findings provide further information and insight on clusters of COVID-19 cases in communities that can aid in the ongoing efforts to curb the global pandemic. Age ranges of probable primary COVID-19 cases in clusters. Age distribution among all COVID-19 cases in Japan is provided as reference. B) Proportions of symptoms among probable primary cases of COVID-19 clusters at transmission (n = 16) and among at laboratory confirmed diagnosis (n = 22). 1, Asymptomatic; 2, fever; 3, fatigue; 4, cough; 5, sore throat; 6, headache; 7, arthralgia or myalgia; 8, runny nose; 9, diarrhea; 10, difficulty breathing. C) Distribution of probable primary cases of COVID-19 clusters by time of transmission compared with illness onset by age groups (n = 16). Six cases were excluded because the time of transmission was undetermined.
This study was supported in part by grants from the Ministry of Education, Culture, Sport, Science and Technology in Japan (no. 16809810) and the Japan Agency for Medical Research and Development (no. 19fk0108104h1101).