Coronavirus Disease Outbreak in Call Center, South Korea

Two hundred sixteen people work on the same floor of a Seoul high-rise. They share an open office, recycled air, the same hours, and the same headsets. A virus gets in. How many of them get sick? The answer that Park and colleagues documented in March 2020 is ninety-four. That’s nearly half the floor. This number sits at the center of one of the most forensically detailed workplace outbreak investigations of the early pandemic and highlights something fundamental about how the environment shapes transmission, not just biology. Building X is a nineteen-story mixed-use high-rise located in one of the busiest urban areas of Seoul. Commercial offices fill the lower floors, while residential apartments occupy the thirteenth through the nineteenth floors. A call center operates across four of the commercial floors — the seventh, eighth, ninth, and eleventh — employing eight hundred eleven people in total. On March 8, 2020, the Seoul Metropolitan Government was notified of a confirmed COVID-19 case linked to the building. The Korea Centers for Disease Control and Prevention and local governments formed a joint response team the next day. Building X was closed on March 9, and testing was offered to all occupants between March 9 and 12. The scale of the investigation was sweeping. Investigators identified one thousand one hundred forty-five persons under investigation — nine hundred twenty-two employees, two hundred one residents, and twenty visitors — and tested one thousand one hundred forty-three of them, which is a completion rate of ninety-nine point eight percent. Real-time reverse transcription polymerase chain reaction confirmed the results, with turnaround times of roughly twelve to twenty-four hours. To expand the effort even further, authorities used cell phone location data to identify anyone who had spent more than five minutes near the building and sent sixteen thousand six hundred twenty-eight text messages instructing them to avoid contact and seek testing. This was contact tracing operating at near-total coverage. Of those one thousand one hundred forty-three people tested, ninety-seven had confirmed COVID-19. This results in an overall attack rate of eight point five percent. However, the headline number conceals the real story, which is almost entirely about one floor. Of those ninety-seven confirmed cases, ninety-four worked in the eleventh-floor call center. The eleventh floor had two hundred sixteen employees and produced an attack rate of forty-three point five percent, with a ninety-five percent confidence interval of thirty-six point nine to fifty point four percent. Every other floor in the building was essentially untouched. Floors one through six reported zero cases. The seventh floor call center, which had one hundred eighty-two people, also had zero cases. The eighth floor call center, with two hundred seven people, reported zero cases. The ninth floor, with two hundred six people, had one case, which results in an attack rate of zero point five percent. The tenth floor reported two cases out of twenty-seven. The residential floors housing two hundred one people reported zero cases. The same building. The same elevators. The same lobby. Near-zero transmission on every other floor, and forty-three point five percent on the eleventh. This contrast is the central fact of this outbreak, and it demands an explanation. Park and colleagues point to the specific characteristics of the eleventh-floor environment. Call center workers talk — loudly, continuously, for hours. They sit close together in open-plan arrangements. And, critically, employees in this building did not generally move between floors. Despite what the paper describes as considerable interaction in elevators and the lobby, transmission was limited almost exclusively to the eleventh floor. The brief contacts in shared vertical circulation apparently weren’t enough. What drove transmission was the prolonged, sustained, face-to-face exposure of people sitting at adjacent workstations, day after day, in a high-density open-plan setting. Even within that floor, most confirmed cases were seated on the same side of the building. Duration and proximity, rather than shared air in passing, appear to have been the main factors driving this outbreak. Now, let’s follow the workers home. Park and colleagues tracked two hundred twenty-five household contacts of the ninety-seven confirmed case patients, resulting in an average of two point three household members per case. What they found was meaningful spread, but nothing like what happened at the worksite. Among symptomatic index patients, thirty-four household members were infected, leading to a household secondary attack rate of sixteen point two percent, with a ninety-five percent confidence interval of eleven point six to twenty-two percent. Compare that to forty-three point five percent on the call center floor. While the home is not a safe space during an outbreak, it presents a substantially less efficient transmission environment than a crowded shared workspace with continuous voice work. The asymptomatic findings further clarify the picture. Of the ninety-seven confirmed cases, investigators classified patients by symptoms at the time of testing and tracked them for fourteen days of quarantine. Eighty-nine, or ninety-one point seven percent, were symptomatic at the time of investigation. Four were presymptomatic: they tested positive before symptoms appeared and then developed illness within fourteen days. Surprisingly, only four, just four out of ninety-seven, remained entirely asymptomatic throughout the entire monitoring period. That’s one point nine percent of confirmed cases — a small fraction. None of the household contacts of those asymptomatic individuals developed an infection during the fourteen-day follow-up. Neither did the household contacts of the presymptomatic cases. Every documented household transmission in this investigation came through symptomatic index patients. Park and colleagues take care to note that aggressive containment — which includes early identification, quarantine, and intensive household monitoring — may have limited the opportunities for asymptomatic individuals to transmit. This means the true transmissibility of asymptomatic cases can't be fully resolved from this data alone. However, within this well-monitored cohort, the evidence indicates clearly that symptomatic people drove household spread, and the crowded call center environment amplified transmission before anyone was aware there was an outbreak to contain. Now, let’s discuss what ultimately stopped the outbreak. The timeline is crucial. The first symptomatic employee in this cluster reported symptoms on February 22. This employee was on the tenth floor. The second symptomatic case, on the eleventh floor, reported symptoms on February 25. By March 8, a case was flagged, and authorities were notified. By March 9, the building was closed, and testing had begun. By March 12, one thousand one hundred forty-three people had been tested. Every negative test result came with a mandatory fourteen-day quarantine, while every positive case resulted in isolation. Every household contact was identified, tested, and monitored for two weeks. The speed of each step was critical, as the virus does not wait. The paper explicitly discusses what this combination of actions achieved. Mass testing of all potentially exposed individuals, intensive contact tracing, rapid polymerase chain reaction turnaround, isolation of confirmed cases, and enforced quarantine for others — together, these measures likely interrupted transmission chains and prevented the outbreak from spreading beyond a single side of a single floor in a nineteen-story building in one of the most densely populated cities on Earth. The study does have clear limits. This case study focuses on one building and one outbreak, investigated during the early weeks of the pandemic. Investigators could not identify a definitive index case or trace the cluster back to another source. Not all clinical data by seat location were available. The findings should not be assumed to generalize to every crowded workplace without qualification. Nevertheless, the lesson Park and colleagues draw is both specific and applicable. The eleventh floor wasn't exceptional in its floor plan — open-plan call centers exist in cities worldwide. What made this outbreak containable was the speed and scope of the response, rather than anything unique about the building's architecture. The attack rate was forty-three percent on one floor, near-zero everywhere else, a secondary household attack rate that was real but substantially lower, nearly all confirmed cases were symptomatic, and there was an aggressive contact tracing operation that tested almost every potentially exposed person in the building within four days of the first notification. The outbreak in Building X provides a small, precise, and unusually complete picture of how SARS-CoV-2 behaved in a dense shared workspace, and what happens when a public health system moves fast enough to get ahead of it. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.