Discussion

The present study revealed that NO-COVID-19 occurred significantly more frequently in hospitalized patients who were in a four-patient room exposed to newly diagnosed COVID-19 hospitalized patients than those who were in a single private room and exposed to newly diagnosed HCPs. Since proactive identification of hospitalized patients or HCPs with COVID-19 as sources of NO-COVID-19 is challenging, infection control measures should focus on mitigating nosocomial transmission. The finding suggests that private room availability for hospitalized patients may have a significant impact on the incidence of NO-COVID-19.

The high risk of NO-COVID-19 observed among hospitalized patients who were exposed to a roommate with COVID-19 is plausible. First, the net time of exposure, from the time of initial infectivity to the time of diagnosis, was likely much longer among roommates compared with HCP–patient contact. Patients hospitalized in a four-patient room are only separated from each other by curtains, which do not adequately prevent inhalation of exhaled breath aerosol if a COVID-19 patient is present in the same room. Air circulation by the ventilation system and air conditioning may cause the spread of air containing SARS-CoV-2 aerosols in multi-patient rooms. In fact, a ventilation fan is typically placed at the entrance of a patient room. The air conditioning system is placed at the center of patient rooms to control the room temperature effectively. Since it is challenging to renovate hospital ventilation and air conditioning systems, alternative solutions, such as placement of air purifiers in the setting of suboptimal air ventilation, might be considered. ^{Reference Lindsley, Derk and Coyle6}

Moreover, some hospitalized patients might be minimally symptomatic or even asymptomatic despite having COVID-19, which also increases the opportunity of exposure. ^{Reference Oran and Topol7} Although universal masking is a hospital policy, hospitalized patients do not necessarily wear masks consistently, especially during the nighttime. In contrast, exposure from HCPs predominantly occurs in a private room when they provide meal assistance, rehabilitation, or other types of patient support. Given the limited encounter time and circumstances, the risk of nosocomial transmission of SARS-CoV-2 from infected HCPs is considered relatively smaller than in infected hospitalized patients.

The present study has some limitations. Although we performed contact tracing to determine the most plausible transmission route, unrecognized exposure to SARS-CoV-2 in the hospital might be present. Because the definition of NO-COVID-19 was based on its onset since the date of hospitalization (ie, the development of COVID-19 in patients hospitalized for over three calendar days), some cases classified as “NO-COVID-19” could have been due to exposure in the community prior to admission, leading to the overestimation on the incidence of NO-COVID-19. The impact of hospital visitors, a potential source of SARS-CoV-2 transmission on the incidence of NO-COVID-19, was not assessed in this study. Since SARS-CoV-2 genomic analysis was not performed, it is difficult to determine if NO-COVID-19 was in fact due to close contact. Lastly, the present study was conducted at a single institution, and the findings might not be generalizable to other institutions with different floor plans and ventilation systems.

NO-COVID-19 remains a significant challenge from the hospital infection prevention perspective, even after COVID-19 has entered an endemic phase. Our study suggests that the hospital occupancy and the source of SARS-CoV-2 contributed to the differential incidence of NO-COVID-19. An ideal hospital setting against the nosocomial spread of COVID-19 would be the universal use of private patient rooms. If the availability of private rooms is limited, then optimization of hospital room structures, including increases in the ratio of private patient rooms and renovation of ventilation/air conditioning systems, should be considered to prepare for future pandemics.