[Skip to Navigation]
Sign In
November 13, 2020

Hospital-Acquired SARS-CoV-2 Infection: Lessons for Public Health

Author Affiliations
  • 1Division of Infectious Diseases, Hospital of the University of Pennsylvania, Philadelphia
  • 2Division of Infectious Diseases, Montefiore Medical Center, Bronx, New York
  • 3Division of Infection and Global Health Research, School of Medicine, University of St Andrews, Fife, Scotland, United Kingdom
JAMA. 2020;324(21):2155-2156. doi:10.1001/jama.2020.21399

From the outset of the coronavirus disease 2019 (COVID-19) pandemic, it was clear that hospitals were an important setting for viral transmission. A review of 2 early case series in China estimated that 44% of 179 severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections were hospital acquired.1 An illustrative example of the devastating potential for health care transmission of SARS-CoV-2 came from St Augustine’s Hospital in Durban, South Africa, a facility with 469 beds, including 18 wards, 6 intensive care units, and 735 clinical staff.2 Through a detailed epidemiologic study supplemented by phylogenetic analyses, investigators documented how a single unsuspected case of SARS-CoV-2 led to 6 major clusters involving 5 hospital wards and an outside nursing home and dialysis unit, with infection ultimately confirmed among 80 staff members and 39 patients, 15 of whom died.2

Patients and health care workers around the world became concerned about the risk of providing routine care for patients with COVID-19, especially given the uncertainties about routes of transmission of SARS-CoV-2. Many settings intentionally limited or decreased elective visits and procedures during surging cases to reduce health care load and potential exposure. In addition, there have been substantial reductions in presentations for urgent or emergency illnesses, including myocardial infarction, tuberculosis, stroke, and hyperglycemic episodes. The consequences of these reductions are likely to contribute to excess mortality associated with the COVID-19 pandemic, considering that these conditions result in hundreds of thousands of deaths each year in the US.

Simultaneously, mounting evidence supports the effectiveness of a relatively simple intervention in reducing hospital transmission of SARS-CoV-2: universal use of surgical masks by health care workers and patients. This intervention appears to be effective, despite a relative inability to completely physically distance within the hospital. A study of 21 000 health care workers found that nosocomial-acquired SARS-CoV-2 infections decreased significantly after implementation of a universal masking policy, whereas community-acquired cases continued to increase, consistent with the overall community incidence.3

Another study of health care workers at 12 hospitals with more than 75 000 employees found that the SARS-CoV-2 test positivity rate among health care workers decreased linearly from 14.65% to 11.46% during a 3-week period after implementation of universal masking.4 A detailed contact tracing study of 226 patients exposed to health care workers with confirmed COVID-19 during the surge in Boston found only 1 possible transmission, and this was in the context of a 30-minute encounter during which both patient and health care worker were unmasked.5 In a complementary analysis of the same hospital system, researchers found no convincing cases of in-hospital transmission among more than 9000 admitted patients after implementation of universal masking for staff and patients.6

Together, these findings suggest that overall hospital transmission of SARS-CoV-2 in the setting of universal masking is likely rare, even during periods of high community prevalence. This has at least 2 important implications. First, it is imperative for the public to understand that hospitals with well-implemented universal masking policies are overwhelmingly safe, and public health messaging should emphasize that avoidance or delay of needed and urgent health care is not necessary. Second, despite the politicization of masking, this is an important and compelling proof of concept for the broader use of universal masking in crowded indoor settings, even with good ventilation. High-quality studies in other contexts can inform the extent to which efficacy differs with cloth masks or face coverings more typically used outside of hospital settings relative to surgical masks (and in some cases, especially during aerosol-generating procedures, use of N95 respirators) in hospitals. Masking will become even more important as the COVID-19 crisis continues into the upcoming influenza season in the northern hemisphere.

However, the World Health Organization recently suggested that health care workers account for up to 1 in 7 cases of COVID-19 worldwide. An analysis of more than 2 million community members and nearly 100 000 frontline health care workers in the US and the UK found an increased risk of having a positive SARS-CoV-2 test result among health care workers compared with community members, with rates of 2747 and 242 cases per 100 000 people, respectively, and a hazard ratio for a positive test result of 3.40 after adjusting for a wide variety of measurable factors.7 If hospital transmission with universal masking is so rare, what is the explanation for ongoing reports that health care workers account for a large proportion of cases worldwide?

The available data that suggest increased risk among health care workers must be interpreted with caution for several important reasons. First, many of these cases occurred either before the institution of universal masking protocols or in settings in which access or adherence to masks has been limited. Second, studies that have suggested higher risk of infection in health care workers have largely not been able to differentiate between community and hospital acquisition of infection. Health care workers may differ from the general public in their social exposure patterns outside the hospital. Third, health care workers are more likely to be tested than community members, and, even in studies that attempt to adjust for this differential in testing, estimates of risk of infection among health care workers relative to the community are almost certainly biased upward.

With these caveats in mind, detailed investigations of hospital outbreaks in the context of universal masking may be highly informative in identifying residual systemic weaknesses that allow transmission clusters to occur despite masking policies. These investigations, during which exhaustive contact tracing is routinely performed, could provide lessons that also could be applied to indoor settings outside the hospital, in which the specific nature of exposures is not as frequently available.

In the few outbreaks that have been reported in hospital settings during universal masking, several recurrent features have emerged. Duke Health in North Carolina found that “unmasked exposure to another [health care worker] rather than exposure to known infected patients resulted in the most [SARS-CoV-2] cases among staff after implementation [of universal masking].”8 Transmission leading to a cluster of at least 55 infections at Baystate Medical Center in Massachusetts in July 2020 was traced back to “staff who convened in a breakroom and removed their masks.”9 In an outbreak at Brigham and Women’s Hospital in Boston in September 2020 that was associated with infections in 42 health care workers and 15 patients, hospital epidemiologists identified a number of important contributing factors, including that “many patients were not masked during clinical care” and “[that there was a] lack of physical distancing among staff while unmasked while eating.”10 This pattern is unsurprising, especially given what is now known about the timing of infectiousness peaking at approximately symptom onset or just before; most patients with SARS-CoV-2 who require hospitalization are admitted at least 5 to 7 days after symptom onset, by which point they are probably minimally or noninfectious.

What is clear is that these hospital-based outbreaks have not revealed a failure of universal masking, but rather challenges in systems such as inadequate support to maintain masking adherence and basic human nature, in which individuals tire of masking. In particular, breakdowns have occurred in small workrooms and during mealtime in facilities that were not designed to allow adequate physical distancing during a respiratory pandemic. Although many in health care have focused on preventing transmission from patients with known SARS-CoV-2, these examples emphasize that residual transmission with universal masking appears to primarily occur after breaches in now-routine preventive measures. Lessons from the experiences of the millions of dedicated health care workers continuing to work during the COVID-19 crisis should be heeded by hospital systems and society in general to improve transmission prevention in all settings.

Hospital systems (and other indoor facilities) should focus on a number of provisions in addition to distributing a sufficient supply of masks to all staff and patients. Adequate, well-ventilated, and ideally dedicated space must be provided for breaks from daily work activities and mealtimes for health care workers, with processes in place to ensure that these are staggered to minimize contact and conversation during these higher-risk periods. Shared patient rooms should be avoided when possible, especially when local prevalence of infection is high, because of the possibility that patients might be admitted during the SARS-CoV-2 incubation period and because they must remove their masks to eat. The marginal benefit of universal eye protection should be evaluated, particularly during clinical encounters. Regular, flexible, and convenient testing with short turnaround times and adequate and statutory sick leave should be made available to all health care workers, with systems in place to ensure progression of training for medical trainees. Through these measures, transmission could be further minimized (and perhaps even eliminated), and emerging evidence could continue to direct policies designed to maintain safety in the hospital setting.

Back to top
Article Information

Corresponding Author: Aaron Richterman, MD, 3400 Spruce St, Philadelphia, PA 19104 (aaron.richterman@pennmedicine.upenn.edu).

Published Online: November 13, 2020. doi:10.1001/jama.2020.21399

Conflict of Interest Disclosures: None reported.

Zhou  Q, Gao  Y, Wang  X,  et al.  Nosocomial infections among patients with COVID-19, SARS and MERS: a rapid review and meta-analysis.   Ann Transl Med. 2020;8(10):629. doi:10.21037/atm-20-3324 PubMedGoogle ScholarCrossref
Lessells  R, Moosa  Y, de Oliveira  T. Report into a nosocomial outbreak of coronavirus disease 2019 (COVID-19) at Netcare St. Augustine’s Hospital. Published May 15, 2020. Accessed September 13, 2020. https://www.krisp.org.za/manuscripts/StAugustinesHospitalOutbreakInvestigation_FinalReport_15may2020_comp.pdf
Seidelman  J, Lewis  S, Advani  S,  et al.  Universal masking is an effective strategy to flatten the SARS-2-CoV healthcare worker epidemiologic curve.   Infect Control Hosp Epidemiol. Published online June 24, 2020. doi:10.1017/ice.2020.313PubMedGoogle Scholar
Wang  X, Ferro  EG, Zhou  G,  et al.  Association between universal masking in a health care system and SARS-CoV-2 positivity among health care workers.   JAMA. 2020;324(7):703. doi:10.1001/jama.2020.12897 PubMedGoogle ScholarCrossref
Baker  MA, Fiumara  K, Rhee  C,  et al; CDC Prevention Epicenters Program.  Low risk of COVID-19 among patients exposed to infected healthcare workers.   Clin Infect Dis. 2020;ciaa1269. Published online August 28, 2020. doi:10.1093/cid/ciaa1269PubMedGoogle Scholar
Rhee  C, Baker  M, Vaidya  V,  et al; CDC Prevention Epicenters Program.  Incidence of nosocomial COVID-19 in patients hospitalized at a large US academic medical center.   JAMA Netw Open. 2020;3(9):e2020498. doi:10.1001/jamanetworkopen.2020.20498PubMedGoogle Scholar
Nguyen  LH, Drew  DA, Graham  MS,  et al; Coronavirus Pandemic Epidemiology Consortium.  Risk of COVID-19 among front-line health-care workers and the general community.   Lancet Public Health. 2020;5(9):e475-e483. doi:10.1016/S2468-2667(20)30164-X PubMedGoogle ScholarCrossref
Advani  SD, Yarrington  ME, Smith  BA,  et al.  Are we forgetting the “universal” in universal masking? current challenges and future solutions.   Infect Control Hosp Epidemiol. Published online July 16, 2020. doi:10.1017/ice.2020.333Google Scholar
Baystate Medical Center (BMC) identified COVID-positive patients and employees on a non-COVID clinical unit, and took immediate steps to respond. Accessed October 5, 2020. https://www.baystatehealth.org/covid19/cluster-information
Brigham and Women’s Hospital. Statement for media regarding COVID-19 cluster. Accessed October 5, 2020. https://www.brighamandwomens.org/about-bwh/newsroom/press-releases-detail?id=3684
1 Comment for this article
How Infectious Are Hospital Patients?
Peter English, MB, ChB, FFPH | Public Health Physician
I wonder, when reading reports such as this, about how infectious hospital patients are, anyway.

The natural history of Covid-19, after all, is that people (if they have symptoms at all) have an initial period of flu-like illness, lasting about a week, and then - usually on about day 10 - a small proportion deteriorate and require hospital admission.

But we also know that viable potentially infectious virus is no longer present after about 8 days after the onset of symptoms. By day 10, the illness is caused by the immune system over-reacting to the infection.
/> While it is possible that people who are very ill with Covid-19 have more persistent viable virus, it is also possible that they are likely no longer to be infectious.

This could explain the observed low transmission rates in intensive care units, for example. It may not be that the staff on those units are better at observing PPE precautions etc - they may just have less infectious patients.

Of course, transmission has occurred in hospitals, and some patients have been infectious - but some of the transmission is staff-to-staff, or from patients admitted much earlier in the disease while still infectious, rather than when the hypoxia kicks in.

So reports like this, looking at evidence of efficacy of interventions to reduce transmission, could be looking in an area where the risks are already (and contrary to many people's expectations) very low.