When outbreaks have been reported in hospitals that are using universal masking, unmasked exposure to other health care workers is often the cause. Transmissions have been traced to break rooms and cafeterias.
Early in the COVID-19 pandemic, mysteries abounded about this ninjalike disease striking a defenseless populace seemingly out of the blue. Who was most at risk? What were the best steps for protecting the vulnerable—when those were identified? Epidemiologic data on transmission were scarce, consisting mainly of small case reports, cohort studies, and governmental reports.1 A review of 40 studies suggested 44% of COVID-19 cases worldwide were nosocomial, but many hospitals were still coming to grips with what exactly they could do to prevent spread.2
As the pandemic took hold, some experts warned that rates of health care–acquired infections (HAIs) might rise because of patients with the virus being sicker, with longer lengths of stay, and other COVID-19–related factors. Over time, however, universal masking, “extreme” hand hygiene, and other intensified measures—when properly adhered to—were countering those predictions. Hospitals around the world were seeing striking drops in many HAIs.
A study at Kerman University of Medical Sciences in Iran, for instance, found the total rate of nosocomial infection for all illnesses was 3.7% during the first part of the pandemic, 20% lower than at the same time the previous year.3 A hospital in Spain saw the incidence of nosocomial Clostridioides difficile infection drop by 70% from March 11 to May 11, 2020, compared with the same time span the year before.4 In Singapore, investigators found enhanced infection prevention and control (IPC) measures had the “unintended positive consequence” of containing respiratory viral infections: Incidence shifted dramatically downward, from 9.69 cases per 10,000 patient days prepandemic to 0.83 cases.5
It is not surprising that intense precautions would also reduce cases of nosocomial COVID-19. A comprehensive infection control program implemented at Brigham and Women’s Hospital in Boston, Massachusetts, in March 2020 included dedicated COVID-19 units, personal protective equipment (PPE) in accordance with Centers for Disease Control and Prevention recommendations, donning and doffing monitors, universal masking, visitor restriction, and reverse transcriptase–polymerase chain reaction (PCR) testing of symptomatic and asymptomatic patients.6 Despite the high burden of COVID-19 in the hospital, only 2 patients were identified as having HAIs, 1 of whom was likely infected by a visiting spouse before visitor restrictions and universal masking. Between March and June 2020, of 8370 patients with non-COVID-19–related hospitalizations, 11 tested positive for the virus. Only 1 of the 11 was deemed an HAI—but with no known exposures inside the hospital.
The protective precautions have become routine. At Penn State Health Milton S. Hershey Medical Center, in Pennsylvania, limited visitation, screening of all employees and visitors upon each arrival, universal mask wearing, contact tracing, hand hygiene, and equipment cleaning “are essential parts of our everyday work,” says Fibi Attia, MD, MPH, CIC, infection control coordinator, and a member of the Infection Control Today® Editorial Advisory Board. All those measures “are definitely working together to limit any exposure.”
The least, and potentially the best, of these measures is masking. According to Aaron Richterman, MD, et al in JAMA: “[M]ounting 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 [HCWs] and patients,” even when physical distancing is not possible.7
It did not take long to find out that masking was a simple but effective measure to flatten the curve of nosocomial COVID-19. From March 15 to June 6, 2020, investigators at Duke Health (a system of tertiary care academic hospitals, community hospitals, and clinics) assessed all HCWs—of 21,014—who had tested positive for the virus. Of those cases, 38% were community acquired, 22% were health care associated, and 40% had no clear source. Of note, the investigators said, 80% of the HCWs did not work in COVID-19 units.8
Of the HAI cases, 70% were related to unmasked exposure to another HCW for more than 10 minutes, less than 6 feet apart; 30% were thought to be secondary (a non–COVID-19 infection) to direct care of patients who tested positive for the virus. However, only 1 week after the hospital system initiated universal mask wearing, the cumulative incidence rate of nosocomial COVID-19 dropped significantly, whereas the community-acquired cases and cases with no clear source continued to mirror the cumulative incidence rates in the communities surrounding Duke Health. At Mass General Brigham in Boston, a study involving more than 75,000 workers also found universal masking could quickly reverse the trajectory of nosocomial spread: Prior to the new rule, new infections among HCWs spiked from 0 to 21%; afterward, the positivity rate declined linearly from 14.65% to 11.46% during a 3-week period.9
Similarly, results of a study at the German Heart Center in Berlin showed that measures including obligatory surgical face masks for patients and employees, SARS-CoV-2 screening for all patients, and symptom-based testing for employees led to an overall 0.4% prevalence of positive tests (5 of 1128 tests). The incidence of new infections was 0.5%. No nosocomial infections were found, despite a mean number of 14.8 in-hospital contacts.10 Notably, the investigators said, during the observation period the number of positive SARS-CoV-2 tests sharply increased in Germany.
Nonetheless, “[m]y experience tells me that nosocomial COVID-19 is rampant,” Manoj Jain, MD, an infectious diseases physician and contributing health writer for the Washington Post, wrote in a January 2021 op-ed, “and hospitals are not sounding the alarm.”11
His concern was raised about a patient who died of nosocomial COVID-19. Jain was perplexed—how could the virus have reached the patient? It was not lack of handwashing. It was unlikely the virus had come from surfaces; the hospital’s environmental services staff had been diligent. All the family members who had visited the patient had tested negative.11
That was not the only patient either. Jain said in his op-ed that he had seen 5 other cases of nosocomial COVID-19 in the previous 4 weeks.
Were the staff the reason? Jain said he “armored” himself with N95 masks covered with a surgical mask and a face shield, plastic gown, and gloves. Still, one of his colleagues, who was taking similar precautions, tested positive. Again, there was no discernable cause. However, Jain noted, in the doctors’ lounge food is served, and health care workers frequently gather, “often taking off their masks to share COVID-19 war stories.”11
When outbreaks have been reported in hospitals that are using universal masking, unmasked exposure to other health care workers is often the cause. Transmissions have been traced to break rooms and cafeterias. Last November, more than 900 Mayo Clinic staff contracted COVID-19 in 2 weeks.
Amy Williams, MD, executive dean of clinical practice, in a briefing said 93% of staff contracted the virus in the community. Most of those who contracted it at work had been eating in a break room with a mask off. She later updated that to say they could have been exposed while eating with a mask off in campus cafeterias.12
One of the ways to prevent these “breaches in now-routine preventive measures,” Richterman et al wrote in JAMA, would be to provide adequate, well-ventilated and, ideally, dedicated space for breaks and meal times, perhaps staggering times to minimize contact and conversation.7
The unprecedented demands of this pandemic call for unprecedented responses. Abbas et al, in their literature review, found that many hospitals are breaking with traditional management of respiratory viral illness, particularly nosocomial outbreaks. In a “paradigm shift,” they said, hospitals are expanding testing criteria: testing asymptomatic patients/residents and health care workers, as well as conducting serial testing or repeat point-prevalence surveys. When Brigham and Women’s instituted liberal universal testing on admission, the investigators observed fewer late-onset cases (diagnoses on hospital day 3 or later).
At 2 Minnesota skilled nursing facilities (SNFs), serial testing identified COVID-19 cases among 64% of residents and 33% of health care workers. Following up with genetic sequencing revealed facility-specific clustering of viral genomes from HCW and resident specimens, suggesting intrafacility transmission.13
However, such information is not useful by itself: It needs to be supported by action. Although transmission was reduced by early identification of asymptomatic infections and prompt implementation of mitigation efforts, there were challenges. The Minnesota Long-Term Care COVID-19 Response Group found that, in the SNFs, continued SARS-CoV-2 transmission was “potentially facilitated” by “low baseline knowledge of and experience with [IPC] and PPE use,” delays of up to 12 days in receiving half of the HCW test results at one facility, and incomplete HCW participation.13
The Minnesota study is an example of how none of the measures is entirely effective on its own (although masking comes close). For instance, serial testing, investigators said, needs to be done until no new cases are detected after 14 days—along with IPC strengthening. Testing should be accompanied by “IPC education, flexible medical leave, and PPE resources targeted to this at-risk workforce.”13
Investigators at Changi General Hospital, Singapore, developed a prediction model to identify patients at low risk for COVID-19 to better guide resource allocation.14 They said that their risk prediction score would have obviated the need for isolation and testing in up to 41% of patients with pneumonia and acute respiratory infection. Missed cases of COVID-19 are “expected trade-offs” of their risk stratification strategy, they conceded, but they pointed to “reassuring reports” that basic infection control measures (eg, masks, hand hygiene) are effective in minimizing the risks of nosocomial spread.
Crucially, the precautions need to be carried out faithfully. As more viral strains evolve and people relax their vigilance because of the vaccines, more waves of the pandemic are only too possible. Can all this experience benefit hospitals and their workers in a future wave of this pandemic—or in a future pandemic?
Rhee et al said that their findings at Brigham and Women’s suggest that “robust and rigorous infection control practices” can minimize the risk of nosocomial spread of COVID-19, and “provide reassurance” as some health care systems reopen services and others continue to face surges.6
The basics can also be supplemented with more sophisticated tools. When there was no routine SARS-CoV-2 screening of asymptomatic HCWs, 5 hospitals in the United Kingdom assigned different locations within the hospitals as either green (virus negative) or red (positive) zones, combined with staff bubbles. That tactic helped keep nosocomial infections down. But as patients tested positive even after spending prolonged periods of time in green areas, the investigators realized that there were unrecognized transmission events between the 2 areas.15
They turned to viral genome sequencing as a realistic possibility to track and identify root causes of nosocomial transmissions. They sequenced SARS-CoV-2 genomes for patients and HCWs, obtaining 173 high-quality genomes. They then integrated patient movement and staff location data into the analysis to understand spatial and temporal dynamics of transmission. They identified 8 patient contact clusters with significantly increased similarity in genomic variants, compared with nonclustered samples.
The investigators found that incorporating the location of the HCWs identified additional links in transmission pathways and enhanced identification of outbreak clusters. Looking forward, they said, adopting genomic approaches in real time (eg, within 48 hours) along with consideration of patient movement data sets will enhance rapid identification of linked nosocomial SARS-CoV-2 infections. Their approach, they suggested, could optimize infection control management strategies, lead to targeted interventions, and ultimately prevent avoidable harm.
In his op-ed, Jain said he told his regional Memphis and Shelby County COVID-19 Joint Task Force, consisting of hospital executives and mayors, that hospital staff need to be regularly tested through “the affordable technique of pooling up to 20 test samples.”11 Such PCR gold-standard testing, he said, would amount to the cost of a cup of coffee and a few doughnuts. Without testing, he said, “we are flying blind.”
Like other HAIs, COVID-19 can be prevented, Jain said: “Hospitals should be places where patients come to get better, not sicker; where diseases are treated, not acquired.”
JAN DYER is a writer and editor specializing in clinical topics. She lives in Suffern, New York.