Emerging multidrug-resistant organisms (eMDROs) typically associated with health care, such as carbapenem-resistant Acinetobacter baumannii (CRAB), carbapenem-resistant Enterobacterales (CRE), carbapenem-resistant Pseudomonas aeruginosa (CRPA), and Candida auris represent a substantial threat to public health due to their ability to spread rapidly.^1,2 These organisms affect vulnerable patient populations and exploit gaps in infection control (IC) to spread silently from patient-to-patient within health care facilities.^3,4 During the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic, several eMDRO threats increased nationally, including clinical cases of hospital-onset CRE, hospital-onset Acinetobacter spp, hospital-onset multidrug-resistant P aeruginosa, and C auris.²

Outbreaks of eMDROs among SARS-CoV-2 patients in acute care hospitals (ACHs) have been reported, including clusters of C auris, CRAB, and CRE.^5–9 Factors thought to have contributed to the size and severity of these outbreaks included the presence of critically ill SARS-CoV-2 patients, staff and personal protective equipment (PPE) shortages, implementation of PPE supply conservation strategies (eg, extended use of isolation gowns), and general disruptions of facilities’ standard IC practices including reduced adherence to hand hygiene and PPE use, lack of cleaning of mobile medical equipment, and reductions in indwelling device maintenance rounds.^5–9 These reports have primarily described single-center outbreaks in ACHs and do not include reports from post–acute care facilities (PACFs) such as nursing homes. In addition, to date, no studies have systematically described the extent of eMDRO clusters and associated IC practices on SARS-CoV-2 treatment and observation units across different health care settings in the United States.

In response to multiple eMDRO clusters on SARS-CoV-2 units across different health care settings reported to the US Centers for Disease Control and Prevention (CDC) and health departments during the initial months of the SARS-CoV-2 pandemic, we aimed to systematically describe these clusters and IC practices in SARS-CoV-2 treatment and observation units among a convenience sample of US health departments.

Twelve HAI/AR programs (in 1 local and 11 state health departments) from 11 states, agreed to complete outbreak report forms that were submitted by March 12, 2021. Eleven of 12 HAI/AR programs reported ≥1 cluster of eMDROs involving SARS-CoV-2 treatment or observation units. One state health department HAI/AR program reported no clusters, and 1 state health department HAI/AR program only reported data for 2/4 (50%) clusters that occurred in their jurisdiction during the period of investigation. Overall, 18 clusters of CRE (n = 10), C auris (n = 6), CRPA (n = 1), and CRAB (n = 1) (Table 1) were reported: 4 clusters began prior to March 1, 2020, and continued during the SARS-CoV-2 pandemic and 14 began on or after March 1, 2020. Among the 17 clusters with information available, 5 (29%) were first recognized in a non-SARS-CoV-2 unit and later spread to a SARS-CoV-2 treatment or observation unit. Seven (41%) clusters occurred in facilities located in communities with moderate-to-substantial SARS-CoV-2 transmission (Appendix) during the cluster period, and 10 (59%) occurred in jurisdictions where the eMDRO was considered endemic or regional, or interregional spread had been identified.

The clusters affected 345 patients in 11 ACHs (median [interquartile range]: 29 [7, 36] patients per cluster) and 52 patients or residents in 6 PACFs (median [interquartile range]: 9 [4.5, 12] patients per cluster), all of which were skilled nursing facilities. One additional cluster in a long-term ACH did not have information available on the number of affected patients. Among the 17 clusters with information available, the pooled proportion of patients that were identified by screening tests was 65% (SE% = 2%), and the pooled proportion of patients with any clinical culture with an eMDRO isolated was 67% (SE% = 2%). The pooled proportion of patients coinfected with SARS-CoV-2 was 54% (SE% = 3%) among the 16 clusters with information available, and the pooled proportion of patients who expired within 30 days of specimen collection was 37% (SE% = 3%) among the 11 clusters with information available. ACHs had a lower proportion of patients in all outbreaks identified by screening tests (pooled proportion [SE]: 63% [3%], n = 11 clusters) compared with PACFs (pooled proportion [SE]: 77% [6%], n = 6 clusters; pooled z-test P value: .024); a higher proportion of patients coinfected with SARS-CoV-2 (pooled proportion [SE]: 62% [3%], n = 10 clusters) compared with PACFs (pooled proportion [SE]: 6% [3%], n = 6 clusters; pooled z-test P value: < .001); and a higher proportion of patients who expired within 30 days of specimen collection (pooled proportion [SE]: 43% [3%], n = 6 clusters) compared with PACFs (pooled proportion [SE]: 2% [2%], n = 5 clusters; pooled z-test P value: < .001). In one cluster, 83% of patients expired within 30 days of specimen collection. This was a large CRAB outbreak that occurred in 2 SARS-CoV-2 intensive care units at an ACH and involved 96 patients, all of whom were coinfected with SARS-CoV-2. Excluding this cluster, the pooled proportion of patients who expired within 30 days of specimen collection was 19% (SE% = 3%).

Reported changes in frontline HCP staffing were common during the cluster period. Ten (71%) facilities, among the 14 with information available, increased use of contracted or agency HCP relative to prepandemic practices, and 8 (53%) reassigned HCP to units with a different patient acuity from those where they typically worked (Table 2). In 7 (58%), cleaning duties were reassigned to HCP providing direct patient care. Shortages of alcohol-based hand sanitizers and difficulty obtaining a facility’s preferred disinfectant product(s) were reported in 5 (31%) and 8 (67%) facilities, respectively.

Among 15 facilities with information available regarding PPE availability during the cluster period, 9 (60%) reported experiencing a shortage of isolation gowns, and 1 (7%) reported experiencing a shortage of gloves (Table 2). Extended use of gowns, where HCP wore the same gown when interacting with more than 1 patient, was widely practiced, irrespective of actual shortage (8 [80%] ACHs and 3 [50%] PACFs); 4 (25%) facilities extended gown use between patients without considering whether patients had a history of multidrug-resistant organism (MDRO) colonization or infection. Gowns were also reused in 6 (67%) ACHs and 2 (33%) PACFs, such that the same gown was donned, doffed, stored, and then redonned for more patient encounters; in 2 (25%) ACHs, HCP extended the use of gowns and reused the same gown for multiple patients. In some facilities, HCP extended the use of gloves without changing them between patients or tasks (n = 3, 19%) or wore multiple layers of gloves (n = 4, 29%). All 3 facilities that reported extended use of gloves also reported performing hand hygiene over the gloves while still on at intervals when they would normally be changed.

ACHs reported substantial decreases in the frequency of auditing IC practices (Fig 1). Compared with practices in similar units prior to the pandemic, hand hygiene audit frequency decreased in 85% of affected units during the cluster period, and approximately half reported decreases in audits of PPE, environmental services, and indwelling device care. In contrast, most units in PACFs reported increases in hand hygiene and PPE audit frequencies during the cluster period (67% and 83%, respectively) (Fig 2).

HAI/AR programs described limited health care facility staff time, staffing shortages, and intermittent shortages of disinfectants, hand hygiene dispensers, and PPE as perceived barriers to implementing changes in outbreak facilities to prevent future eMDRO transmission. Facilities reported that even in the absence of an active SARS-CoV-2 surge, practices to conserve PPE supplies had become habitual since the onset of the pandemic and were employed in anticipation of future shortages. Competing priorities from other activities (eg, SARS-CoV-2 testing) impeded implementation of measures to mitigate eMDRO transmission, such as serial point prevalence surveys, contact tracing, and audits of IC practices, even though the frequency of hand hygiene and PPE practice audits increased in PACFs. Continued awareness of eMDRO transmissibility and close collaboration between health care facilities and health departments were cited as factors that mitigated transmission in some eMDRO clusters.

During the first 10 months of the SARS-CoV-2 pandemic and among a convenience sample of 11 states, we identified a total of 18 clusters of eMDROs in SARS-CoV-2 treatment and observation units. Most eMDRO clusters occurred in jurisdictions where the eMDRO associated with the cluster had been regularly identified in the region prior to identification of the cluster; only 1 cluster occurred in a jurisdiction that had not previously identified cases of the eMDRO and the only state that reported no eMDRO clusters has a very low underlying prevalence of these organisms. Although nearly half of reported clusters occurred in health care facilities in communities experiencing lower levels of community SARS-CoV-2 transmission, a far-greater proportion indicated implementing contingency and crisis capacity strategies¹³ to manage actual or anticipated shortages of PPE. Together, these results suggest that changes in PPE practices in outbreak units, including extended use and reuse of isolation gowns, were common and were not limited to facilities facing SARS-CoV-2 surges. Changes in frontline HCP staffing such as reassigning HCP to units they did not typically work and use of auxiliary or supplemental staffing, decreases in IC audits in ACHs, and shortages of alcohol-based hand sanitizer, disinfectant products, and isolation gowns were also common and might have contributed to eMDRO transmission.

In ACHs, multiple outbreaks of emerging eMDROs among SARS-CoV-2 patients have been reported.^5–9 Similar to previous reports, we found that shortages of resources combined with the widespread use of PPE supply conservation strategies and reduced frequencies of IC audits were common. While staffing and PPE shortages were widespread at health care facilities during the SARS-CoV-2 pandemic,¹⁴ multiple facilities that reported eMDRO clusters on SARS-CoV-2 treatment or observation units implemented PPE supply crisis strategies, such as extended use of gowns, when they were not experiencing a PPE shortage. These PPE supply crisis strategies are intended to be used temporarily during periods of known PPE shortages.¹³ In addition, 3 ACHs, only 1 of which was experiencing an actual shortage of isolation gowns at the time, implemented extended gown use, regardless of patients’ infection status of pathogens other than SARS-CoV-2. Unless no other options exist, extended use of gowns should only be considered if there are no additional co-infectious diagnoses transmitted by contact (such as Clostridioides difficile, C auris).¹⁵

Although multiple eMDRO outbreaks on SARS-CoV-2 treatment or observation units in ACHs have been reported in the literature,^5–9 we were unable to find similar reports in PACFs. However, our investigation demonstrates that these did occur and accounted for 39% of the reported clusters. Interestingly, many practices reported during PACF clusters would have been expected to decrease eMDRO transmission, such as increases in auditing frequencies for hand hygiene and PPE use. However, any benefits from these practices were likely offset by practices that facilitated cross-transmission, such as PPE supply conservation strategies, use of auxiliary or supplemental staff, and use of double layers of gowns and gloves, which has the added effect of exacerbating PPE supply shortages. PPE supply conservation strategies were commonly used nationwide early in the pandemic, with 94% of nursing homes implementing these measures.¹⁶ This may have contributed to unrecognized broad transmission of MDROs in these settings and is similar to findings from our investigation where at least 86% of PACFs implemented 1 or more PPE supply conservation strategies.

While our investigation was limited to a relatively small number of health departments reporting eMDRO clusters, nationwide, several urgent and serious antimicrobial resistance threats increased substantially during the SARS-CoV-2 pandemic.² These included clinical cases of hospital-onset carbapenem-resistant Acinetobacter spp (78% increase in 2020 vs 2019), hospital-onset CRE (35% increase in 2020 vs 2019), hospital-onset multidrug-resistant P aeruginosa (32% increase in 2020 vs 2019), and combined hospital- and community-onset C auris (60% increase in 2020 vs 2019 and 95% increase in 2021 vs 2020).^2,17 In ACHs, elevated rates of health care-associated MDRO infections have been associated with increases in hospitalized SARS-CoV-2 patients¹⁸ linked to their prolonged critical illness, overcrowding, higher antibiotic use, and shortages of HCP and PPE supplies.⁹ Changes in IC practices resulting from the strain placed on the US health care system also may have contributed. These nationwide increases in MDROs that were realized during the pandemic have the potential for long-term negative impact on the prevention and control of antimicrobial-resistant organisms in the United States.

There are several limitations to this investigation. First, data sources varied substantially between participating HAI/AR programs and responses are subject to recall bias. Only 10 (59%) facilities (6 ACHs and 4 PACFs) received on-site visits by health departments or regulatory agencies during the cluster period, for the remainder, HAI/AR programs relied on a mixture of facility self-report, telephone assessment, and video walkthroughs to evaluate practices and experiences during the cluster period. This likely led to underestimation of the prevalence of IC gaps. Second, we were not able to assess and describe additional factors that could impact eMDRO transmission, such as communication of patient MDRO status at transfer, screening, cohorting practices, and antibiotic use. Third, we collected data only on facilities with eMDRO clusters associated with SARS-CoV-2 treatment and observation units. We were not able to compare practices in these facilities to other facilities that did not experience eMDRO clusters, or those with eMDRO clusters not associated with SARS-CoV-2 units, so we were unable to assess whether these practices were also common in other units that did not experience eMDRO clusters. Fourth, for patients who expired in the 30 days after specimen collection, we were not able to assess the cause of death and the role eMDRO or SARS-CoV-2 infection may have played in their death, nor to differentiate between eMDRO-colonized and -infected individuals. Finally, we collected data from a limited number of HAI/AR programs for a short time in 2020. Although we intended this to include all eMDRO clusters that occurred on SARS-CoV-2 treatment and observation units during the period of investigation, 1 HAI/AR program was only able to submit data for 2/4 (50%) clusters due to competing interests during the SARS-CoV-2 pandemic. As such, the results are not nationally representative and do not assess whether practices changed later in the pandemic after PPE supply chain shortages were resolved.

In both acute care and PACFs with eMDRO clusters, basic IC practices had often been modified because of the SARS-CoV-2 pandemic. In the United States, although acute shortages in PPE have been alleviated, staffing challenges may continue to affect health care facilities¹⁹ and the long-term impact of the pandemic on IC practices and eMDRO transmission remains to be seen. While PPE supply conservation strategies were essential for keeping HCP safe during the SARS-CoV-2 pandemic, an unintended consequence of widespread implementation, combined with changes in staffing and other IC practices, might have been an increase in MDRO transmission. Now more than ever, due to the increasing prevalence of eMDROs, health care facilities require strong, well-supported IC programs to reverse the MDRO trends seen in recent years.