key: cord-0710592-bahvaprp authors: Rosser, Joelle I.; Tayyar, Ralph; Giardina, Richard; Kolonoski, Peter; Kenski, Diane; Shen, Peidong; Steinmetz, Lars M.; Hung, Li-Yuan; Xiao, Wenzhong; Bains, Karen; Morrison, Timothy; Madison, Alexandra; Chang, Sang-ick; Tompkins, Lucy; Pinsky, Benjamin A.; Holubar, Marisa title: Case-control study evaluating risk factors for SARS-CoV-2 outbreak amongst healthcare personnel at a tertiary care center date: 2021-09-15 journal: Am J Infect Control DOI: 10.1016/j.ajic.2021.09.004 sha: e1e312d6a6f03270d0252b3ae24a865ae8de315d doc_id: 710592 cord_uid: bahvaprp BACKGROUND: Despite several outbreaks of SARS-CoV-2 amongst healthcare personnel (HCP) exposed to COVID-19 patients globally, risk factors for transmission remain poorly understood. METHODS: We conducted an outbreak investigation and case-control study to evaluate SARS-CoV-2 transmission risk in an outbreak among HCP at an academic medical center in California that was confirmed by whole genome sequencing. Results: A total of 7/9 cases and 93/182 controls completed a voluntary survey about risk factors. Compared to controls, cases reported significantly more patient contact time. Cases were also significantly more likely to have performed airway procedures on the index patient, particularly placing the patient on high flow nasal cannula, continuous positive airway pressure (CPAP), or bilevel positive airway pressure (BiPAP) (OR=11.6; 95% CI=1.7-132.1). DISCUSSION: This study highlights the risk of nosocomial infection of SARS-CoV-2 from patients who become infectious midway into their hospitalization. Our findings also reinforce the importance of patient contact and aerosol-generating procedures as key risk factors for HCP infection with SARS-CoV-2. CONCLUSIONS: Re-testing patients for SARS-CoV-2 after admission in suspicious cases and using N95 masks for all aerosol-generating procedures regardless of initial patient SARS-CoV-2 test results can help reduce the risk of SARS-COV-2 transmission to HCP. Keywords: SARS-CoV-2, COVID-19, healthcare personnel, outbreak, case-control study, whole genome sequencing, contact tracing, epidemiology, aerosol generating procedure Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) was first documented to cause hospital outbreaks amongst healthcare personnel (HCP) early in the course of the pandemic (1) . HCP are at higher risk of SARS-CoV-2 infection due to close contact with infected patients and increased exposure to their respiratory droplets (2, 3) . In addition, inadequate supply of personal protective equipment (PPE) may have exacerbated HCP's risk when caring for patients with coronavirus disease 2019 (COVID-19) (4). There have been several reports of increased SARS-CoV-2 infections among HCP worldwide since the beginning of the pandemic (5) (6) (7) (8) . In a large prospective cohort study of over 2 million community individuals and nearly 100,000 HCP in the UK and USA, HCP were found to have a 3.4 times increased risk of SARS-CoV-2 infection compared to the general population (6) . Several SARS-CoV-2 outbreaks amongst healthcare workers have also been reported and directly linked to a single index patient, often a patient with a delayed diagnosis of COVID-19 due to negative tests early in the hospital course (8) (9) (10) (11) (12) (13) . Although these outbreaks are often relatively small, they provide critical information about risk factors for nosocomial transmission and can help guide infection control practices. The CDC stratifies the risk of COVID-19 transmission to HCP by duration or type of exposure. A prolonged exposure is defined as any cumulative exposure to SARS-CoV-2 infected patients or visitors for greater than 15 minutes in a 24-hour period. However, risk factors like exposure in confined spaces or performing an aerosol-generating procedure (AGP) are considered high risk for transmission regardless of exposure time (14) . HCP spend a variable amount of time in patients' rooms depending on their jobs and departments. A study in the UK showed increased rates of COVID-19 amongst doctors and nurses working in emergency medicine and critical care (15) . In the US, nursing staff have been shown to have higher COVID-19 rates and associated hospitalizations compared to other HCP; other clinical HCP also had higher rates compared to non-clinical support and infrastructure personnel (7, 16) . Although our understanding of COVID-19 hospital outbreaks and infection control measures have improved dramatically since the start of the pandemic, healthcare worker outbreaks continue to occur and the risk of SARS-CoV-2 infection for HCP exposed to various types of AGPs remains unclear. Evidence about the risks associated with exposure to AGPs performed on patients with COVID-19 who didn't require intubation is particularly lacking. Better understanding of risk factors for SARS-CoV-2 transmission, particularly in non-intubated patients, is critical to formulating infection control policies to prevent nosocomial transmission and protect HCP. We conducted an outbreak investigation and case-control study of nosocomial SARS-CoV-2 transmission amongst HCP that occurred in a well-resourced academic medical center that was found to originate in a non-intubated patient transferred from an outside hospital. We evaluated risk factors for transmission including work type, PPE use, participation in various AGPs and non-aerosol generating procedures (non-AGPs), and duration of time spent with the patient. A 72-year-old female with a pertinent past medical history of atrial fibrillation on amiodarone and rivaroxaban, chronic obstructive pulmonary disease requiring oxygen at baseline, and remote history of lung cancer status post radiation was transferred to an academic medical center for a carotid endarterectomy after being diagnosed with a left middle cerebral artery stroke at an outside facility. She was admitted to the intensive care unit in a 214 square foot private room with excellent ventilation (approximately 20 air changes per hour). She had a negative nasopharyngeal SARS-CoV-2 real-time polymerase chain reaction (RT-PCR) test on the day of transfer and another negative nasopharyngeal SARS-CoV-2 RT-PCR upon arrival to the hospital. On hospital day (HD) 3, she developed acute hypoxic respiratory failure; diagnostic work-up included a chest CT which revealed bilateral patchy ground glass and consolidative opacities and a negative nasopharyngeal viral respiratory pathogen PCR panel on HD4 which did not include SARS-CoV-2. Alternative etiologies were considered, and the patient was started on prednisone 60 mg daily for presumed amiodarone-induced lung toxicity. She required increasing oxygen support and received high flow nasal cannula (HFNC), continuous positive airway pressure (CPAP), and bilevel positive airway pressure (BiPAP). Due to clinical deterioration, she had a repeat nasopharyngeal SARS-CoV-2 PCR test on HD 12 which was positive. We performed a case-control study as part of an outbreak investigation conducted by the Infection Prevention and Control Department (IPCD). This quality improvement project was reviewed and deemed exempt by the Stanford University School of Medicine Panel on Human Subjects in Medical Research. All exposed HCP were tested for SARS-CoV-2 by RT-PCR by occupational health and were invited to complete a voluntary survey to characterize their exposure to the index patient. All individuals with a negative RT-PCR test who completed the survey and reported at least one day working with the index patient were included in the final analysis as controls. Survey data were directly entered by the employees into REDCap data management tool (17, 18) . The survey asked questions regarding demographic and work characteristics, COVID-19 symptoms, PPE use, care activities, and time spent with the patient. Potential demographic and work predictors of cases were assessed in univariate analysis by Fisher's exact test, linear regression, and ordinal regression where appropriate. Multivariable analysis was not performed due to the small sample size. Instead, given that most cases were nurses or therapists, a secondary restricted analysis was performed including only individuals who fell into occupational categories that contained at least one case. All data were analyzed using RStudio version 1.1.456. Whole genome sequencing was performed on samples from the index patient and all positive HCP. A multiplex PCR based technology developed at Stanford Genome Technology Center was used to perform the whole genome sequencing (19, 20) . A total of 62 PCR amplicons of ~540bp were designed to cover the entire SARS-COV-2 genome from position 46 to 29793 (NC_045512) (manuscript under preparation). The samples were pooled and sequenced on MiSeq. The SNP analysis was performed using MiSeq Reporter and an in-house software. Nextclade (https://clades.nextstrain.org) and GISAID EpiCoV (https://www.epicov.org) were then used to track the origin of this strain and assign it to its corresponding clade (21) . On the patient's HD 11, IPCD was notified of 3 HCP who worked on the same unit with positive nasopharyngeal SARS-CoV-2 RT-PCR tests. Of these 3 HCP, 1 was also in contact with another HCP from the same unit outside of work. On HD 12, the index patient was re-tested because of worsening clinical condition. After this result was available, the index patient's previous (non-SARS-CoV-2) viral respiratory sample that was collected on HD 4 was tested for SARS-CoV-2 which was then detected. Seven additional patients on the unit at the same time as the index patient were tested and all 7 were negative. The exposure window was defined as the patient's entire 12-day hospitalization. A total of 191 HCP were identified as potentially exposed and were evaluated by the occupational health department. HCP were tested twice within 14 days of the last exposure. Of the 191 HCP, 9 ultimately tested positive for SARS-CoV-2. The timing of the patient's clinical course in relation to the work schedule, symptom development, and testing of HCP cases was mapped in the outbreak investigation ( Figure) . Whole genomic sequencing was performed on available samples from the index patient and 7 of the 9 HCP. A total of 12 SNPs were found and shared among all 8 samples. This showed 100% similarity amongst SARS-CoV-2 strains of the index patient and 7 HCP cases. In comparison, the frequency of these SNPs ranged from 0.005% to 87.5% when searched across 96,303 SARS-CoV-2 sequences from the GESS whole genome sequencing database for the U.S. as of May 7 th , 2021 ( Table 1 ). The strain identified in this cluster of cases was identified as belonging to Nextstrain clade 20A and traced back to a likely importation from Europe in March 2020. In total, 7/9 (77.8%) of cases and 93/182 (51.1%) of controls responded to the survey. There were no significant differences in demographic characteristics or PPE practices between cases and controls. (Table 2 ) A total of 5/7 (71.4%) of cases performed any AGP compared to only 19/93 (20.4%) of controls (OR = 9.44; 95%CI = 1.42-106.44) ( Table 3) . Several aerosolizing and non-aerosolizing procedures were found to be significantly higher in cases than controls. Given that several of these activities are common procedures for certain work categories like nurses and nursing assistants, a secondary analysis was performed restricted to only those work categories which included at least one case (e.g. nurse, nursing assistant, respiratory/speech therapist, and physical therapist). Activities that were significantly higher in cases in both the primary and restricted analyses included: placing the patient on HFNC/CPAP/BiPAP, performing suctioning, being present during suctioning, and adjusting oxygen mask/tubing. All of the cases spent at least 2 days with the index patient compared to only 37/93 (39.8%) of controls (Table 4 ). Similarly, 6/7 (85.7%) of cases had at least 3 patient contacts per day compared to only 31/93 (33.3%) of controls. In ordinal regression analysis, cases spent significantly more time with the index patient across all time measurements compared to controls. In a secondary analysis restricted to only individuals in work categories including at least one case, the total number of days with the index patient, longest single contact with the patient, and time with the patient during an AGP remained significantly higher amongst cases. After exposure to an index patient with a delayed COVID-19 diagnosis, 9 of 191 exposed HCP tested positive for SARS-CoV-2. The index patient tested negative on admission and standard infection control measures were taken throughout her hospital stay. Compared to controls, positive HCP cases were more likely to report more direct patient contact time, performing AGPs (particularly placing the patient on HFNC/CPAP/BiPAP), and being involved in non-AGPs like adjusting oxygen, performing or being present during suctioning. Our study reinforces the risk of SARS-CoV-2 transmission to HCP due to delayed diagnosis of COVID-19 cases. Delayed or missed diagnosis of COVID-19, often due to alternative diagnoses or limited testing capabilities or initial negative tests, has been linked to nosocomial transmission and hospital outbreaks among HCP in multiple settings. In the Solano County outbreak in February 2020, the lack of readily available SARS-CoV-2 testing and the low clinical suspicion of community transmission led to the delayed diagnosis of the index patient and 3 PCR-confirmed HCP cases out of 121 exposed (22) . In another case in Boston, 44 HCP were exposed (with 2 attributable HCP cases) to a patient who presented with abdominal pain and shortness of breath who was initially diagnosed with cholecystitis and only tested and diagnosed with COVID-19 on HD 13 after developing acute respiratory failure; the patient was retrospectively thought to have undiagnosed infection on admission (11) . In the aforementioned cases, HCP had unprotected exposures to the index case yet resulted in fewer cases compared to our outbreak in which all HCP were wearing masks at all times and frequently also wearing eye protection. One key difference may be the different stages of infection in the index patients. Our index patient had a negative PCR on admission but developed new progressive respiratory symptoms between HD 3-6. This timing suggests that our index patient was infected in the community, was in the incubation period of her infection on HD 1 and likely became most infectious around HD 3-6, when the SARS-CoV-2 viral load in upper respiratory tract is thought to be highest (23) . This is also consistent with the timing of patient contact with cases: all 7 cases with direct patient contact took care of the patient on at least one of the days between HD 3 to 6, with 6 cases taking care of the patient on HD 4. A significant challenge in evaluating nosocomial transmission risk is the small number of cases that result from an exposure. Having survey responses from only 7 cases does limit the power of our study and our ability to perform multivariable analyses. However, compared to other outbreaks as described above, we did have a relatively large number of both cases and exposed controls which allowed for both a univariate analysis and a restricted sub-analysis to evaluate risk factors. Another critical concern in any outbreak investigation is determining the relationship between the index and secondary cases. The timeline of HCP cases in relation to the patient's testing and clinical course is consistent with the patient being the index case. Furthermore, a major strength of this study is the corroborating molecular evidence that these cases were all related. Whole genome sequencing is being increasingly used as an important epidemiologic method to investigate SARS-CoV-2 outbreaks including healthcare facility transmission events (12, 13) and provides very strong evidence that this cluster of cases is due to a single outbreak. However, we could not rule out the possibility of HCP to HCP transmission after exposure to the index patient. There is less demonstrated risk associated with other AGPs administering nebulizer treatments or manipulating non-invasive ventilation or high flow nasal canula (27, 28) . In our study, placing the patient on HFNC/CPAP/BiPAP was significantly associated with risk of transmission. The other AGPs evaluated in our study (e.g performing endoscopy and giving nebulizer treatments) were not significantly associated with transmission, however the power of this analysis was severely limited as very few people performed these procedures in either group. Curiously, performing or being present during suctioning and adjusting oxygen were also activities statistically significantly associated with cases in both the full and restricted analyses. However, there are several possible explanations for this. First, suctioning is an activity that occurs quite frequently and may be a proxy for increased time spent with the patient. Additionally, there is significant overlap among patient care responsibilities that involve the airway, which is particularly true for the small number of cases. Amongst cases, 3 of the 4 (75%) individuals who were present during suctioning also performed suctioning and placed the patient on BiPAP/CPAP/HFNC. Given the limited sample size and substantial overlap of activities and time, it is unclear whether these other airway procedures truly pose an increased risk above and beyond the risk attributed to AGPs. Nevertheless, it is possible that there is an increased risk associated with these other airway procedures that place HCP in close contact with airways secretions for extended periods of time. This would be most likely in a critically ill, nonintubated patient in which suctioning occurs in an open system (unlike in intubated patients where suctioning occurs in a closed system). Therefore, it is possible that in non-intubated patients, suctioning or other airway procedures that are not classified as strictly AGPs may in fact be an important, under-appreciated mode of SARS-CoV-2 transmission. In response to this outbreak, we implemented some major changes to minimize risk of transmission. Prior to this outbreak, the universal hospital protocol included universal procedure masks and eye protection in all patient encounters. Prior to CDC changing its recommendations (14) , a new protocol was implemented in our hospital requiring N95 masks for all contact with a patient undergoing intermittent or continuous AGPs even with negative SARS-CoV-2 testing on admission. This case also reinforced the importance of repeated SARS-CoV-2 re-testing after admission in suspicious cases. Our hospital laboratory also updated test codes so that it was not possible to order the respiratory virus panel without also testing for SARS-CoV-2. Our study underlines the risk of HCP transmission due to unidentified COVID-19 cases. It also highlights that patients who are admitted during the incubation period with a negative PCR test on admission may in fact be particularly high risk because they may become most infectious during their hospitalization. Finally, our study demonstrates that prolonged contact and placement or manipulation of HFNC/CPAP/BiPAP are a key risk factors for transmission. Healthcare Personnel (Cases) Cases included all potentially exposed HCP who tested positive for SARS-CoV-2 by RT-PCR. Controls included potentially exposed HCP who tested negative for SARS-CoV-2 twice within 14 days after their last exposure. *Categorical variables were collapsed into binary variables for univariate analysis using Fisher's exact test. Odds ratio compares the indicated group to all other groups in the category. None of the univariate analyses above were significant. There was significant overlap in many activities. Notably, all the individuals who performed suctioning also placed the patient on HFNC/CPAP/BiPAP. All individuals who performed a particular procedure were also included as present during that procedure. One case reported being present during suctioning but did not perform suction and reported no AGP exposures. COVID-19 in medical personnel: observation from Thailand Relative contributions of transmission routes for COVID-19 among healthcare personnel providing patient care Centers for Disease Control and Prevention Shortage of personal protective equipment endangering health workers worldwide Occupation and risk of severe COVID-19: prospective cohort study of 120 075 UK Biobank participants Risk of COVID-19 among front-line health-care workers and the general community: a prospective cohort study. Lancet Public Health Infections Among Healthcare Personnel in the United States Veterans Health Administration Transmission risk of SARS-CoV-2 to healthcare workers -observational results of a primary care hospital contact tracing Cluster in an Acute Care Hospital COVID-19-Associated Hospitalizations Among Health Care Personnel -COVID-NET, 13 States COVID-19 infections among HCWs exposed to a patient with a delayed diagnosis of COVID-19 Whole-genome Sequencing to Track Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Transmission in Nosocomial Outbreaks Serial Testing for SARS-CoV-2 and Virus Whole Genome Sequencing Inform Infection Risk at Two Skilled Nursing Facilities with COVID-19 Outbreaks -Minnesota MMWR Morb Mortal Wkly Rep Centers for Disease Control and Prevention Characteristics and transmission dynamics of COVID-19 in healthcare workers at a London teaching hospital COVID-19-Associated Hospitalizations Among Health Care Personnel -COVID-NET, 13 States Research Electronic Data Capture (REDCap) -A metadata-driven methodology and workflow process for providing translational research informatics support The REDCap consortium: Building an international community of software platform partners Next-Generation Molecular Testing of Newborn Dried Blood Spots for Cystic Fibrosis Transplant Virus Detection Using Multiplex Targeted Sequencing Nextstrain: realtime tracking of pathogen evolution Transmission of COVID-19 to Health Care Personnel During Exposures to a Hospitalized Patient SARS-CoV-2, SARS-CoV, and MERS-CoV viral load dynamics, duration of viral shedding, and infectiousness: a systematic review and meta-analysis A COVID-19 infection risk model for frontline health care workers. Netw Model Anal Health Inform Bioinforma What Is an Aerosol-Generating Procedure? JAMA Surg Aerosol-generating procedures and infective risk to healthcare workers from SARS-CoV-2: the limits of the evidence Evaluation, and Treatment of Coronavirus Aerosol Generating Procedures and Risk of Transmission of Acute Respiratory Infections to Healthcare Workers: A Systematic Review None.