The potential for nosocomial spread of severe acute respiratory coronavirus virus 2 (SARS-CoV-2) is a primary concern of public health experts, hospital epidemiologists, clinicians, healthcare institutions and patients, particularly because SARS-CoV in 2003 was associated with substantial nosocomial spread ^{Reference Yu, Xie and Tsoi1} and SARS CoV-2 has a considerably high reproductive number. ^{Reference Del Rio and Malani2–Reference Temime, Gustin and Duval4} The reasons for efficient person-to-person transmission are multifactorial, including high-level viral shedding in the upper respiratory tract and documented presymptomatic, asymptomatic, and paucisymptomatic spread. ^{Reference Arons, Hatfield and Reddy5–Reference Wolfel, Corman and Guggemos9} In this study, we describe the infection control measures implemented and the relationship with SARS-CoV-2 test results in hospitalized patients.

The University of Wisconsin Health System (UW Health) includes 3 hospitals, with 672 beds and >120 clinics; it serves >600,000 patients in the Upper Midwest. The infection control program includes a special pathogens prevention multidisciplinary program that led the coronavirus disease 2019 (COVID-19) preparedness and response, including measures to prevent nosocomial transmission of SARS-CoV-2. The infection control measures instituted, time of implementation, and description of each intervention are listed in Table 1. Each intervention fell into 1 of 6 general categories: (1) personal protective equipment guidance and training, (2) testing guidance and algorithms, (3) monitoring of patients, visitors, and staff for signs and symptoms, (4) improving communication and patient care processes for patients with suspected or proven COVID-19, (5) implementation of electronic medical record decision support aids, and (6) control of physical environment with cohorting of suspected patients or patients and maintaining physical distancing.

Table 1. Implementation, Timing and Description of Infectious Control Measures Instituted in Response to the COVID-19 Pandemic

Note. AII, airborne infection isolation; APP, advanced practice provider; CTL, care team leader; ED, emergency department; EHS, employee health services; EMR, electronic medical record; ICU, intensive care unit; IMC, intermediate care; MD, medical doctor; NP, nasopharyngeal; OR, operating room; PPE, personal protective equipment; RT-PCR, reverse-transcriptase polymerase chain reaction; PUI, person under investigation; RN, registered nurse.

As a measure of the success of these interventions, we examined the positivity rate for SARS-CoV-2 RT-PCR testing of inpatients from March 13, 2020, to June 25, 2020. All testing was performed using nasopharyngeal swabs with emergency-use authorization (EUA)-approved RT-PCR testing methods. Patients who were tested as outpatients, those tested in the emergency room or urgent care clinics, and those tested within the first 24 hours of an admission were excluded. Notably, repeated inpatient testing of individuals was, in general, directed toward those undergoing procedures, those in whom signs or symptoms suggested possible COVID-19, those with acute changes in status requiring intensive care unit (ICU) or intermediate (IMC) care, and/or based on provider judgment.

In total, 720 patients were tested >24 hours after admission to an inpatient unit, and the total number of inpatient SARS-CoV-2 tests was 1,007. The median age was 59 years (IQR, 40–69) and 52% were male. The reason for testing was skewed toward asymptomatic screening preceding procedures (71%). This finding was expected because repeat preprocedural testing was directed to be done within 48 hours prior to any aerosol-generating procedure. Of 1,007 inpatient tests, 59 tests (5.9%) were positive and 58 were known to be positive prior to inpatient testing (eg, positive prior to admission or as part of admission work-up). Thus, only 1 patient (0.1%) tested positive during an inpatient stay in which that patient was not known to have a history of a positive test. Over the study period, we had a sizeable COVID-19 inpatient population (112 inpatients with 1160 inpatient days) and a large at-risk pool of inpatients without COVID-19 (37,096 inpatient days).

For the single positive inpatient without a prior history of SARS-CoV-2, chart review revealed that this adult patient lived in a community setting, had mild symptoms (sinus congestion, eye pain, and cough) that started 10 days prior to admission, and was self-isolating at home. The patient presented with a myocardial infarction before universal admission testing was instituted, and the prior mild respiratory symptoms were not noted. On hospital day 4, the patient tested positive as part of pre-procedure screening. We believe that infection was present from community exposure prior to admission; therefore, we did not find any laboratory-confirmed cases suggestive of possible nosocomially acquired SARS-CoV-2 infection despite a substantial inpatient population with and without COVID-19. It has been suggested that false-negative results may occur, but negative-to-positive conversion has rarely occurred at our institution (<1%).¹⁰ Importantly, we were able to achieve these results without routine, serial testing of asymptomatic healthcare workers (HCWs), and we had a low threshold for testing HCWs with symptoms with a 1% rate of infection in our HCWs.

Our study has several limitations. First, this was a retrospective observational study. Second, because testing was limited to inpatient setting, we were not able to ascertain symptom onset after discharge, which may have resulted in testing elsewhere. However, we examined all positive ambulatory tests and did not find any positive results in patients within 7 days of discharge from our hospital. Finally, we were unable to examine the relative effect of each individual infection control measure.

Our study has a number of strengths. As the single positive case we found demonstrates, it can be difficult to identify all potential positive patients by history taking alone. Thus, we strongly believe that universal testing of patients admitted to the hospital should be performed. This testing should be followed by targeted testing based on daily, protocol-driven screening questions to determine whether any symptoms have changed that suggest possible COVID-19. These first 2 measures aim to rapidly identify patients that should be placed in transmission-based isolation and to help prevent inadvertent spread. However, additional measures are obviously necessary to prevent nosocomial spread from known SARS-CoV-2–positive patients who may need complex medical care including intensive care, multiple-specialty care, invasive procedures or surgery, and intrahospital transport. These measures include meticulous infection control measures described here. In conclusion, using iterative implementation of infection control measures we were able to care for numerous COVID-19–infected and –uninfected patients without any cases of nosocomial spread.