Tracking Risk Factors Related to an Outbreak of COVID-19 Among Healthcare Workers in a General Medicine Ward

Background An outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection occurred in a medical ward involving patients and hospital staff from May to June 2020. Aim The aim of this study is to determine risk factors related to the outbreak of SARS-CoV-2 in six healthcare workers (HCWs) in a medical ward with initially unrecognized coronavirus disease 2019 (COVID-19) positive patients. Methods A retrospective cross-sectional study was conducted using a comprehensive questionnaire and personal interviews to determine the risk factors for COVID-19 infection in HCWs. Findings A total of 6/34 HCWs were diagnosed with COVID-19 in a medical ward. There were no differences between COVID-19 negative HCWs and COVID-19 positive HCWs in terms of mean duration of hours worked in the unit during the cluster event (180.2 vs 177.5 hours) (p>0.05), mean total time spent in contact with COVID-19 positive patients (12.8 vs 10.5 hours) (p>0.05), mean total time spent on aerosol-generating procedures (1.9 vs 0.9 hours) (p>0.05), and mean total time spent on non-aerosol generating procedures (10.9 vs 9.6 hours ) (p>0.05). There was no difference in exposure to COVID-19 positive family members among the HCWs (33% vs 3.7%, p=0.08). In contrast, exposure to COVID-19 positive contacts in the community was significantly greater in infected vs non-infected HCWs (16.7% vs 0%, p=0.03). Conclusion There was no significant difference in risk factors for contracting SARs-CoV2 among HCWs due to hospital exposures. COVID-19 positive HCWs were more likely to be exposed to positive individuals in their households and community, indicating that the source of SARS-CoV-2 infection came from outside the hospital.


Introduction
The first human case of severe acute respiratory syndrome coronavirus 2 (SARS-COV-2) was reported in December 2019 in Wuhan, China, and it was subsequently declared as a global pandemic by the World Health Organization (WHO) in March 2020.Since then, SARS-COV-2 has been contracted by more than 100 million population worldwide and has caused more than 2 million deaths [1].High infection rates among healthcare workers (HCWs) were initially reported in China, and concern about HCWs became a worldwide issue [2].In March 2020, WHO released guidelines on infection prevention and control in healthcare facilities [3].The first national guidelines on the clinical management and treatment of coronavirus disease 2019 (COVID- 19) were released in the UAE in March 2020, and updated in April 2020, when the total cumulative cases in the

Study timeline
To analyze the progression and spread of the COVID-19 infection outbreak within this unit, a timeline of events was mapped (Figure 2).A total of six HCWs (HCW-A, HCW-B, HCW-C, HCW-D, HCW-E, HCW-F) tested positive in the unit, all preceding the patients testing COVID-19 positive on polymerase chain reaction (PCR).However, all patients were clinically symptomatic at the time of admission, prior to the HCWs testing COVID-19 positive on PCR.HCW-A, HCW-B, and HCW-C were nurses in the ward, and HCW-D, HCW-E, and HCW-F were nursing assistants.A total of three patients (patient A, patient B, and patient C) eventually tested positive on the ward.Questions to be answered by this investigation included the following: Were the patients the source of this outbreak within the ward?Did the HCWs acquire the infection outside of the hospital or from other workers in the ward?

Case descriptions
Patient A is an 87-year-old male with known ischemic heart disease, diabetes, and past cerebrovascular accident, who was admitted to the ward on April 25 with cough, lethargy, reduced fluid intake, and urinary tract infection.His chest X-ray (CXR) on admission showed right upper lobe infiltrations, and right lower lobe bronchiectasis and infiltrates; subsequently, he was diagnosed with right-sided pneumonia.His SARS-CoV-2 PCR was negative on April 25, and his second and third SARS-CoV-2 PCR tests were also negative on April 30 and May 11, respectively; however, he was detected to have COVID-19 on May 30 with a positive PCR test.
Patient B was admitted to the ward with methicillin-resistant Staphylococcus aureus sepsis and had symptoms of fever and weakness with a high C-reactive protein.His CXR revealed right upper lobe reticular nodular opacifications with no other findings, and the SARS-CoV-2 PCR was negative on May 11 and May 22. On May 24, his CXR revealed bilateral mild and lower zone infiltrations with right pleural effusion, and his SARS-CoV-2 PCR was positive.Due to his unstable condition, he had to be intubated and put on a ventilator.
Patient C is an 85-year-old male with a previous history of hypertension, cerebrovascular accident, and dementia, who was admitted on May 13 with cough, tachypnea, lethargy, and an oxygen saturation of 83%.His management included piperacillin-tazobactam, azithromycin, and frequent nebulization.His CXR on admission was clear, and a repeat CXR on May 19 showed left lower lobe infiltrates.SARS-CoV-2 PCR on admission was not detected, and he remained negative on repeat test three days later; he was found to be SARS-CoV-2 PCR positive on May 23.His symptoms deteriorated prompting the requirement of intubation on May 22.He was transferred to a different hospital four days after his intubation.

Patient's timeline of events
Patient A was admitted to the ward on day 1 and was tested for COVID-19 a total of three times.His tests were negative during his admission to the hospital for two weeks, where he underwent a total of three COVID-19 PCR tests (days 1, 6, and 17).During the same period, he underwent four CXRs, which showed signs of progressing lung infection that transformed into pneumonia on day 7 and remained unchanged until day 24.The patient was found to be positive on day 36, that is, 11 days after the first HCW (HCW-B) was detected to be positive (day 25).He was then transferred to another facility for further care.
Patient B was admitted on day 17 and was tested three times.His first two COVID-19 PCR tests were negative on days 17 and 29.His CXR meanwhile showed signs of lung infection (opacifications and pleural effusions).His first positive PCR was on day 30, that is, five days after the first positive PCR of our HCW (HCW-B day 25

Healthcare workers' timeline of events
All HCWs underwent routine COVID-19 screening and were additionally tested if they had any symptoms.They all filled out a questionnaire to identify in-hospital and community exposure (see the Appendix for the questionnaires).In this case, all the nurses tested positive after experiencing symptoms.HCW-B was the first HCW to have a positive PCR result on day 25.HCW-F then tested positive the following day (day 26).HCW-E and HCW-A tested positive one day apart (days 31 and 32, respectively).Finally, HCW-D and HCW-C tested positive on days 35 and 37.All the HCWs and patients tested positive in the span of 10 days.

Nurses' symptomatology
All six HCWs (HW-A, HW-B, HW-C, HW-D, HW-E, and HW-F) tested positive for SARS-CoV-2 on PCR.In all cases, symptom onset preceded PCR positivity by 1 to 10 days.
HW-B and HW-D were the first to develop symptoms among the cohort.HW-A (nurse N) developed symptoms on May 22, which lasted for seven days.Her symptoms consisted of myalgia, sore throat, headache, and shortness of breath.
HW-B (nurse O) developed symptoms on May 17, such as fever, chills, myalgia, anosmia, headache with nausea, vomiting, and diarrhea.Her symptoms persisted for seven days.
HW-C (nurse P) reported symptoms on May 29, which lasted for less than a week.Her symptoms were fever, myalgia, rhinorrhea, sore throat, headache, and cough.
HW-D (nursing assistant G) showed symptoms of rigors and myalgia on May 17.She then developed anosmia eight days later.
HW-E (nursing assistant H) developed symptoms of rhinorrhea and wheezing on May 21, which lasted for six days.

Exposure duration
HW-B, HW-D, and HW-E were exposed to all three patients in the unit.HW-A was only exposed to patient A and patient B. HW-F was only exposed to patient B and patient C. HW-C was only exposed to patient B.

Outcome measures
In this study, we looked at the difference in exposure duration, personal protective equipment (PPE) use, and community exposures between the COVID-19 positive HCWs (cases) and COVID-19 negative HCWs (controls).We measured the total time spent in the COVID-19 ward and the total time spent in contact with the COVID-19 positive patients.The time spent with the COVID-19 positive patients was further categorized as total time spent conducting AGPs and total time spent conducting non-aerosol generating procedures (NAGPs).
The PPE use between the cases and controls when in contact with each patient was measured by comparing the use of masks (single surgical mask, multiple surgical masks, and N95 mask), aprons, gowns, gloves, face shields, shoe covers, and hand washing before and after patient contact.
Community risk of acquiring COVID-19 was compared between cases and controls by retrospectively looking at the presence or absence of risk factors during the time of the cluster event, such as traveling, large group gatherings, visiting hotels, restaurants, or swimming pools.Community risk was also measured by comparing the housing conditions between the HCWs by assessing the number of adults and children they lived with.Direct community exposure was defined as contact with a COVID-19 positive patient in the community or at home.
All these variables were compiled into a single questionnaire, which was distributed to the nurses and answered retrospectively based on their memory of how much time they spent on each of the variables.

Analysis
Categorical data were presented as frequencies and proportions, and continuous data were presented as means and standard deviations.Baseline characteristics of HCWs who were infected and those not infected were compared using the Mann-Whitney test for continuous data.Meanwhile, exact Fischer's test was used for categorical variables.Average ratios and prevalence ratios between infected and non-infected, as well as their corresponding confidence intervals, were calculated for all the variables in the study when applicable.A p-value of less than 0.05 was considered statistically significant in all the tests.Analysis was performed using Statistical Package for the Social Sciences (SPSS) for Windows Version 25 (IBM Corp., Armonk, NY).

Demographics
A total of 34 HCWs were screened in the unit while identifying the cluster event: 16 were nurses, nine were physician hospitalists, and nine were nursing assistants.All HCWs had contact with at least one of the three COVID-19 positive patients (patient A, patient B, and patient C).Of the HCWs, 82% were female, and 74% were under the age of 40 years.A total of six out of 34 HCWs were positive for SARS-CoV-2 on PCR: 3/16 nurses, 0/9 doctors, and 3/9 nursing assistants.There was no statistically significant difference in terms of risk of acquiring COVID-19 between HCWs who were >40 years and <40 years of age (p>0.05) or in terms of gender of the HCWs (p>0.05).

Exposures
The measured exposures included the difference in exposure duration, PPE use, and community exposures between the COVID-19 positive HCWs (cases) and COVID-19 negative HCWs (controls).
There was no difference between COVID-  There was also no difference in specific AGPs and NAGPs between cases and controls when in contact with patient A (Table 2).There was also no difference in PPE use between cases and controls when in contact with patient A (

PPE, personal protective equipment
The community exposures between cases and controls are tabulated in Table 8 and Table 9.Interestingly, the only significant exposure that the COVID-19 positive HCWs had that was greater than the COVID-19 negative HCWs was in the community exposure, with the prevalence of contact with a COVID-19 positive individual in the community being 16.7% in the cases versus 0% in the controls (p=0.03).Moreover, there was a tendency toward a greater prevalence of the cases having a COVID-19 positive spouse compared to controls (33% vs 3.7%; p=0.08).There was no difference between cases and controls in the prevalence of other community exposures such as staying at a hotel (0% vs 7.4%; p=0.50), going out to restaurants (0% vs 7.4%; p=0.50), or going out swimming (0% vs 3.4%; p=0.64).None of the HCWs traveled or had a gathering of more than 10 people, and all the HCWs went out shopping; therefore, these risk factors could not be compared.There was also no statistically significant difference between cases and controls in their housing conditions, with the mean number being similar in terms of adults they lived with (3.83

Discussion
In this cohort, 34 HCWs had contact with at least one COVID-19 positive patient; six of these 34 HCWs tested positive for SARS-CoV-2 on a nasal swab PCR.There was no statistical difference in contracting COVID-19 and the age of the HCWs, nor in their gender.There was no statistically significant difference in the exposure duration or PPE use and contracting COVID-19 among any of the three patients.Exposure assessment included time spent in the COVID-19 ward and total time spent in contact with the COVID-19 positive patients, with the latter including total time spent conducting AGP and total time spent conducting NAGP.There was no statistical significance in exposure and contracting COVID-19 from any of the patients.
However, exposure to a COVID-19 positive individual in the community was associated with a statistically significant higher prevalence of contracting COVID-19 in HCWs (p=0.03).There was a greater prevalence of COVID-19 positive spouses among the cases in comparison to the controls (33% vs 3.7%; p=0.08).Other community exposures, including staying at a hotel, going out to restaurants, or going out swimming, were not different between cases and controls.There was also no statistically significant difference between cases and controls in their housing conditions.
During the first phase of the pandemic, healthcare systems were overburdened by the rising COVID-19 cases.
HCWs struggled with increased psychological stress, fatigue, and long working hours.Additionally, there was a global shortage of PPE available for HCWs [6], and they were exposed to suspected COVID-19 infected patients with minimal to no PPE.Therefore, during the initial stages of the pandemic, HCWs were at an increased risk of acquiring COVID-19 [7,8].The present study was conducted in HCWs without any PPE shortage.
Lessons from the first phase of the pandemic taught HCWs the precautionary steps needed to prevent an outbreak, such as screening for SARS-CoV-2, isolating and quarantining COVID-19 positive patients, screening of HCWs who work with COVID-19 positive patients, maintaining social distancing [9][10][11], using PPE in wards housing COVID-19 positive patients, and continuous use of mask in the hospital [12,13].The HCWs in the present study were equipped with all these lessons learnt from the initial stages of the pandemic.
The literature has shown that the risk of HCWs acquiring COVID-19 in the intensive care setting is low [14,15], possibly owing to the extensive PPE use and awareness of the patients' COVID-19 infection status.However, there are growing concerns that the risk to HCWs may be arising from contact with patients at the early stages of the infection who are pre-symptomatic or asymptomatic [11].
Several studies in the literature that showed an increased risk of HCWs contracting the COVID-19 infection were conducted during the first wave of the pandemic, before the establishment of universal masking protocols, or in settings where access to face masks was limited [6,16,17].In these studies, the in-hospital and community risk factors were not measured.A study by Wilkins et.al. on 6,510 HCWs in a well-resourced healthcare system showed that the risk of HCWs testing positive for COVID-19 was lower than previous estimates when community exposure was accounted for [18].Another study that investigated the potential source of infection in HCW observed a trend where those with a larger household size tended to have more frequently detectable antibodies against SARS-CoV-2 [19].A meta-analysis conducted by Gómez-Ochoa et al. evaluated the potential sources of COVID-19 infections among healthcare workers, assessing both nosocomial and community-acquired risk factors [20].There seem to be no significant differences in infection rates of HCW groups working in high, intermediate, and low-exposure risk settings [21,22].Additionally, the proportion of HCWs who tested positive for COVID-19 was the same in HCWs with patientfacing roles and non-patient-facing roles.The meta-analysis also revealed that community transmission has a relevant role in HCW infection rates [19,20,22,23].The findings of our study are consistent with this metaanalysis.We encountered the consistent condition of increased community exposure being a significant risk factor for acquiring the infection.Our literature search revealed a substantial need for data on the risk factor disparities in HCW [24].
Finally, most of the published literature does not identify the specific workplace risk factors such as time spent working with COVID-19 positive patients.Our study used a novel systematic survey of potential hospital and community risk factors that was applied to both infected and non-infected HCWs.We believe that this survey could be of use to other hospitals investigating similar incidences of COVID-19 infections identified in HCWs (Appendix).

Strength and limitations
One of the strengths of this study was the extensiveness of the survey, which attempted to cover the majority of patient interaction variables.Also, there was a high completion rate of the survey, which limited any gap in information.A limitation is the small sample size of the cluster event, where only six of the HCWs could be included as cases; this reduces the power of the study, but it is still valuable in observing and analyzing the trends in this event.Since the survey was administered retrospectively to all HCWs involved, there is a possibility of recall bias in accurately documenting the information in the survey, but the administration was done soon after the cluster event, reducing the risk of bias.
It is possible that other important risk factors were not surveyed in the present study.Socialization between HCWs, such as for face-to-face meetings and meals, is a possible mechanism of transmission between HCWs [25]; however, this was not measured in the present study.This may have been a small factor because hospital policy prohibited these types of interactions.A previous study showed that socializing between HCWs and the lack of PPE was strongly correlated with the risk of HCWs acquiring COVID-19 from patients but conducting AGPs was not [26].
Our study can further ease the HCW's fear of acquiring COVID-19, from false COVID-19 negative patients admitted to the hospital, due to the lack of significant difference in risk from patient contact between cases and controls.Continued vigilance and monitoring of hospital infection control practices related to COVID-19 remain increasingly important with the continued spread of potentially more transmissible new variants [27].

Conclusions
In conclusion, our paper reports a cluster event of COVID-19 infection in a hospital ward associated with initially unrecognized COVID-19 positive patients.No significant difference was found in risk factors for contracting SARS-CoV-2 among HCWs due to hospital exposures.COVID-19 positive HCWs were more likely to be exposed to positive individuals in their households and community, indicating that the source of SARS-CoV-2 infection likely came from outside the hospital.However, these data do not rule out a yet-tobe-identified mechanism of transmission independent of the duration and type of exposures measured in this study.

Appendices
Questionnaires

2 FIGURE 1 :
FIGURE 1: Contact tracing between COVID-19 positive patients and healthcare workers during a cluster event in the medical ward

FIGURE 2 :
FIGURE 2: Healthcare workers and patients' timeline of events PCR, polymerase chain reaction; CXR, chest X-ray 19 positive HCWs (cases) and COVID-19 negative HCWs (controls) in terms of mean duration of hours worked that month in the unit during the cluster event (177.50 [SD: 28.42] hours vs 176.89 [SD: 30.73] hours; average ratio (AR): 1.00; p=0.56) or mean total time spent in contact with the 3 patients (10.53 [SD: 9.74] hours vs 9.15 [SD: 9.00] hours; AR: 1.15; p=0.54).There was no difference in mean total time spent on AGPs for three patients (0.95 [SD: 0.94] hours vs 1.45 [SD: 1.39] hours; AR: 0.65; p=0.51) and mean total time spent on NAGPs for three patients (9.58 [SD: 10.06] hours vs 7.70 [SD: 8.87] hours; AR: 1.24; p=0.46).When comparing the different types of AGPs and NAGPs, there was no difference between the cases and controls (Table1).There was also no difference in PPE use between cases and controls when in contact with any of the patients.None of the HCWs wore multiple surgical masks, and all HCWs washed their hands before and after patient contact.

TABLE 1 : Total time spent on AGP and NAGP by healthcare workers around COVID-19 positive patients.
AGP, aerosol-generating procedures; NAGP, non-aerosol-generating procedures; SD, standard deviation Individual

TABLE 3 : PPE use by healthcare workers when in contact with patient A.
PPE, personal protective equipmentThere was no difference in mean total time spent in contact with patient B (2.24 [SD: 1.49] hours vs 1.68 [SD: 2.17] hours; AR: 1.33; p=0.154), mean total time spent on AGPs for patient B (0.36 [SD: 0.31] hours vs 0.33 [SD: 0.27] hours; AR: 1.11; p=0.52), and mean total time spent on NAGPs for patient B (1.88 [SD: 1.61] hours vs 1.36 [SD: 1.86] hours; AR: 1.38; p=0.11).There was also no difference in specific AGPs and NAGPs between cases and controls when in contact with patient B (Table4).There was also no difference in PPE use between cases and controls when in contact with patient B (Table5).Cases,

TABLE 5 : PPE use by healthcare workers when in contact with patient B
PPE, personal protective equipmentThere was no difference in mean total time spent in contact with patient C (6.36 [SD: 7.64] hours vs 5.76 [SD: 6.95] hours; AR: 1.10; p=0.981), mean total time spent on AGPs for patient C (0.28 [SD: 0.60] hours vs 0.60 [SD: 0.92] hours; AR: 0.46; p=0.50), and mean total time spent on NAGPs for patient C (6.08 [SD: 7.73] hours vs 5.16 [SD: 6.63] hours; AR: 1.18; p=0.90).There was also no difference in specific AGPs and NAGPs between cases and controls when in contact with patient C (Table6).There was also no difference in PPE use between cases and controls when in contact with patient C (Table7).Cases ,n Cases, mean (SD) Controls, n Controls, mean (SD) Ratio of cases:controls p-Value

you experienced any of the following symptoms in the last four weeks?
Symptoms Circle if Yes or NoIf yes, please define symptom onset (✔):

use while interacting with COVID-19 positive patient PPE
use Did you use the following PPE during interaction with the patient?

TABLE 13 : Personal protective equipment use questionnaire
PPE, personal protective equipmentCommunity exposure in the last four weeks

TABLE 14 : Community exposure to COVID-19 questionnaire for healthcare workers Additional Information Disclosures Human
subjects: Consent was obtained or waived by all participants in this study.Mediclinic Middle East