key: cord-0796554-0f2dsxry
authors: Lepak, Alexander J.; Buys, Ashley; Stevens, Linda; LeClair-Netzel, Megan; Anderson, Laura; Osman, Fauzia; Brennan, Meghan B.; Bartels, Christie M.; Safdar, Nasia
title: COVID-19 in Health Care Personnel: Significance of Health Care Role, Contact History, and Symptoms in Those Who Test Positive for SARS-CoV-2 Infection
date: 2021-07-02
journal: Mayo Clin Proc
DOI: 10.1016/j.mayocp.2021.06.019
sha: 860bdcc454a7c9278d74330b8cd0b3074cc8c34a
doc_id: 796554
cord_uid: 0f2dsxry

Objective To identify significant factors that help predict whether healthcare personnel (HCP) will test positive for SARS-CoV-2. Patients and Methods We conducted a prospective cohort study among 7015 symptomatic HCP from March 25, 2020, through November 11, 2020. We analyzed the associations between healthcare role, contact history, symptoms, and a positive nasopharyngeal swab SARS-CoV-2 PCR test results, using univariate and multivariable modelling. Results 624 (9%) HCP were positive over the study period. On multivariable analysis, having a healthcare role other than physician or advanced practice provider (APP), contact with family or community member with known or suspected COVID-19, and seven individual symptoms (cough, anosmia, ageusia, fever, myalgia, chills, and headache) were significantly associated with higher adjusted odds ratios for testing positive for SARS-CoV-2. For each increase in symptom number, the odds of testing positive nearly doubled (OR, 1.93; 95% CI, 1.82-2.07, p<0.001). Conclusions Symptomatic HCP have higher adjusted odds of testing positive for SARS-CoV-2 based on three distinct factors: (1) non-physician/APP role, (2) contact with a family or community member with suspected or known COVID-19, and (3) specific symptoms and symptom number. Differences among healthcare roles, which persisted after controlling for contacts, may reflect the influence of social determinants. Contacts with COVID-19 positive patients and/or HCP were not associated with higher odds of testing positive, supporting current infection control efforts. Targeted symptom and contact questionnaires may streamline symptomatic HCP testing for COVID-19.

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of coronavirus disease 2019 (COVID-19), continues to surge globally. Healthcare personnel (HCP) are vital to the response, necessitating a healthy workforce. Sources of transmission to HCP can include patients, healthcare visitors, other HCP, and close contacts such as family and community members. Within each contact type, the risk of transmission may vary based on use of personal protective equipment, hand hygiene, physical distancing, aerosol exposure, ventilation management, and frequency and duration of contact. SARS-CoV-2 acquisition among HCP may also be driven by social determinants, especially outside of the healthcare environment.

Monitoring for COVID-19 symptoms in HCP is important for early diagnosis and isolation, preventing spread in the healthcare system and among vulnerable populations. However, signs and symptoms can be subtle and mimic other common respiratory viral illnesses. 1 Therefore, healthcare systems have struggled to balance screening HCP for possible COVID-19, promoting illness absenteeism, and ensuring appropriate staffing for patient care, especially during surges. Data on prevalence of, and risk factors for, COVID-19 in HCP exist; 2-6 however, there is little information on the relationship between (1) healthcare role, (2) contact history, and (3) symptoms with subsequent SARS-CoV-2 PCR testing results for symptomatic HCP. [7] [8] [9] [10] [11] We performed a prospective cohort study of symptomatic HCP at a large, tertiary care center to determine these associations. 

We constructed a prospective cohort of symptomatic HCP working in a large, integrated health system serving more than 600,000 patients annually in the Upper Midwest. The system employed approximately 13,000 personnel at 4 hospitals and more than 80 outpatient sites. Data collection took place between March 25, 2020 and November 11, 2020.

Starting March 25, 2020, Employee Health Services (EHS) offered SARS-CoV-2 nasopharyngeal (NP) swab PCR testing for any symptomatic HCP at no cost to employees. All HCP were eligible and encouraged to use the testing site if they had symptoms compatible with COVID-19. EHS-trained staff prospectively collected HCP self-reported symptoms, sociodemographics, contact history, and duration of contact using a standardized telephone survey performed at the time of testing. All data collection was performed prior to test results to prevent recall bias. Queried symptoms included: cough, chills, shortness of breath, chest tightness, fever, anosmia (loss of smell), ageusia (loss of taste), pharyngitis (sore throat), rhinorrhea (runny nose), nasal congestion, headache, fatigue, myalgia (muscle pain), nausea and/or vomiting, and diarrhea. Ten questions focused on potential contacts with SARS-CoV-2-infected was defined as being within 6 feet of this person, who did not wear a surgical mask for any portion of the contact time. Additionally, contact with a non-masked healthcare visitor was ascertained. If there was reported contact, the estimated duration was recorded (<10 minutes, 10-20 minutes, >20 minutes, or indeterminate). Additional variables included: date of testing, test result, work location (inpatient, ambulatory setting, home, affiliated healthcare facility) and healthcare role. Roles were grouped as follows: (1) physician or advanced practice provider (APP); (2) nurse, medical assistant (MA), or therapist; (3) other HCP involved in direct patient care (e.g., radiology technician); (4) other HCP not involved in direct patient care (e.g., custodian); (5) medical/nursing trainee (e.g., medical student, resident, nursing student); (6) pharmacy.

Re-testing a symptomatic HCP was restricted to the following criteria: (1) those that had previous negative test, but the symptoms had resolved completely and they now had new symptoms, (2) those that had a previous negative test, their previous symptoms had not completely resolved but have developed new symptoms or significant worsening of previous symptoms, and (3) for situations outside of the previous two the case was discussed with employee health nurse or APP. HCP included in the dataset who had more than one test over the study period were considered independent observations for each lab test-data pair. HCP who tested positive were not included if they were tested again at a later period to avoid counting a single infection twice.

Throughout the study period, the healthcare system's infection control measures for confirmed COVID-19 included negative pressure room isolation and donning fit tested N-95 respirators, gowns, gloves, and face shields for room entries. 12 For all other HCP-patient interactions, HCP wore surgical masks made to ASTM standards and face shield. Surgical masks were re-used for up to 3 days and N-95 for 7 days. HCP received in-person PPE training and emails regarding best practices to reduce transmission. All hospital visitors were instructed to wear a mask, although some did not comply completely with this policy early in the pandemic. Universal masking of staff, patients, and visitors was instituted on April 24, 2020, and remained in effect throughout the remaining period. Hospital visitation became increasingly restricted over the study period; for a significant portion no visitors were allowed except under special circumstances such as end of life. This study was considered quality improvement and was exempt from IRB review. 

We used chi-squared and Fisher's exact tests to summarize demographic information, HCP contact history, and symptoms. We conducted univariate logistic regression to identify factors associated with a positive SARS-CoV-2 PCR test. Factors with p-values ≤0.05 in univariate analysis were entered into a multivariable model. We used a post-estimation odds plot with confidence intervals to graph the results of the multivariable model and robust standard error estimates. 13 We also constructed an unweighted symptom score (0-7) using the seven symptoms that remained positively correlated with a positive test result in our multivariable model. We created a second multivariable model to assess the association between testing positive for SARS-CoV-2 and the symptom score, controlling for factors that were statistically significant in the univariate model but excluding individual symptom variables to avoid collinearity. All 

A total of 7015 SARS-CoV-2 RT-PCR tests were performed on symptomatic HCP with 624 (9%) testing positive over the study period. Trends in the number of HCP tested and positivity rate paralleled regional surges, with very low prevalence and relatively lower testing volume early, a modest surge in June and July, and a substantial surge in September through November (Figure 1) .

Testing, demographic, contact, and symptom data stratified by HCP role are presented in Table 1 The majority of HCP worked in inpatient (n=3120, 44.5%) or ambulatory settings (n=3296, 47.0%). HCP not involved in direct patient care frequently worked from home.

Four thousand fourteen HCP (57.2%) reported 7080 contacts. Contact with an individual known to have COVID-19 was reported 2507 times, while contact with an individual suspected to have COVID-19 was reported 1654 times. Patient contacts were highest amongst those expected to have more patient interactions (e.g physician/APP and nurse/MA/therapist groups). Contact with another HCP with suspected or known COVID-19 was similar across healthcare roles. Physicians/APP had fewer contacts with family members with known COVID-19 compared to other healthcare roles. When contact time could be estimated, a majority (70%) recalled spending >20 minutes within 6 feet of the index person. The proportion reporting this contact duration was similar across healthcare roles. Finally, the most common symptoms, in descending order, included: pharyngitis, headache, and nasal congestion. All were reported in >50% of symptomatic HCP. Rhinorrhea, cough, fatigue, myalgia, chills, chest tightness, nausea/vomiting, diarrhea, shortness of breath, and fever were reported in 10-50%; J o u r n a l P r e -p r o o f whereas, loss of smell or loss of taste as a symptom were rare (<10%) (Figure 1 ).

Symptoms were reported with similar frequency across HCP roles.

On univariate analysis, all healthcare roles had significantly higher odds of testing positive compared to the referent physician/APP group ( Table 2) . Working from home and contact with a family or community member with known or suspected COVID-19 were associated with higher odds of testing positive for SARS-CoV-2. In the healthcare environment, contact with a patient who had a respiratory illness or nonmasked visitor was protective. Contact times longer than 20 minutes were associated with increased odds of positive test results. Nine symptoms (cough, chills, fever, anosmia, ageusia, nasal congestion, headache, fatigue, and myaligia) were associated with increased odds of testing positive for SARS-CoV-2, while two (pharyngitis and nausea/vomiting) were associated with decreased odds. The proportion of HCP who tested positive versus negative for SARS-CoV-2 for each symptom are shown in Figure   2 .

In multivariable analysis ( there was no significant difference in the odds of a positive test across work locations or contact duration, nor was there a significant difference based on contact with a patient who had a respiratory illness or contact with a non-masked visitor. However, contact with a family or community member with known or suspected COVID-19 remained highly associated with a positive SARS-CoV-2 test (OR, 4.03; 95% CI, 3.13-5.18; and OR, 1.65; 95% CI, 1.24-2.19; respectively). Seven symptoms remained statistically significant in the multivariable model. These included, in descending order, cough, anosmia, ageusia, fever, myalgia, chills, and headache. The two symptoms that were negatively correlated were pharyngitis and nausea/vomiting.

We created a second multivariable model using the composite symptom score based on the seven symptoms that were associated with an increased odds of testing positive in the main multivariable model (cough, anosmia, ageusia, fever, myalgia, chills, and headache). Among HCP with none of the seven symptoms, 3.1% tested positive. Over 40% of HCP with five or more of the seven symptoms tested positive ( Table 4 ). The odds of testing positive almost doubled (OR, 1.93; 95% CI, 1.82-2.07) for each additional symptom.

We report relationships between (1) healthcare role, (2) with family or community exposures, but not with exposures within the healthcare environment. There may be several explanations for this finding and it is important to acknowledge we could not definitively distinguish between SARS-CoV-2 infection acquired occupationally or from the community. A possibility we theorize, but could not test, was whether differences in sociodemographics and/or socioeconomic status (SES) played a role. For example, symptomatic physicians and APPs had significantly reduced odds of testing positive for SARS-CoV-2 compared to all other healthcare roles. The one exception to the trend was pharmacy. This may reflect the extremely high volume of interactions many pharmacy personnel have with persons that have an unknown infectious (i.e., COVID-19) history, as well as PPE, hand hygiene, and physical distancing challenges that are unique to their specific work environment. Pharmacists may also work at multiple locations with variation in community SARS-CoV-2 transmission. However, pharmacy was the smallest group with a wide confidence interval around our point estimate. Therefore, the true odds may be aligned with our hypothesized socioeconomic trend and poorly estimated in our model.

disadvantage. 14-21 The healthcare system can serve as a unique microcosm for understanding COVID-19 disparities. In our study population, all HCP received email reminders or best practices to remain safe in and out of the healthcare environment.

Testing for symptomatic individuals was accessible and free regardless of specific job roles. Therefore, better education or access to testing is unlikely to explain the lower odds of a positive test among physicians/APPs compared to other HCP groups. Instead, we suspect physicians/APPs may have increased resources to more effectively reduce J o u r n a l P r e -p r o o f their risk of SARS-CoV-2 infection. For example, higher earning individuals may be more likely to live in detached, single-dwelling homes where distancing is feasible. They may be able to afford in-home care or education options for their dependents, which may reduce exposure risk. Lower earners may have additional risk due to second jobs or dependence on public services such as transportation. Additionally, social behaviors and interactions that are independent of economic factors may differ between these groups. Further studies incorporating income and other social determinants are necessary to examine the influence of these factors on differences in SARS-CoV-2 infection across healthcare roles.

Our study and others suggest that SARS-CoV-2 transmission to HCP within the hospital environment is relatively rare if strict infection control practices are followed. 12, 21-24 Neither patient nor HCP contacts were associated with higher odds of testing positive for SARS-CoV-2. However, HCP who had contact with a COVID-19 positive family or community member experienced a >4-fold increased odds of a positive test.

Given that SARS-CoV-2 is highly contagious, and contact with a family or community member is not likely to occur under the same PPE standards as in the healthcare environment, one would expect these interactions to lead to transmission. In addition, the protective association on univariate analysis between contact with a patient with a respiratory illness and contact with a non-masked visitor is logical when put into context. First, if HCP reported contact with a patient who had a respiratory illness, but not known or suspected COVID-19, it most often meant that the patient had a negative SARS-CoV-2 test result. Secondly, non-masked visitors were primarily an event of the first few months of the pandemic when the local prevalence was very low and mask mandates J o u r n a l P r e -p r o o f were not strictly enforced. For the rest of study period, hospital visitation was more restrictive. Thus, the odds of HCP getting COVID-19 from either of these two contact situations should be low.

In our cohort, many specific symptoms were associated with a symptomatic HCP testing positive for SARS-CoV-2, including: cough, anosmia, ageusia, fever, myalgia, chills, and headache. We also demonstrated the importance of multiple symptoms from this group, as for each additional symptom the odds of testing positive for SARS-CoV-2 doubled. We believe this data is helpful to healthcare institutions who would like to create streamlined symptom questionnaires, prioritize testing queues, and triage HCP prior to definitive PCR testing. However, we do not think that symptoms alone may be a substitute for testing. 

Despite our study's strengths, including the size and breadth of the population, the duration of evaluation through periods of low, moderate, and high COVID-19 activity within our local community, the prospective approach to symptom and contact data collection, and equal and free access to testing, we also acknowledge limitations. Contact and symptom data are self-reported and we cannot conclusively determine all contact events. Likewise, genome sequencing was not performed routinely to link specific contacts with infection events. There is heterogeneity within some healthcare role groups in terms of job title, responsibilities, experience, and income. We were unable to examine specific SES and sociodemographic factors such as race. Finally, this is a single center cohort study, which may limit generalizability. Future studies incorporating social, behavioral, and economic determinants will be important to elucidate why we observed differences in SARS-CoV-2 infection across healthcare roles.

We demonstrated significant factors associated with positive SARS-CoV-2 testing in symptomatic HCP across three distinct areas: healthcare role, contact history, and symptoms. Our findings of different risk across HCP roles, after controlling for known contacts, reinforce the need for further investigations into COVID-19 disparities.

They suggest equal education and testing will not be enough to overcome differences.

Our findings highlight that existing infection control practices limit transmission to HCP within the healthcare environment. Lastly, the association between seven symptoms and positive test results, as well as the increased odds of testing positive with each J o u r n a l P r e -p r o o f additional symptom, may be useful in augmenting HCP testing strategies. While this study limited the analysis to HCP, it may be reasonable to apply some of our findings, including contact history with family/community members with known or suspected COVID-19 and the seven symptoms, to the screening process for other populations. J o u r n a l P r e -p r o o f 

Clinical symptoms, comorbidities and complications in severe and non-severe patients with COVID-19: A systematic review and meta-analysis without cases duplication

COVID-19 in Healthcare Workers: A Living Systematic Review and Meta-analysis of Prevalence, Risk Factors, Clinical Characteristics, and Outcomes

Prevalence of COVID-19 Infection and Outcomes Among Symptomatic Healthcare Workers in

risk assessment, surveillance and management of SARS-CoV-2 infection in health workers: a scoping review

Epidemiological and clinical characteristics of severe acute respiratory coronavirus virus 2 (SARS-CoV-2) infection among healthcare workers in Hubei Province

What do we know about SARS-CoV-2 transmission? A systematic review and meta-analysis of the secondary attack rate and associated risk factors

COVID-19 symptoms predictive of healthcare workers' SARS-CoV-2 PCR results

Diagnosis of COVID-19 Based on Symptomatic Analysis of Hospital Healthcare Workers in Belgium: Observational Study in a Large Belgian Tertiary Care Center during Early COVID-19 Outbreak

High prevalence of SARS-CoV-2 infection among symptomatic healthcare workers in a large university tertiary hospital in Sao Paulo

Strong associations and moderate predictive value of early symptoms for SARS-CoV-2 test positivity among healthcare workers, the Netherlands

COVID-19 symptoms, duration, and prevalence among healthcare workers in the New York metropolitan area

Implementation of infection control measures to prevent healthcare-associated transmission of severe acute respiratory coronavirus virus 2 (SARS-CoV-2)

Plotting regression coefficients and other estimates in Stata: University of Bern Social Sciences Working Papers

Assessment of Community-Level Disparities in Coronavirus Disease 2019 (COVID-19) Infections and Deaths in Large US Metropolitan Areas

Variation in racial/ethnic disparities in COVID-19 mortality by age in the United States: A cross-sectional study

Racial and Ethnic Disparities in Population-Level Covid-19 Mortality

Neighbourhood income and physical distancing during the COVID-19 pandemic in the United States

Mayo Foundation for Medical Education and Research

We would like to thank Mallory Wagner and Theron Schultz for assistance in data collection for this project. We would like to thank Jamie LaMantia for her editorial assistance.