1 INTRODUCTION

The SARS-CoV-2 (COVID-19) virus has disrupted life as we know it both at work and at home. Although the Centers for Disease Control and Prevention (CDC)1 along with state and local governments in the United States and the World Health Organization2 emphasized the importance of personal protective behaviors (PPBs) to help control the spread of COVID-19, much of the United States saw increases in infections throughout the fall and winter of 2020.3 The infection rate in the 25–49 age groups increased dramatically throughout the United States during summer and fall of 2020, while the infection rate in persons over the age of 65 remained relatively stable.3 In the United States, 25–49 age group was also less likely to report wearing face coverings, practicing social distancing and avoiding congregating with groups of <10 compared to persons ≥65 years of age.4 Fifty-two percent of practicing nurses in the United States are under the age of 505 and thus belong to the demographic with the highest rate for community spread of COVID-19 in 2020. However, little is known about their own behaviors as it relates to engagement in PPBs outside their work setting and whether their behavior is influenced by their professional identity.

2 BACKGROUND

In the early days of the pandemic, the CDC began publishing guidelines to help individuals, employers, and communities protect themselves from the spread of COVID-19.1 In the summer and fall of 2020, the CDC on their main COVID-19 prevention homepage, promoted the following behaviors: frequent hand hygiene, masking in public places, social distancing of 6 feet or greater, avoiding groups of more than 10 people, daily sanitizing of home surfaces, isolating from others in household when feeling sick, and wearing face mask in house when others are sick. These protective behaviors were widely disseminated across healthcare and community settings in an effort to curb the spread of COVID-19 both in healthcare settings and in communities.

As a vital member of the healthcare team, nurses' health and safety is paramount to the functioning of the healthcare system. Protection of the healthcare worker and patient is the cornerstone of infection control in the inpatient setting. Research regarding compliance with infection control measures in inpatient settings has generally focused on compliance with personal protective equipment (PPE) such as facemasks, gloves, shields.6-8 Fatigue,6 personal fear of infection, exposure to virus, years' experience,8 discomfort and work location have been associated with PPE compliance in the hospital setting.6-8 Less attention has been paid to the factors that inform nurses' use of PPBs outside the work setting. Reynolds et al.9 reported only 24% of healthcare workers complied with all quarantine orders after severe acute respiratory syndrome (SARS) exposure and only 29% expressed concern for infecting others. During the H1N1 pandemic, one study found that 54.9% of healthcare workers expressed worry over infecting family and friends while only 6.6% restricted their social contacts.10 Higher levels of worry over becoming infected were associated with limiting social contacts.10 These findings suggest that healthcare workers may find it difficult to comply with infection control practices outside the work setting.

Health promotion and ability to protect the health of populations during public health emergencies are key elements of professional nursing education.11 In addition, the American Nurses Association (ANA),12 International Council of Nursing13 and the National Health System of England14 have endorsed views supporting nurses as role models for healthy living. In 2017, the ANA12 launched the Healthy Nurse, Healthy Nation Campaign™, with the goal of impacting the nation's health by encouraging nurses to engage in healthy lifestyle behaviors to positively impact both the personal health of nurses and the health of the communities in which they serve. Although nursing promotes health promotion and disease prevention as a key role of the professional nurse,11, 13 there is little in the literature to support that nurses have integrated this concept into their professional identity outside the work setting.14-16

Nursing, as the largest healthcare profession, is vital to the efficient functioning of the healthcare system in the United States and globally, especially in times of crisis. However, little is known as to nurses' personal engagement with PPBs outside work, which may reduce the spread of infection or decrease their risk of acquiring infection within their community. The COVID-19 crisis in the provides a unique opportunity to explore nurses' engagement in infection prevention practices when not at work, and personal and professional factors that may inform this engagement. Understanding these factors is necessary in ensuring the health of our nursing workforce, healthcare system, and our communities while we deal with COVID-19 and future health crises. Thus, the purpose of this study is to understand nurses' engagement in PPBs outside the healthcare setting and factors that may influence their use during the COVID-19 pandemic.

3 OBJECTIVES

Specific Aim 1: To characterize nurses' use of CDC promoted COVID-19 PPBs outside the work setting.

Specific Aim 2: To understand factors that inform nurses' use of CDC promoted COVID-19 PPBs outside the work setting including fatigue, vulnerability to infection, and view of self as role model.

4 METHODS 4.1 Design and setting

We conducted a cross-sectional online survey in a large academic healthcare system in the southeastern United States. The healthcare system includes four community hospitals serving both rural and metropolitan areas across five counties, including a large metropolitan academic teaching hospital.

4.2 Sample and recruitment

All registered nurses, including advanced practice registered nurses, employed full or part-time were invited to participate in the study. Survey links were emailed to all eligible participants the week of September 14, 2020. The links remained active for a period of 2 weeks, participants could only respond once to the survey, and their responses could not be linked to their email address. Reminders were sent three times over the course of 2 weeks. Participants were given the opportunity to participate in a random drawing to win one of three $50 gift certificates after completing the survey. If respondents chose to participate in the random drawing, they were directed to a separate site to enter the drawing; this ensured that their contact information could not be linked to their responses on the survey. Study data were collected and managed using REDCap (Research Electronic Data Capture).17 We calculated that we needed 355 responses based on a confidence level of 0.95 and a margin of error of +/−5%.

The survey was composed of questions with a variety of response formats (including Likert format, free text, and multiple choice). The nursing research committee of the organization developed the personal behavior questions and focused on CDC recommended preventive behaviors and behaviors that had been described in the media and posted on social network sites by healthcare workers. The committee piloted the survey before distribution. Fatigue, vulnerability to disease, and select demographic and work history variables were measured based on previous research related to PPE compliance.

4.3 Measures

Respondents were queried regarding the following variables: the makeup of their household (children, adults) exposure to patients with COVID-19, if they or someone they lived with was at risk for higher morbidity or mortality related to COVID-19 based on the CDC risk groups, their exposure to patients with COVID-19, and personal history of COVID-19 infection. Demographic data collected included age, race, education level, years' experience as a RN, and current work environment.

To assess vulnerability to disease, the 7-item Perceived Infectability (PI) subscale of the Perceived Vulnerability to Disease Scale (PVD) was used.18 This subscale measures individual's belief in their own susceptibility to disease. Responses were based on a 7-point Likert scale (strongly disagree to strongly agree), mapped to values of 1 through 7 and summed to calculate a total score. Three items were reversed coded so that higher values represented stronger agreement with respect to vulnerability. The scale has been used in a variety of research settings and has shown discriminate and predictive validity and reliability across various populations.18 In the current study, the reliability of the PI subscale was 0.84.

Fatigue was measured using the Fatigue Assessment Scale © FAS (Fatigue Assessment Scale): ild care foundation (http://www.ildcare.nl/).19 The FAS is a 10-item general fatigue questionnaire composed of two subscales: mental fatigue (5 items) and physical fatigue (5 items). Each FAS question is assessed on a 5-item Likert-scale with responses from 1 (“never”) to 5 (“always”), with 2 items reverse coded so that higher scores correspond to greater frequency of fatigue (range of scores 10–50). This scale has been used to measure fatigue in nurses20 and other healthcare workers.21 The FAS has good convergent and divergent validity.22 The FAS was treated as a unidimensional scale22 for this study. Internal consistency in our sample was 0.87.

There is not an available or extensively validated instrument to measure nurses' perception of themselves as role models. Other studies have asked participants to rate their agreement with statements related to their self-perception as a role model.15 For this study, participants were asked to rate their perception of oneself as role model using a 100 mm visual analogue scale provided in the REDCap platform and weighted with higher score reflecting greater identification of self as role model. Two separate questions were used to determine the degree to which a respondent perceived themselves as a role model for their patients and community. Scores could range from 0 to 100.

4.3.1 Personal protective behaviors

Participants rated the frequency in which they engaged in the CDC most recommended PPBs (frequent hand hygiene, mask in public places, and social distancing of 6 feet or greater, avoiding groups of more than 10 people, daily sanitizing of home surfaces, isolate from others in household when felling sick, and wearing face mask in house when others are sick)1 outside the hospital setting using a 5-point Likert response format (1-never, 2-rarely, 3-sometimes, 4-usually, 5-always). Participants were also given the option of “not applicable” if they did not have the opportunity to practice the behavior in the previous two weeks.

The Institutional Review Board (IRB) approved the study (IRB00068198) with a waiver for written informed consent. Email invitations to participants contained informed consent.

4.4 Data analysis

Statistical analysis was performed in R (version 4.02). Our primary outcome(s) of interest were PPBs responses and our primary predictors included: age, education, work place in the hospital, provision of direct patient care (yes/no), years of experience, concern of acquiring COVID, perceived self-risk, FAS score, PI score, confidence in social distancing, confidence in masks, confidence in hand hygiene, whether or not participants had had COVID, and whether or not nurses saw themselves as a role model for their patients or for their community. Confidence measures (for social distancing, in masks, and hand hygiene were all continuous responses from 0 to 100, with 0 representing no confidence and 100 representing absolute confidence). Descriptive statistics for our primary predictors of interest were computed using means and SD for continuous variables, and frequency (percentage) for categorical variables.

4.4.1 Missing data

Missing data was scarce, with 7% or less among each predictor variable of interest and <2% across PPBs outcome responses. Missing data was assumed to be missing at random. We used multiple imputation by chained equations (across 10 imputed data sets) to impute missing data for each PPB question and predictor variables of interest.23 Statistical analysis was performed on the imputed data sets using pooled estimates.

4.4.2 Analysis

There were in total 8 CDC promoted PPBs. Our primary analysis investigated two outcome measures of interest for behaviors: (A) The average “frequency” with which nurses practice PPBs across their set of applicable PPBs, and (B) the proportion of PPBs practiced “often” or “always” out of all PPBs applicable. For (A), we calculated the mean response for each participant out of the number of “applicable” questions to them. We then fit a linear regression model for the CDC protective behaviors, modeling mean frequency response by the set of our defined predictors of interest. For (B), we dichotomized each PPB response to “always or often” versus “never, rarely, or sometimes” and for each respondent we summed the number of PPBs with which they responded “always” or “sometimes.” We modeled the proportion of applicable PPBs practiced via logistic regression using a quasi-likelihood framework to account for overdispersion.24 For all models, we determined statistical significance at a 0.05 ɑ level.

5 RESULTS

A total of 4621 nurses (RN and advanced practice RNs) were sent links to the survey and 958 responded (response rate 20.7%). The mean age of respondents was 45 (SD = 12.6) and average experience as an RN was 17.5 years (SD = 12.5). Most respondents self-identified as White (83.6%) and 76.4% had a bachelor's degree or higher, 85.9% provided direct patient care and 55% reported providing direct care to a COVID-19 patient. Only 2.8% and 2.5% of respondents had either tested positive for COVID-19 or lived with someone who had, respectively, at the time of the study. Only 13.6% of nurses reported being very concerned about contracting COVID-19. Of the 8 CDC promoted PPBs, covering mouth to cough and wearing facemask (86.6% and 78.2% reporting “always”, respectively) were the two most frequently practiced behaviors (see Figure 1). See Tables 1 and 2 for complete demographic and factor descriptives.

Personal protective behaviors most frequently practiced in descending order. CDC promoted behaviors on their website in September 2020 [Color figure can be viewed at wileyonlinelibrary.com] Table 1. Demographic characteristics and other predictors N (%) Mean (SD) Median (Min, Max) Missing (%) Age 44.6 (12.6) 44.6 (12.6) 45.0 (21.0, 74.0) 10 (1.0%) Race 3 (0.3%) American Indian or Alaska native 5 (0.5%) Asian 36 (3.8%) Black/African American 67 (7.0%) Don't know/not sure 2 (0.2%) Native Hawaiian or other Pacific Islander 1 (0.1%) Other 19 (2.0%) Prefer not to answer 24 (2.5%) White 803 (83.6%) Education 2 (0.2%) Some college or technical school 4 (0.4%) Associate's degree 220 (22.9%) Bachelor's degree 453 (47.2%) Some graduate school 59 (6.1%) Graduate degree 222 (23.1%) Missing 2 (0.2%) Years clinical experience 17.5 (12.5) 15.0 (0, 52.0) 17 (1.8%) Direct patient care 2 (0.2%) No 133 (13.9%) Yes 825 (85.9%) Work area 0 (0%) Adult Med-Surg 182 (19%) Adult ICU 91 (9.5%) Adult ED 66 (6.9%) Ambulatory Clinic 168 (17.5%) Designated COVID Unit 17 (1.8%) Pediatrics inpatient 84 (8.8%) Pediatrics outpatient 22 (2.3%) Other 330 (34.4%) Family/self at high risk for complications from COVID-19 6 (0.6%) No 576 (60%) Yes 378 (39.4%) Concern for COVID-19 5 (0.5%) No Concern 122 (12.7%) A little 237 (24.7%) Somewhat 260 (27.1%) Moderate 209 (21.8%) Very 127 (13.2%) Fatigue 20.6 (7.37) 19 (10,49) 57 (5.9%) Perceived Infectability 22.1 (8.10) 21 (7.0,49) 40 (4.2%) Confidence In Distancing 63.4 (28.6) 70 (0,100) 65 (6.8%) Confidence In Masking 70.2 (29.7) 78 (0,100) 52 (5.4%) Confidence in Hand Washing 83.5 (19.4) 91 (0,100) 54 (5.6%) Live with someone who has tested positive 8 (0.8%) No 922 (96%) Yes 30 (3.1%) Has tested positive for COVID-19 5 (0.5%) No 917 (95.5%) Yes 38 (4%) Role model for patients 82.1 (21.1) 90 (0.0, 100) 51 (5.3%) Role model for community 79.0 (22.7) 86.0 (.0, 100) 55 (5.7%) Table 2. Linear regression model for mean frequency of practice of CDC personal protective behaviors with select descriptive statistics Variable Effect estimate 95% CI lower 95% CI upper p value Model R2 (Intercept) 2.287 0.45 Age 0.006 0.00 0.01 0.021 Work area Emergency Depart. Referent Adult ICU 0.191 0.02 0.36 0.030 Direct Patient Care No No REF Yes −0.114 −0.22 −0.01 0.036 Concern for COVID No concern Referent A little 0.322 0.19 0.45 0.000 Somewhat 0.517 0.38 0.65 0.000 Moderate 0.588 0.45 0.73 0.000 Very 0.754 0.59 0.92 0.000 Fatigue −0.007 −0.01 0.00 0.010 −0.007 −0.01 0.00 0.010 Confidence in masking 0.006 0.00 0.01 <0.001 Self as role model for community 0.004 0.00 0.01 0.060 Note: Fitted model adjusted for all covariates presented in Table 1; only variables with p < .10 presented. 5.1 Mean frequency of PPBs practice

Regression results are found in Table 2. Assessing the mean practice frequency of PPBs, we found that increasing confidence in masks (p < 0.001), increasing concern about COVID (p < 0.001), and increasing age (p < 0.021) were significantly related to higher mean practice frequency, explaining 45% of the variability adjusted for all predictors. We found that participants involved in direct patient care had significantly lower mean PPB practice frequency (p < 0.04) and participants with higher fatigue (FAS) scores had significantly lower mean PPB practice frequency (p = 0.01).

5.2 Proportion of PPBs practiced

We found that increased confidence in masks (p < 0.001) and increasing concern about COVID (p < 0.001) were significantly related to higher proportions of PPBs practiced (See Table 3). We additionally found that decreasing fatigue score (p = 0.004) was significantly related to higher proportion of practiced behaviors.

Table 3. Logistic regression estimated odds ratios for proportion of CDC personal protective behaviors engaged in Odds ratio Odds ratio 95% CI lower Odds ratio 95% CI upper p value Age 1.010 1.000 1.021 0.053 Work in COVID unit (relative to ED) 1.768 0.965 3.240 0.064 Work in adult ICU (relative to ED) 1.498 1.069 2.098 0.019 Direct patient care 0.731 0.573 0.933 0.012 Concern for COVID Not at all Referent A little 1.