Knowledge, attitude, and use of protective measures against COVID-19 among nurses: a questionnaire-based multicenter cross-sectional study

The oPt has been affected by the ongoing COVID-19 pandemic since March 5, 2020. Later, cases were reported in all regions and governorates. This study was a cross-sectional survey that was conducted among nurses in the oPt from October 2020 to December 2020 utilizing a paper-based questionnaire. The study involved nurses from multiple healthcare centers/hospitals from all governorates in the oPt (Fig. 1). The study is reported in adherence to the guidelines for reporting cross-sectional studies in which a questionnaire was used as the study tool (Supplementary Table S1) [36‐38].

The target population in the present study was nurses who practice their jobs in various healthcare institutions in the oPt during the ongoing pandemic. A sample size calculator that is freely accessible online at (www.​raosoft.​com) was used to compute the sample size to be recruited in this study. Assuming a maximal population of 20,000 nurses practicing in the oPt during the ongoing pandemic, the sample size was computed at a 95% confidence interval (CI) with a margin of error of ≤5%. The sample size to be included in this study was estimated as 385 nurses. Quota sampling was used to recruit the study participants for this study. Quotas were proportionate to the population of each governorate. The latest statistics of the Palestinian Central Bureau of Statistics were used. To achieve the sample size to be included in this study, it was decided a priori that 500 nurses would be invited to participate in this study. The decision to invite this number of nurses for this study was informed by previous studies conducted among nurses and other healthcare providers [39‐41].

The study objectives and methodology were explained to the nurses. The nurses had to provide their written informed consent before they could participate in this study. The inclusion criteria were: 1) a practicing nurse, 2) providing written informed consent to participate in this study, and 3) expressing willingness to respond to a questionnaire. Nurses who did not provide a written consent to participate in this study, and those who expressed unwillingness to respond to a questionnaire were excluded from the current study. Participation in this study was voluntary and the nurses were not offered any financial incentives as a compensation for their participation in the current study.

The study tool was adopted from previous studies that were conducted among healthcare providers [27, 42‐44]. The paper-based questionnaire was handed to potential participants through a personal contact in the healthcare centers/hospitals. The nurses were instructed that the principal investigator and the personal contacts were available and could be contacted in case the nurses needed to clarify any point. In the first section of the questionnaire used in this study, nurses were asked to report their gender. Previous studies reported differences in knowledge, attitudes, and use of protective measures against COVID-19 among male and female healthcare providers [22, 27, 29, 41, 43, 45‐47]. The nurses were also asked to self-rate their satisfaction with their financial status, social status, academic achievements, and perceived knowledge about COVID-19 using a Likert-scale of 1–3 (1 = low, 3 = high). Previous studies have shown that satisfaction with financial status, social status, and academic achievements were associated with knowledge, attitudes, and behaviors [48]. During the ongoing pandemic, many healthcare centers suffered shortages of personal protective equipment and healthcare providers had to purchase their own sanitizers, gloves, masks, and other personal protective equipment [49]. Probably, dissatisfaction with one’s financial status might influence purchasing sanitizers, gloves, masks, and other personal protective equipment. Lately, social networks have emerged as an important source of information to the general public as well as to healthcare providers. During the pandemic, the latest news and information about COVID-19 went viral on social media networks [50]. Probably, socially active nurses had larger social networks and could have received more information that influenced their knowledge, attitudes, and use of protective measures against COVID-19. Satisfaction with academic achievements and perceived knowledge about COVID-19 were collected to investigate if there was an association between these variables and performance of the nurses in knowledge, attitudes, and use of protective measures against COVID-19 items. The nurses were also asked to report whether they have contracted COVID-19 before or not. This variable was collected to investigate if contracting COVID-19 affected nurses’ knowledge, attitudes, and use of protective measures against COVID-19. Additionally, nurses were asked to provide their sources of information about COVID-19. In the second section, nurses were asked to respond to a knowledge test of 8 items on the causative agent of COVID-19, signs and symptoms, similarity with flu/cold, treatment, high risk patients, risk for infection, protection, and myths. On each item, the participants had to respond either by true, false, or I do not know. Attitude of nurses was measured using 2 items relevant to finally controlling COVID-19 and confidence in the health authorities in the oPt to win the battle against the disease. On each item, the participants had to choose either disagree, neutral/not sure, or agree. Use of protective measures was examined using 3 items relevant to using soap/sanitizer, physical distancing, and wearing personal protective equipment. On each item, the participants had to report their use either by yes which meant always or most of the time or no which meant not always/most of the time.

The questionnaire was pilot tested for readability and comprehension with 12 nurses who did not participate in the full study. In this pilot testing, the nurses read the questionnaire and provided their feedback on the clarity and comprehensibility of the items. Based on the feedback received from the nurses in this pilot, some items were reworded to enhance readability and comprehensibility. To ensure stability of scores over a short period of time, the test-retest method was used. A total of 25 nurses who did not participate in the larger study were asked to respond the questionnaire. After a period of 30 min to 1 h, the same 25 nurses were asked to respond to the questionnaire again. Pearson’s correlations were used to correlate the scores of the 25 nurses in the 2 rounds. It was decided a priori that a correlation coefficient of more than 80% would be needed to ensure stability of scores over the short time period that was let between the 2 rounds [51‐53]. The Pearson’s correlation coefficient was 0.96 (95% CI of 0.91 to 0.98) with a p value of < 0.001 which indicated excellent stability of scores. The internal consistency of the items in the questionnaire was assessed using Cronbach’s alpha. Internally consistent tools have a Cronbach’s alpha of ≥0.70 [54]. When all items were included, the Cronbach’s alpha was 0.75. When the Cronbach’s alpha was computed for each domain separately, knowledge items had a Cronbach’s alpha of 0.76, attitude items had a Cronbach’s alpha of 0.97, and use of protective measures had a Cronbach’s alpha of 0.87. Which indicated that the items were internally consistent across all domains.

The items used in the knowledge test were psychometrically evaluated by their difficulty index as calculated by the ratios of correct answers for each item [53, 55, 56]. Questions were psychometrically attributed as: 1) 0% ≤ “very difficult” < 30%, 2) 30% ≤ “difficult” < 60%, 3) 60% ≤ “moderate” < 80%, 4) 80% ≤ “easy” < 90%, and 5) 90% ≤ “very easy” < 100% [53].

For each knowledge item, the nurses were awarded 1 point for each correct answer and 0 for each incorrect/I don’t know answer. Points were summed (possible points could range from 0 to 8) and were transformed into percentages of correct answers (possible percentages could range from 0 to 100%). Nurses could rate each attitude item on a Likert scale of 1–3. The possible raw ratings (scores) on the attitude items could range from 2 to 6. Attitude scores were transformed into percentages. The practice items were dichotomous (no/yes) and scores could range from 0 to 100%. Data obtained in this study were entered and analyzed statistically using IBM SPSS for Windows, version 21.0 (IBM Inc., Armonk, NY). As the sample size was more than 300, normality of distribution was assessed using absolute skewness and kurtosis values [57, 58]. To fulfil the criteria for normal distribution, the absolute skewness had to fall within the range of − 2.0 and + 2.0 and the absolute kurtosis had to fall within the range − 7.0 and + 7.0. As the criteria for normal distribution were fulfilled, the data were expressed as mean (SD). Differences between knowledge, attitude, and use of protective measures scores among the nurses were investigated using either Student’s t-test or Analysis of Variance (ANOVA) with Bonferroni test as appropriate. Knowledge, attitude, and use of protective measures scores were correlated using Pearson’s correlations. To control potentially confounding variables, predictors of higher knowledge, attitude, and use of protective measures were identified using multiple linear regression analyses. The variables with a p-value of < 0.25 in the student’s t-test, ANOVA, and/or Pearson’s correlations were retained in the multiple linear regression models. Enter method was used. For each multiple regression model, the adjusted R-squared with a p-value of < 0.05 was used to evaluate the goodness-of-fit. Tolerance and variance inflation factor (VIF) values were used to assess the multicollinearity of the regression models. Absence of multicollinearity was ensured by tolerance values of > 0.1 and VIF values of close to 1 [59, 60]. In this study, p-values of ≤0.05 were considered statistically significant.

This study was conducted in adherence with the principles of the Declaration of Helsinki and the ethical principles followed at An-Najah National University. The current study received ethical approval from the Institutional Review Board (IRB) of An-Najah National University. The nurses provided written informed consent before they participated in the current study.

In this, of the 500 nurses invited, 455 (91.0%) completed the questionnaire. The sociodemographic and other variables of the study participants are shown in Table 1. Of the study participants, 285 (62.6%) were female in gender, 385 (84.6%) self-rated their financial status as moderate or high, 430 (94.5%) self-rated their social status as moderately or highly satisfactory, 355 (78%) self-rated their academic achievements as moderately or highly satisfactory, and 425 (93.4%) self-rated their knowledge about COVID-19 as moderately or highly satisfactory. Of the participants, 40 (8.8%) reported that they have contracted COVID-19.

When the study participate were asked to provide their sources of knowledge about COVID-19, 415 (91.2%) of the participants indicated that they obtained information about COVID-19 through the internet/social media, 230 (50.5%) obtained information about COVID-19 through TV/radio. Friends/family/acquaintances, awareness brochure/leaflet, courses taught at the university, and newspapers/magazines were also cited as sources of information for the participants. Details of the sources of information about COVID-19 are provided in Table 2.

The mean knowledge score was 75.7% (SD: 12.4%). Of the nurses, 140 (30.8%) scored 80% and above in the knowledge test. Of the 8 knowledge items, 4 (50.0%) were attributed as “very easy”, 1 (12.5%) was attributed as “easy”, 1 (12.5%) was attributed as “moderate”, and 2 (25.0%) were attributed as “difficult”. No item was attributed as “very difficult”. Of the participants, 390 (85.7%) could correctly identify the causative agent of COVID-19 disease as a virus, 445 (97.8%) could correctly identify the signs and symptoms of COVID-19 disease. However, only 100 (22.0%) of the participants knew that signs and symptoms of COVID-19 disease were similar to those of flu or cold, and 355 (78.0%) knew that there was no current effective treatment for COVID-19 disease. Of the participants, 410 (90.1%) could correctly identify immunocompromised and older individuals as have higher risk for infection and complications of COVID-19, 450 (98.9%) could identify crowded places as source of infection, and 415 (91.2%) knew that using masks appropriately can prevent the spread of COVID-19. However, only 155 (34.1%) knew that antibiotics cannot prevent transmission of COVID-19. Detailed answers of the participants are shown in Table 3.

The mean attitude score was 75.1% (SD: 17.7%). Of the nurses, 215 (47.3%) scored 80% and above on the attitude items. Of the participants, 290 (63.7%) were positive that COVID-19 will finally be successfully controlled and only 115 (25.3%) had confidence that the health authorities in the oPt would win the battle against COVID-19. Detailed responses of the participants are shown in Table 4.

The mean use of protective measures against COVID-19 score was 91.6% (SD: 18.2%). Of the nurses, 430 (94.5%) scored 80% and above on the practice items. Of the participants, 415 (91.2%) reported using soap or sanitizer to wash their hands and faces, 395 (86.8%) avoided unnecessary close contact and practiced physical distancing, and 440 (96.7%) reported wearing necessary personal protective equipment during interaction with the patients. Detailed responses of the participants are provided in Table 5.

Spearman’s correlations showed that there was a significant low positive correlation between knowledge scores and use of protective measures against COVID-19 (Pearson’s r = 0.27, p value < 0.001). Similarly, there was a significant low positive correlation between attitude scores and use of protective measures against COVID-19 (Pearson’s r = 0.13, p-value = 0.007). Details of the correlations between knowledge, attitude, and use of protective measures against COVID-19 are shown in Table 6.

Table 7 shows differences in knowledge, attitude, and use of protective measures against COVID-19 among the participants. In this study, knowledge scores were significantly lower for nurses who were male, self-rated their financial status as low, self-rated their social life as low, and self-rated their knowledge about COVID-19 as low compared to nurses who were female and those who self-rated their financial status, social life, and knowledge about COVID-19 differently (Table 7). Attitude scores were significantly lower for nurses who were female, having contracted COVID-19 before, self-rated their financial status as high, self-rated their academic achievements as low, and self-rated their knowledge about COVID-19 as high, compared to nurses who were male, have not contracted COVID-19, self-rated financial status, academic achievements, and knowledge about COVID-19 differently (Table 7). Use of protective measures scores were significantly lower for nurses who self-rated their academic achievements as low and self-rated their knowledge about COVID-19 as low compared to those who self-rated their academic achievements and knowledge about COVID-19 differently (Table 7).

Multiple linear regression showed that higher knowledge scores were predicted by being female (p-value = 0.004) and self-rating social status as high (p-value = 0.005). Attitude scores were predicted by being female (p-value = 0.005), self-rating academic achievement as high (p-value = 0.007), and having contracted COVID-19 (p-value = 0.001) (Table 8). Other variables were no longer significantly associated. Use of protective measures scores were predicted by self-rating academic achievements as high (p-value = 0.010) (Table 8). Other variables were no longer significantly associated. The tolerance values of the regression models were > 0.1 (in the range of 0.88 to 0.99) and the VIF values were in close to 1 (in the range of 1.01 to 1.14). These values indicated absence of multicollinearity among the predictors.

This is the first study among nurses in general and among healthcare providers in the oPt with regard to their knowledge, attitude, and use of protective measures against COVID-19 during the ongoing pandemic. In this study, the response rate was 91.0%. The response rate obtained in this study was high when compared to response rates reported in previous studies in which a questionnaire was used as a study tool among healthcare providers including nurses [39, 41, 48, 52]. Interestingly, the number of nurses who responded to the questionnaire was larger than the sample size needed for this study. This should have minimized the potential bias associated with low response rates. Additionally, the nurses who responded in this study were from both genders and had variable financial, social life, academic achievements, and self-rated knowledge of COVID-19. The sample also included nurses who previously contracted COVID-19. This diversity might have added validity, depth, and width to the findings of this study. Although the tool used in this study was adopted from previous studies, the tool was revalidated in a pilot testing using appropriate tests [27, 42‐44]. Findings of the pilot testing phase indicated that the tool was suitable to be used to assess knowledge, attitude, and use of protective measures against COVID-19 nurses [51‐53]. This might have allowed exposing the current knowledge, attitude, and use of protective measures against COVID-19 among nurses practicing across the healthcare centers/hospitals in the West Bank of the oPt.

The findings of this study should also be interpreted considering the following limitations. First, this study was a cross-sectional study. The findings might change with time and knowledge might increase as the pandemic continue unfolding. Additionally, the findings could have been more interesting should an intervention to improve knowledge, attitude, and use of protective measures was attempted. However, findings of this study might be informative to decision makers who wish to intervene by designing appropriate measures aiming to protect future nurses by improving knowledge, correcting attitude, and promoting adequate use of protective measures against COVID-19 among nurses. Second, the self-rated financial status, self-rated social status, self-rated academic achievements, and self-rated knowledge about COVID-19 were measured using a three-point Likert scale. Although the Likert scale is popularly used in medical research, the number of scale points to be used is still highly controversial [71]. Previous studies have used Likert scales with a number of points that ranged from 3 to 11. In a previous study, Leung administered the Rosenberg Self-Esteem Scale among 1217 students in Macau using different number of points and showed that there were no significant differences in Cronbach’s alpha, item–item correlations, item–total correlations, factor loadings, mean scores, and standard deviation of the scores [71]. Although the study of Leung advocated the use of large number of points (> 6 points), five-point Likert scales are commonly used in medical research. In this study, the use of a three-point Likert scale might have influenced the number of nurses who self-rated their financial status, social status, academic achievements, and knowledge about COVID-19 as moderate. This could have limited generalization and/or comparison of the findings to other settings. Third, the number of items measuring knowledge, attitude, and use of protective measures against COVID-19 among nurses with regard to COVID-19 was relatively small. Additionally, attitudes of the nurses were measured using only 2 items. However, the tool was previously used to assess knowledge, attitudes, and practice among healthcare providers in other settings studies [27, 42‐44]. Despite the inherent disadvantages, the use of small number of items in a questionnaire has many advantages including increasing participation, avoiding participant fatigue, and saving the time of the participants [72, 73]. Additionally, the items used to assess knowledge ranged from very easy to difficult. However, no question was attributed as very difficult in this study. Fifth, the use of protective measures against COVID-19 items collected perceived practice behavior. Although, social desirability bias cannot be excluded, it is noteworthy mentioning that the study participants were nurses who cared for infected patients during an ongoing pandemic. This could also, at least in part, explain the reportedly high use of use of protective measures against COVID-19. Finally, a nonprobability sampling technique was used to recruit the nurses to this study. Compared to probability sampling, nonprobability sampling techniques are inherently biased. This might limit generalization of the findings to the entire population of nurses. Fourth, knowledge of the nurses might have been underestimated as a results of recall bias. During the pandemic, nurses as well as other healthcare providers were exhausted and had to work for extended shifts. Previous studies conducted elsewhere including neighboring Jordan have reported high prevalence of burnout among healthcare providers [12‐14, 74]. It is possible that the exhaustive work conditions during the pandemic have affected the results.