Handwashing knowledge, attitudes, and practices during the COVID-19 pandemic in Saudi Arabia: A non-representative cross-sectional study

Handwashing (HW) with water and soap is one of the cheapest and most effective ways of protecting oneself and others against the coronavirus. Here, the HW knowledge, attitudes, and practices of Saudi adults were assessed during the COVID-19 pandemic using a cross-sectional study conducted between May 8 and June 8, 2020, during a partial lockdown period. A web-based validated questionnaire was distributed through different social media platforms, and the sociodemographic characteristics of the participants, seven items related to knowledge, four items related to attitudes, and thirteen items related to the practice of HW were assessed. A total of 1323 (51% male and 49% female) adults from all regions of Saudi Arabia responded to the questionnaire. The overall mean (± SD) was 5.13 (± 1.18) for knowledge of HW and COVID-19, 2.79 (± 0.77) for attitude toward HW, and 7.8 (± 2.56) for HW practice. A multiple linear regression analysis revealed factors associated with knowledge to be age and family income. Sex, educational level, family income, and HW knowledge were associated with negative and neutral attitude, whereas age, sex, family income, and HW knowledge were associated with practice. These results suggest that HW knowledge was strongly associated with positive attitudes toward HW and correct HW practice in Saudi adults during the COVID-19 lockdown.

Questionnaire preparation and description. A self-administered questionnaire written in Arabic language (S1) was developed by the investigators using information published recently to address the objectives of the study 12,15,22 . A pilot study was conducted with 25 males and 25 females of different ages, who were asked to complete the questionnaire and then report on whether it was easy to understand and what the estimated time for completion was. These results are not included in this report, but some modifications were made to ensure clarity and ease of understanding the questions. The face and content validities of the questionnaire were reviewed by three experts (two epidemiologists and one consultant in infection control; all had more than 10 years of experience in their fields). Each expert was approached individually by the principal investigator to review the questionnaire and add their comments. The experts performed the reviewing task separately and made notes and responses, with an average of two times before final approval. The questionnaire scoring system is described as follows: first, in the knowledge and practice sections, 1 point was assigned for each correct response/active practice and zero for incorrect response/passive practice. Second, in the attitude section, 1 point was assigned for positive attitude, while zero was assigned for negative and neutral attitude.
Cronbach's alpha was calculated for each section: knowledge (0.64), attitude (0.63), and practice (0.62) of the questionnaire. The final version of the questionnaire comprised 34 multiple-choice questions in four sections: sociodemographic data (10 items), knowledge of HW (7 items), attitudes regarding HW (4 items), and HW practices (13 items). Data collection. The web-based questionnaire, which included forced-fields questions, was placed on Google-Forms platform. Then distributed via several social media platforms. Its title, the objective of the study, the voluntary nature of participation, declarations of confidentiality and anonymity, estimated duration (5-6 min), and URL link were included on the cover page. The inclusion criteria were Saudi adults aged 18 years or older living in the Kingdom of Saudi Arabia. The questionnaire was randomly shared on social media and personally shared with the investigators' Saudi national contacts lists. The frontpage of the questionnaire showed the study title, purposes, and inclusion criteria (Saudi National and ≥ 18 years) to ensure respondent eligibility. The study procedure was approved by the institutional review board at Qassim University, and conducted in accordance with the ethical principles for medical research involving human subjects as described in the Declaration of Helsinki. Informed consent was obtained from all participants in this study. This study's ethical approval was obtained from the Qassim Region Research Ethics Committee on May 5, 2020 (Reference No. 19-11-02).
Statistical analysis. The data were received in Excel sheet and manually surveyed to exclude duplicated responses, after that exported to STATA version 16 for statistical analysis. The data are presented as a mean (SD) for continuous variables or as a number (percentage) for categorical variables. Comparisons between continuous variables were performed using either the Student's t-test or one-way ANOVA for more than two groups, and the chi-squared test was used to compare the categorical variables. A multivariable regression analysis was per- www.nature.com/scientificreports/ formed to identify the determinants of HW knowledge, attitudes, and practices. Variables such as the respondent's age category, sex, educational level, marital status, the presence of elderly people in the same household, and chronic disease were considered independent variables; knowledge, attitudes, and practices were the dependent variables. Variables with p < 0.25 in the bivariate analysis were included in the multivariable analysis. A multiple linear regression analysis was performed to determine the factors that affected the knowledge score. A multiple logistic regression analysis was performed to determine the factors that affected attitudes and practices. Odds ratios (ORs), beta coefficients, and 95% confidence intervals (CIs) are reported, as appropriate. A p value of less than 0.05 was considered strong evidence against the null hypothesis.

Sex
Female 228 (35) 595 (92) 628 (97)  Handwashing practice. Bivariate analysis revealed that factors significantly associated with never/sometimes follow of HW techniques were male sex (p = 0.031), family income + 5.333 USD (p = 0.008), and mean of HW knowledge score of 5.08 (p = 0.010). Factors associated with washing the hand for less than 40 s were healthy persons versus those with chronic disease (P = 0.009) and a mean HW knowledge score of 4.92 (p ≤ 0.001). Factors that were associated with the frequency of HW < 10 times/day were age group (18-29) (p = 0.005), male sex (p ≤ 0.001) and mean HW knowledge score 5.07 (p = 0.002).
Bivariate analysis. Factors that were associated with not washing hands after visiting public places were age group ≥ 60 years (p = 0.002), educational level ≤ secondary school (p ≤ 0.001), and mean HW knowledge score 4.18 (p ≤ 0.001). Not washing one's hand after touching a high touch surface outside was significantly associated with an age group ≥ 60 years (p ≤ 0.001), male sex (p = 0.001), educational level ≤ secondary school (p ≤ 0.008) and mean HW knowledge score of 4.47 (p ≤ 0.001). Not washing hands after removing gloves was associated with an age group ≥ 60 years (p = 0.001), male sex (p ≤ 0.001), educational level ≤ secondary school (p ≤ 0.003), mean HW knowledge score of 4.48 (p ≤ 0.001), and mean HW attitude score of 2.71 (p = 0.033). Not washing hands after removing masks was significantly associated with an age group ≥ 60 years (p = 0.017), male sex (p = 0.032), educational level ≤ secondary school (p ≤ 0.029) and mean HW knowledge score of 4.47 (p ≤ 0.001). Table 6.
Following the correct handwashing technique. Using the correct HW technique was assessed by seven items, which are listed together with their scores in Table 2. The correct HW technique was always observed by 394 (30%) participants. Multiple logistic regression analysis (Table 7) showed that never or sometimes following the correct HW technique was more common among males (vs. females, aOR 1.31, p = 0.031), those with family incomes above + 5333 USD (vs. < 2666 USD, aOR 1.92, p ≤ 0.001) and those with family incomes from 2666 USD to < 5333 USD (vs. < 2666 USD, aOR 1.37, p = 0.022), those with HW knowledge score (aOR 0.85, p = 0.003) and among those aged 18-29 years (vs. all other age groups).
Following the correct handwashing duration. A multiple logistic regression analysis showed that only male sex was a risk factor for washing hands for less than 40 s (aOR 1.27, p = 0.038), while other factors, such as educational level (less than secondary vs. a bachelor's degree), chronic illness (yes vs. no), and handwashing knowledge scores showed protection against washing hands for less than the recommended duration, as shown in Table 7. www.nature.com/scientificreports/ Daily handwashing frequency. Among the factors investigated, only male sex (aOR 1.91, p < 0.001) indicated a risk for washing one's hands less than ten times daily, whereas age (40-49 years vs. 18-29 years) and age group (50-59 years vs. 18-29 years) and handwashing high knowledge score predicted HW more than ten times daily, as shown in Table 7.  www.nature.com/scientificreports/ Not washing hands after removing a face mask. Among the investigated factors, only the hand-washing knowledge score (aOR 0.63, p = 0.001) was significantly associated with washing hands after removing face mask.

Discussion
Previous studies have investigated HW as a preventive measure against many infectious respiratory system diseases, such as SARS, H1N1 influenza, and avian influenza [23][24][25] , but most of these studies were conducted on healthcare staff [26][27][28] . The current study was community-based and involved respondents from all regions of the Kingdom of Saudi Arabia.
The major finding of this study was that the overall mean knowledge level in this sample was 5.13. This indicates that the majority of the study population had good knowledge of HW to prevent COVID-19 infection. This finding is in agreement with that of Mahdi et al. 17 who conducted a study about hand hygiene knowledge, attitude, and practice (KAP) among domestic Hajj pilgrims and in contrast to Mahdi et al. 18 who conducted a similar KAP study among visitors to the Prophet's Mosque in Al Madinah, which reported a moderate level of knowledge. Yet, this comparison should be considered with caution, as both mentioned studies investigated both Saudi and non-Saudi subjects. It is worth mentioning that during the COVID-19 pandemic, the health authorities in Saudi Arabia launched a broad, multi-language health education campaign that targeted all populations, including both Saudi and non-Saudi nationals. This campaign included national TV programs, the Ministry Table 7. Multiple logistic regression of factors associated with handwashing practice (technique, duration, frequency, and key time). Follow HW technique (never/sometimes vs. always); duration of HW (< 40 s vs. 40 s +); frequency of HW (≤ 10 times vs. > 10 times). *p < 0.05. www.nature.com/scientificreports/ of Health website, social media platforms, mobile phone SMS messages, and direct public awareness sessions. The broadcast materials contained information about the virus, its transmission, and precautionary measures, including HW 29 .
The WHO recommends HW-along with many other measures-for reducing the transmission rate of COVID-19 infections worldwide 30 . Alcohol-based hand rubbing solution is one of the methods recommended by the WHO for hand cleansing, and it has determined the effective alcohol concentration to be a minimum of 60% and ideally 80% 1,31 . In this study, only 30% of the respondents correctly identified the effective alcohol concentration of alcohol-based hand sanitizers. The level of knowledge reported in this study was associated with many variables in the bivariate analysis, including age, educational level, gender, family income, and the medical history of the respondents. It was noted that those aged 30-39 and those with high educational levels had significantly higher mean knowledge scores than those aged over 60 or those with average educational levels. Moreover, women scored significantly better than men-in the multivariable analysis. A high monthly family income was significantly associated with a higher mean knowledge score, which is in line with a similar survey about SARS in Hong Kong 32 . High income likely allows individuals to obtain smart phones, smart TVs, and secure internet access, ensuring access to more methods of health education and awareness that are broadcast in the media. Those with low incomes may have limited resources for such access. Old age decreased the knowledge score compared to younger participants. This may be because younger participants are more expert and familiar with using smartphones and the internet and are more active in social media than older participants, and most of the health education programs and campaigns are easily accessible through these technological innovations 33 .
In this study, the attitude level was typically aligned with the level of knowledge, as the overall attitude score was 70%. Less than half of the respondents (42%) thought that they were vulnerable to COVID-19. This finding is in contrast to a study from Hong Kong, in which 84.8% felt they or a member of their family would contract SARS 34 . This low vulnerability prediction is attributed to the fact that at the time of data collection, the number of cases globally was around 4 million cases, and in Saudi Arabia only 44,000 cases in 14th May 2020 35 . It is worth mentioning that male respondents felt that they had a greater risk of contracting COVID-19 than females did. This is in contrast to a study in Hong Kong, which reported that females felt they were more likely to contract Swine flu than males 36 . This may be because women were more likely to stay at home because their children were banned from visiting public places during the partial lockdown. In this study, female respondents had more appropriate attitudes toward touching their faces while wearing gloves and washing their hands after removing their gloves. This attitude reflects the psychology of females toward dealing with the risk of infection 32 , and it may help women feel less vulnerable to COVID-19 infection, as was observed in this study.
Another finding is that high educational levels increase the feeling of being vulnerable to contracting COVID-19, not touching your face while wearing gloves, and actively directing family members to wash their hands when needed. This may be explained by more education being correlated with a greater awareness of the seriousness of this disease and an increased sense of vulnerability. However, increased education was also associated with positive attitudes, as most of the moderately educated and highly educated respondents believed that HW reduces infection risk, which is consistent with a recent study from Saudi Arabia by Al-Hanawi 15 . Notably, educational level, high income, and high knowledge were linked in the present study.
Another important finding in this study is that only 47% of the participants thought that they could actively ask their family members to cleanse their hands after outdoor activities; those family members might prefer staying home rather than going outside, or they may follow precautionary measures without the need for extra instructions. In the bivariate analysis, many factors were predictive of attitude, including age, sex, educational level, and family income. More specifically, an elderly age group (60 + years) was more likely than other age groups to think that touching one's face while wearing gloves is allowed. This negative attitude is not surprising, as this age group had the lowest knowledge score.
This study found that 71.8% of respondents always followed the best HW technique, but only 48% washed their hands for at least 40 s. 27% wash their hands more than ten times daily. This result is far lower than the 43% reported by Lau et al. for Hong Kong, although the respondents in that study performed HW more than ten times only when traveling across mainland China 37 . Interestingly, 94% of the respondents washed their hands after visiting public places, which is far higher than the 63.6% reported from China at the time of the SARS outbreak 38 . Furthermore, 81% of our respondents washed their hands after touching a high-touch surface, consistent with the 81.2% reported for Hong Kong participants but higher than the 48.3% reported for Singaporeans at the time of SARS 39 . In this study, younger respondents, male respondents, and those with low HW knowledge scores had the lowest HW frequency. This remained significant in the multivariable analysis. Women had better knowledge than men, reflecting that women practice HW more accurately and more frequently than men, which is in line with three recent studies from Saudi Arabia that reported women as more likely to adhere to self-protective measures than men 15,16,39 . Women are known to have a lower threshold for uncleanliness than men, and they may therefore be more concerned about infections and cleansing knowledge and behavior 40 . In the multivariable analysis, a larger monthly income was negatively associated with the accurate practice of HW, contrary to a study from the Philippines 41 . Perhaps, those with high family income have a job environment that does not necessitate frequent hand washing compared to the other jobs. The presence of chronic illnesses did not affect the technique or frequency of washing hands, but it affected the duration of HW, as the majority of those with chronic illness washed their hands for more than 40 s. This remained significant in the multivariable analysis. Washing one's hands for 40-60 s is recommended by the WHO 42 . It is important to understand that it is not only water and soap that cleanse the hands, but also mechanical forces and the techniques of HW. To correctly perform the techniques requires at least 40-60 s 42 .
In the literature, older persons living in the house may encourage other family members to practice HW frequently, and other family members may practice more HW out of caution and concern about transmitting the infection to an older person 43 . In this study, the presence of elderly people living in the home did not affect www.nature.com/scientificreports/ HW practices either positively or negatively. There is no clear explanation for this. When interpreting the results of this study, there are some limitations that should be considered. First, this study was conducted during a period of social distancing and was a self-administered questionnaire conducted online. The distribution of the questionnaire depended on the researchers' connections; it was therefore a convenience sample rather than a random one. Thus, selection bias, report bias, and cohort differences cannot be ruled out. Second, although non-Saudi represents around 37% of the total Saudi Arabia populations 21 , the study population was limited to the Saudi nationals only. Therefore, this study is non-representative. Third, responses to some of the questions in the questionnaire might be easily predicted, which may affect the reliability of the results. Fourth, yes/no/may-be types of questions were used over using a Likert scale. This approach is powerful in addressing the respondent's knowledge, yet it may underestimate the respondent's attitude and practice views. Fifth, the anxiety levels of the respondents were not measured, which may have explained more of the findings. Sixth, the study design was cross-sectional, so causality cannot be determined. Therefore, further studies using different designs with revered/distracted questionnaire answers are needed.