Study design and participants
We conducted two rounds of cross-sectional surveys among the residents in Hangzhou, China. Hangzhou is a well-developed city in eastern China, where there is a considerable number of migrant workers and tourists from all parts of the country. According to national data, Hangzhou had more than 6000 public toilets, ranked fourth nationwide in 2018 [34]. The first round of data collection was conducted to evaluate the hygiene perceptions and practices among the Hangzhou residents before the COVID-19 pandemic, between November and December 2018. To evaluate the improvement in the hygiene perceptions and practices during the pandemic, we conducted the second round of data collection between December 2020 and January 2021 when the pandemic was still in place but people’s life almost returned to normal. The target population were individuals over 16 years old residing in Hangzhou. Mandarin is recognized as the official spoken language, while the Hangzhou dialect is a regional spoken language. As our investigators could not speak Hangzhou dialects (i.e., unofficial language), we excluded elderly residents who only understood Hangzhou dialects but not Mandarin. All six major districts in Hangzhou (i.e., Shangcheng District, Xiacheng District, Jianggan District, Gongshu District, Xihu District, Binjiang District), where there are high density of population and public toilets [35], were purposely chosen. We chose two communities in each of six districts from its community lists based on random numbers as our survey sites. We conducted both rounds of data collection at the same survey sites. In each round of data collection, we recruited 20 to 30 residents in each community using a street-intercept method [36] that our research assistants invited passers-by from the study sites to participate in our study. The respondents of two rounds were recruited from the same communities but not the same individuals because we were not able to contact the same participants of the first round of data collection without their contact information. Self-administered paper questionnaires were distributed by our research assistants to the participants. Most participants completed the questionnaires by themselves and handed them back to the research assistants. For some elderly participants who had difficulties reading, our research assistants were on site to support and assist their reading and understanding of the questionnaire.
The questionnaire was developed based on literature review and the qualitative study adapted from our previous studies [37]. In this study, we converted the original questionnaire from the traditional Chinese version (i.e. writing system used in Hong Kong) into the simplified Chinese version (i.e. writing system used in mainland China) and modified several wordings of the questions to fit the cultural context in mainland China. Despite that Hangzhou has mixed settings of public squat and sitting toilets, with preponderance of squat toilets, we decided to keep the questions related to using sitting toilets in the original questionnaires. The questionnaire was pilot-tested by face-to-face interviews with five laymen in Xihu district, Hangzhou, who did not have an education background of medicine, pharmaceutics, biology or public health, and showed good validity. There were 18 different research assistants for each round of data collection. Before data collection, all the research assistants had standard training for 30 minutes concerning how to recruit study participants, how to elaborate the questions when necessary, and how to check the validity of collected questionnaires. In both rounds of data collection, two researchers (JX & XX) were responsible for training and supervision.
We estimated the proportion of having hygiene related behaviors in this study at 70% according to the data from a similar study conducted among the Chinese residents in Hong Kong [37]. Based on the estimate of 70% [37], we set a target effective sample size of 291 for each round of data collection, providing 80% power to detect the ±10% margin of difference, with a two-sided type I error rate of 0.05, according to the sample size formula for comparison of two proportions [38]. The total sample size was adjusted to 364 for each round considering a potential non-response rate of 20% from our prior survey experience. There were 312 and 314 valid questionnaires in the first and second round of data collection, respectively. All participants were informed that their participation was confidential, voluntary, anonymous, and that they could quit at any time. A study compensation worth five RMB (US$ 0.77) was provided as a token of gratitude for their time participating in the study.
We conducted an additional observational study at toilets in different public places located at the study sites (about five public toilets in each community) to record toilets’ provision of hygiene amenities, between February and March 2021, during the second round of data collection. Four trained data collectors (two males and two females) were responsible for observing all of the sampled public toilets. Those who were responsible for managing public toilets might replenish hygiene amenities when they noticed an investigation taking place. Therefore, our data collectors pretended as public toilet users and recorded the data using a cellphone to fill the digital structured checklist immediately upon arrival to avoid the Hawthorne effect [39]. The structured checklist recorded toilets’ provision of hygiene amenities, including paper tissue, alcohol disinfectants, tap water, soap, paper towels, and working hand dryers, as well as types of toilets. The observational study lasted for 2 weeks, as the four data collectors visited five public toilets in one community for 8 h per day on average, from 12 noon to 8 pm. In total, 70 public toilets were observed.
Measures
During the pandemic, many approaches were adopted to prevent disease transmission among the general population. More intensive cleaning and disinfection were conducted to mitigate the transmission caused by touching contaminated surface in public places [20]. Public places, such as hotels, provide toothpicks for their guests as a tool to press the button on the elevator. Besides, the government provide hygiene amenities and hygiene promotion activities/messages to raise residents’ hygiene awareness and promote hygiene practices during the pandemic. Alcohol disinfectants were distributed to residents, along with advocation of disinfecting their home [40]. Guidance on personal hygiene and health education materials (e.g., pamphlets, posters, booklets) were widely publicized [22, 41].
Hygiene behaviors in public toilets
Respondents’ hygiene behaviors when using sitting toilets were collected by asking how often they would: (1) clean the toilet seat with alcohol; (2) clean the toilet seat with tissue paper; (3) put tissue paper on the toilet seat before using; (4) flush with the toilet lid closed. Respondents’ hand hygiene behaviors after using the toilet were collected by asking their frequency to (1) wash their hands after using the toilet; (2) wash their hands with soap after using the toilet; (3) dry their hands with paper towels; (4) dry their hands with a dryer. Response choices were “always,” “sometimes,” or “never”. We defined those respondents who chose “sometimes/never” responses as having low compliance with hygiene behaviors, thus they were combined for analysis.
Hygiene awareness of hygiene behaviors in public toilets
Respondents were asked to choose all of the hygiene behaviors in public toilets that they believed would increase the risk of disease transmission with “Yes” and “No” options among the following choices: (1) touching contaminated toilet facilities; (2) flushing the toilet without the lid closed; (3) not washing one’s hands after using the toilet; (4) not using soap to wash one’s hands after using the toilet; (5) not drying one’s hands after washing them.
Risk perception of using public toilets
Residents’ perceived risks of getting infected when using public toilets were measured by three items using a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree): (1) toilet hygiene has no direct relationship with infectious diseases; (2) worried about getting infected in the public toilet; (3) public toilet is a breeding ground for infectious agents. Residents who chose “4–agree” and “5–strongly agree” were categorized as one group that indicated agreement with the statement; residents who chose “1–strongly disagree,” “2–disagree,” and “3-neutral” were categorized as another group that did not indicate agreement.
Sociodemographic factors
Sociodemographic factors included gender (male/female), marital status (unmarried/married), age (18–39 years/40–59 years/60 years or older), education level (middle school and under/high school/college and above), and monthly household income (less than ¥5,000/ ¥5,001–¥18,000/ and more than ¥18,000).
Statistical analysis
Descriptive analysis was conducted to show frequencies and percentages. Chi-squared test was used to evaluate the similarity of sociodemographic characteristics between the respondents in the first and the second rounds. Bivariate analysis was used to compare the differences in hygiene behaviors in public toilets, hygiene awareness, and risk perceptions between the two rounds of data. We used multivariable logistic regressions to evaluate the differences of the residents’ toilet hygiene behaviors/perceptions between the two rounds, with the first-round respondents treated as the reference group. Regression models were adjusted for gender, marital status, age, education level, and household income. All statistical analyses were performed using SPSS 24.0 with the statistical significance set at p < 0.05.