key: cord-0695245-mt0a1wmj authors: Anan, Tomomi; Ishimaru, Tomohiro; Hino, Ayako; Nagata, Tomohisa; Tateishi, Seiichiro; Tsuji, Mayumi; Ogami, Akira; Matsuda, Shinya; Fujino, Yoshihisa title: Association between COVID-19 infection rates by region and implementation of non-pharmaceutical interventions: a cross-sectional study in Japan date: 2021-12-06 journal: J Public Health (Oxf) DOI: 10.1093/pubmed/fdab385 sha: 32d28b5e2f8792296ef8fc7085f11cdf479151dd doc_id: 695245 cord_uid: mt0a1wmj BACKGROUND: During a pandemic, non-pharmaceutical interventions (NPIs) play an important role in protecting oneself and others from infection. There are large regional differences in COVID-19 infection rates in Japan. We hypothesized that the local infection incidence may affect adherence to individual NPIs. METHODS: This cross-sectional study was conducted online among full-time workers in Japan in December 2020. The questionnaire asked the respondents to identify their habits regarding seven common NPIs (wearing masks, washing hands after the bathroom, disinfecting hands when entering indoors, gargling when returning home, ventilating the room, disinfecting or washing hands after touching frequently touched surfaces, carrying alcohol sanitizers when outdoors). RESULTS: A total of 27 036 participants were analyzed. Compared with the region with the lowest infection rate, five of the seven NPIs showed statistically significant trends across regional infection levels, the two exceptions being wearing masks and washing hands after the bathroom. Multivariate adjustment did not change these trends. CONCLUSIONS: This study found that NPIs were more prevalent in regions with higher incidence rates of COVID-19 in Japanese workers. The findings suggest that the implementation of NPIs was influenced not only by personal attributes but also by contextual effects of the local infection level. COVID-19 has spread rapidly all around the world since December 2019. The World Health Organization declared COVID-19 a pandemic on 11 March 2020. 1 Various efforts have been made at a policy level in many countries to prevent the spread of infection during the pandemic, which is ongoing. Many countries have implemented lockdowns, curfews of restaurants and bars, physical distancing, bans on social gatherings and school closures. By the end of 2020, vaccines against COVID-19 had been developed and their roll out had commenced. The Japanese government declared a 'state of emergency' in response to record numbers of cases and asked the public to refrain from going out, close restaurants or restrict their opening times, work remotely, and limit or cancel events. 2 During a pandemic, non-pharmaceutical interventions (NPIs) play an important role in protecting oneself from infection and preventing the spread of infection to others. NPIs are defined as behaviors or actions that individuals or communities can take to slow the spread of pathogens, such as washing hands, wearing masks and physical distancing; they do not include taking medication or receiving vaccines. 3 Centers for disease control and prevention (CDCs) recommend wearing a mask, staying at least 6 feet away from others, avoiding crowds and poorly ventilated spaces, washing hands frequently, cleaning and disinfecting frequently touched surfaces, and monitoring one's health daily. 4 The Japanese government has called on the public to focus on avoiding the 'three Cs': closed spaces, crowded places and close-contact settings and has reminded people to wash their hands, wear masks, ventilate their houses and avoid eating out with others. 5 Although the implementation of NPIs is important for preventing the spread of infection, some people do not follow the recommended interventions, and some even oppose them. It is already known that adherence to individual NPIs is influenced by demographic factors such as gender, age and place of residence; 6 social factors such as education, income 7, 8 and sources of information 9 and psychological factors such as anxiety, fear, political ideology and health beliefs. 7, 8, 10, 11 In the USA, many people in urban areas resist wearing masks, partly due to political ideology. 6, 12 Unlike in other countries, the Japanese government's infection control actions do not include mandatory measures such as lockdowns, but mostly request cooperation from the public. Therefore, individuals' attitudes toward the NPIs are a key issue for combatting the spread of infection. There are large regional differences in COVID-19 infection rates in Japan, which may lead to differences in people's risk perception. The risk perception regarding COVID-19 differs widely among places and individuals and can affect the spread of the virus. 13 One study reported that Japanese people placed most trust in local information 9 about infection status and hospital bed occupancy rates. As the infection rate in a region rises, people start to hear about cases close to them, and infection prevention measures are strengthened in workplaces and public places. We assume that people's risk perceptions are influenced by contextual and environmental factors. We hypothesized that the local infection incidence rate may affect adherence to individual NPIs. We considered that the incidence rate of COVID-19 in a given region is related to the implementation of such interventions through people's risk perception. To our knowledge, however, very few reports have examined the rate of implementation of NPIs in relation to local infection rates. Here, to test this hypothesis, we examined the association between regional differences in COVID-19 infection levels in Japan and individual NPIs. We conducted a survey from 22 to 26 December 2020, during the third wave of the COVID-19 pandemic, as a part of the Collaborative Online Research on the Novel-coronavirus and Work (CORoNaWork) Project. 14 This cross-sectional study was conducted online among full-time workers in Japan. The study was approved by the ethics committee of the University of Occupational and Environmental Health, Japan (reference No. R2-079 and R3-006). Participants provided informed consent by completing a form on the survey website. A total of 33 087 participants answered an online, selfadministered questionnaire. After excluding invalid responses, 27 036 were eligible for analysis. The age-range of the target population was 20-65 years. To take account of regional characteristics, the 47 Japanese prefectures were classified into four levels based on the level of infection. Completed questionnaire samples were extracted from these four regions to obtain equal sample sizes. Region 1 consisted of Fukushima, Yamaguchi, Aomori, Ehime, Yamagata, Nagasaki, Iwate, Tokushima, Shimane, Kagawa, Niigata, Tottori and Akita prefectures; region 2 of Nagano, Saga, Tochigi, Oita, Toyama, Okayama and Fukui prefectures; region 3 of Gunma, Ishikawa, Gifu, Kumamoto, Ibaraki, Miyagi, Hiroshima, Shiga, Mie, Kochi, Shizuoka, Wakayama, Miyazaki, Yamanashi and Kagoshima prefectures and region 4 of Tokyo, Kanagawa, Saitama, Chiba, Okinawa, Osaka, Hokkaido, Aichi, Hyogo, Fukuoka, Kyoto and Nara prefectures. The questionnaire asked respondents to report on their habits regarding seven types of NPIs in the last month, namely wearing a mask in public, washing hands after using the bathroom, disinfecting hands with alcohol sanitizers when going inside, gargling when returning home, opening windows or doors to ventilate the room, disinfecting or washing hands after touching frequently touched surfaces such as doorknobs or railings and carrying alcohol sanitizers when going out. Participants answered from the following options: 'always do', 'mostly do', 'not often' or 'almost never'. The infection level in the region where participants lived was measured by the incidence rate for the entire period since the pandemic was declared (per 1000 population), the number of people infected for the entire period, the incidence rate for one month before the survey (per 1000 population) and the number of people infected over the one month. Covariates included demographic and socioeconomic factors including age, sex, marital status, household income, educational background, job type, smoking status and the number of employees in the workplace. Age was used as a continuous variable. Marital status was categorized in three groups: currently married, divorced or widowed and never married. Annual equivalent household income was categorized into three groups: 4 70 000-26 50 000 Japanese yen (JPY), 26 50 000-45 00 000 JPY and 45 90 000-1 05 00 000 JPY. Educational background was categorized in three groups: graduated from junior high school; high school and university, graduate school, vocational school or junior college. Job type was categorized in three groups: mainly desk work (clerical or computer work), mainly work involving interpersonal communication (customer service, sales, selling, etc.) and mainly labor (physical work, nursing care, etc.). The number of employees in the workplace was classified into four categories: 1-29, 30-99, 100-999 and ≤1000. We show the participants' demographic information using counts and percentages (Table 1) . We compared seven representative NPIs by the four regions using one-way analysis of variance (Table 2) . Age-sex and multivariate adjusted odds ratios (ORs) of incidence rate of COVID-19 in areas associated with each NPI, defined by those who answered 'always do', were estimated with a multilevel logistic model and nested by area of residence (cities, towns and villages). In the multilevel model, the incidence rate of COVID-19 in an area was used as an area-level factor. The multivariate model was adjusted for age, sex, marital status, education, job type, annual equivalent household income, smoking status and number of employees in the workplace. A P-value less than 0.05 was considered statistically significant. All analyses were conducted using Stata (Stata Statistical Software: Release 16; StataCorp LLC, TX, USA). Table 1 shows basic characteristics of the participants. Regions are classified by the incidence of COVID-19 for the entire period since the pandemic was declared (per 1000). Region 4 had the highest infection rate, at 1.91, followed by region 3 (0.79), region 2 (0.51) and region 1 (0.28). Region 4 includes the Tokyo metropolitan area, which has more urban lifestyle characteristics than the other three regions, such as more single people, higher household incomes, higher education levels, more hospitality workers, fewer manual laborers and more workers in large companies. Table 2 shows the implementation status of the seven NPIs by region according to the incidence rate of COVID-19. Among the interventions, wearing masks in public places had the highest overall implementation rate, at 86.2%, followed in decreasing order by washing hands after using the bathroom (85.3%), disinfecting hands with alcohol sanitizers when going indoors (55.5%), gargling when returning home (50.9%), opening windows and doors for ventilation (45.0%), disinfecting or washing hands after touching frequently touched surfaces (37.2%) and carrying alcohol sanitizers when going out (31.0%). Six of the NPIs showed a trend in implementation according to area incidence rate of COVID-19 (all Ps < 0.001); the exception was handwashing after using the bathroom. Table 3 shows the multivariate analyses of the implementation of the seven NPI items. Compared with the region with the lowest infection rate (region 1), the OR for the region with the highest rate (region 4) was 1.24 (95%CI: 1.10-1.40, P < 0.001) for wearing a mask in public, 1.08 (95%CI: 0.97-1.20, P = 0.157) for washing hands after using the bathroom, 1.17 (95%CI: 1.01-1.35, P = 0.031) for disinfecting hands with alcohol sanitizers when going indoors, 1.54 (95%CI: 1.31-1.82, P < 0.001) for gargling when returning home, 1.45 (95%CI: 1.20-1.75, P < 0.001) for opening windows and doors for ventilation, 1.33 (95%CI: 1.18-1.51, P < 0.001) for disinfecting or washing hands after touching frequently touched surfaces and 1.32 (95%CI: 1.17-1.49, P < 0.001) for carrying alcohol sanitizers when going out. Five items showed statistically significant trends with regional infection levels: disinfecting hands with alcohol sanitizers when going indoors (P for trend = 0.030), gargling when returning home (P < 0.001), opening windows and doors for ventilation (P < 0.001), disinfecting or washing hands after touching frequently touched surfaces (P < 0.001) and carrying alcohol sanitizers when going out (P < 0.001). The two NPI items showing no significant trend with regional infection level were wearing a mask in public and washing hands after using the bathroom. Multivariate adjustment did not change any trends. This study revealed that people living in areas with higher levels of infection were more likely to engage in various nonpharmaceutical interventions. Non-pharmaceutical interventions have been recommended as a measure to prevent or slow the spread of COVID-19, along with organizational efforts such as emergency declarations and lockdowns. It has been reported that individual characteristics, such as sociodemographic factors, 6-8 sources of information 9 and psychological factors, 7, 8, 10, 11 are associated with non-pharmaceutical behaviors. In this study, the association between community infection level and preventive behaviors remained robust even after adjusting for individual factors. The results suggest that the implementation of NPIs is influenced by the regional context. There are at least three possible reasons why the regional infection level may affect the implementation of NPIs. First, the higher the level of infection in a region, the more people's fear of infection and their risk perception is likely to be affected, which in turn will lead to greater engagement with NPIs. Japanese people trust local government information more than other sources of information 9 and are relatively knowledgeable about the status of infection in their communities. Second, people living in areas with higher levels of infection will encounter more social situations that require NPIs, such as workplaces, restaurants and public facilities. Third, majority synching bias 15 may impact people's implementation of NPIs; this bias is the idea that it is safe to act in the way that people around you are acting and has been demonstrated in a study of factors associated with decisionmaking regarding evacuation during a disaster. 15 Moreover, a recent report found that the predominant reason for wearing a mask during a pandemic among Japanese people was that wearing a mask was a social norm, whereas the original purpose-preventing personal infection and transmitting it to others-was less important. 16 Majority synching bias is likely to be enhanced as the number of people performing NPIs increases in areas with high infection levels. Among the NPIs studied here, some showed a doseresponse relationship with the local infection level, whereas others-specifically, wearing masks and hand-washing after using the bathroom-did not. We assume that the latter two were not associated with the local infection level because their implementation levels were high, at ∼85%, regardless of infection level. In Japan, wearing masks has been widely practiced during epidemics of common colds, influenza, seasonal allergies and infectious gastroenteritis, even since before the current global pandemic. 17 A 2015 survey by the Consumer Affairs Agency, a Japanese government agency, reported that ∼85% of respondents said that they washed their hands after using the bathroom. 18 It is likely that the lack of a dose-response relationship in relation to COVID-19 and community infection levels is due to Japanese people being accustomed to wearing masks and washing their hands after using the bathroom, which start with hygiene education in childhood. In contrast, disinfecting hands with alcohol sanitizers when going indoors, gargling when returning home, opening windows and doors for ventilation, disinfecting or washing hands after touching frequently touched surfaces and carrying alcohol sanitizers when going out were found to have a doseresponse relationship with community infection levels. These were newly recommended approaches for helping to prevent the spread of COVID-19. In previous infectious disease epidemics, carrying alcohol-based disinfectant and washing or disinfecting hands after touching frequently handled doorknobs and handrails were not common prevention behaviors. In view of the widespread recognition that alcohol disinfection is effective against SARS-Cov2 19,20 and that COVID-19 is contact-transmissible, 21 it is likely that such behaviors were reinforced as the infection spread, especially in regions with higher infection rates. Although ventilating rooms is also widely acknowledged as a measure for preventing infection, it has not been routinely adopted because it causes the room temperature to drop during the winter, and many work environments cannot be ventilated appropriately due to the structure of buildings. During this pandemic, avoiding the 'Three Cs'-closed spaces with poor ventilation, crowded places and close-contact settings-has been strongly recommended, especially in Japan. 5 As many people are aware that ventilation is highly effective in preventing COVID-19 infection, it is not surprising that this NPI is more commonly performed in regions with higher infection rates. In Japan, people are taught from a young age that gargling is a preventive measure against infectious diseases, starting in schools and households. Gargling is practiced by many people during infectious disease epidemics. However, in the present survey, it emerged that gargling was not so widespread, with a practice prevalence of ∼50%. Although gargling may be effective in preventing upper respiratory tract infections in healthy adults, 22 it is not recommended in the CDC guidelines for infection prevention 4 or even in the Japanese guidelines for preventive measures against COVID-19. 5 Because of their belief in gargling's effectiveness, people in areas with higher levels of community infection may be more concerned about possible infection while away from home and, therefore, practice more proactive gargling behavior. The present results suggest that individual adherence to NPIs depends not only on individual characteristics but also on the contextual effect of local infection level. Considering that the waxing and waning of infection will be repeated in the future, it is important for people to actively implement NPIs even in areas with low infection levels, where organized campaigns to promote awareness of NPIs may be effective. There are several limitations in this study. First, we conducted a survey of full-time workers only; we did not include parttime workers, housewives, the elderly or people under 15 years of age. Accordingly, the results do not present a complete picture of NPI practices in the community; some reports show different rates of implementation of preventive measures in different age groups. 6 Second, given the cross-sectional nature of the study, it is not possible to assign causation. In other words, it is not possible to compare rates of NPI implementation before and after the increase in COVID-19 infection rate in a given region. Therefore, we cannot conclude that the rate of implementation has increased because of the increase in infectious rate. Third, we used a simple question to measure NPI adherence; details about frequency and timing remain unknown. Archived: WHO timeline -COVID-19 Office for Novel Coronavirus Disease Control, Cabinet Secretariat, Government of Japan. COVID-19 information and resources COVID-19 how to protect yourself & others Information on health and medical consultation Who is wearing a mask? Gender-, age-, and location-related differences during the COVID-19 pandemic Improving the impact of nonpharmaceutical interventions during COVID-19: examining the factors that influence engagement and the impact on individuals The use of facemasks to prevent respiratory infection: a literature review in the context of the health belief model Japanese citizens' behavioral changes and preparedness against COVID-19: an online survey during the early phase of the pandemic Social learning theory and the health belief model Health belief model for coronavirus infection risk determinants Republicans, democrats move even further apart in coronavirus concerns Risk perception and COVID-19 Protocol for a Nationwide Internet-based Health Survey of Workers During the COVID-19 Pandemic in 2020 Evacuation support system to control emotional level of crowd: simulation-based evaluation under various scales and ratios of guide robots for evacuees. Artif Life Robot Why do Japanese people use masks against COVID-19, even though masks are unlikely to offer protection from infection About half of all respondents wore masks, up 23 points in two months! (written in Japanese) [Internet] Ministry of Health Labour, and Welfare Hand sanitizers: a review of ingredients, mechanisms of action, modes of delivery, and efficacy against coronaviruses The pandemic of COVID-19 and its implications for the purity and authenticity of alcohol-based hand sanitizers: the health risks associated with falsified sanitizers and recommendations for regulatory and public health bodies transmission, diagnosis, and treatment of coronavirus disease 2019 (COVID-19): a review Prevention of upper respiratory tract infections by gargling: a randomized trial The current members of the CORoNaWork Project, in alphabetical order, are Yoshihisa Fujino ( The authors declare no conflicts of interest associated with this manuscript. This study was approved by the ethics committee of the University of Occupational and Environmental Health, Japan (reference No. R2-079 and R3-006). Informed consent was obtained in the form of the website.