A cross-sectional analysis was conducted using a web-based questionnaire survey on January 27‐29, 2021, when the COVID-19 pandemic was ongoing in Japan. The central government had declared a state of emergency during the survey period in 11 of the 47 prefectures (Tochigi, Saitama, Chiba, Tokyo, Kanagawa, Gifu, Aichi, Kyoto, Osaka, Hyogo, and Fukuoka) according to the number of infections, situation of the medical system, and the level of development of monitoring systems [15-17]. It was the second declaration, 8 months after the first emergency declaration in Japan [15-17].

An internet-based research company, Macromill, Inc, Japan was engaged to recruit participants and collect responses and the company’s registered monitors include over 30 million individuals living in Japan. The survey included 1236 participants aged between 20 and 79 years from across Japan. The company was responsible for recruiting participants and collecting responses. The participants were randomly selected, and an email was sent inviting them to take the survey. An equal number of participants from both genders and 10-year age groups were recruited, with registrations closing once the targeted sample size was achieved. The participants completed the questionnaires through email and received a small reward, as determined by the company.

Cross-sectional data from the web-based questionnaire included information on participants’ backgrounds, health-related items, changes in outing behaviors due to the pandemic, and personality traits (Table 1). Economic status was reported to influence outing behavior during the pandemic [37], and the analysis included participants’ income as an adjustment variable. Changes in the frequency of three types of outings were included: medical institution visits, eating out, and traveling. Scoring was performed on a five-point Likert scale (1=Increased, 2=Somewhat increased, 3=No change, 4=Somewhat decreased, 5=Decreased).

The survey referred to the Japanese version of the Personality Inventory, which is based on the “Big Five” model to assess personality traits [38]. Japanese personality traits “sense of morality,” “diligence,” and “strong sense of responsibility” were also included because these could influence preventive behaviors toward COVID-19 [26,39]. To minimize the burden on participants, the items were grouped into six categories: “diligent,” “sociable,” “cooperative,” “honest,” “strong sense of responsibility,” and “strong sense of morality.” Regarding the response, the original seven-point Likert scale was modified to a six-point scale for aggregation and analysis purposes [15,40]. The participants were asked to indicate their agreement levels for each personality trait using the scale (1=Strongly disagree, 2=Disagree, 3=Somewhat disagree, 4=Somewhat agree, 5=Agree, 6=Strongly agree).

A multinomial logistic regression was performed to analyze the association between personality traits and changes in outing behaviors. After verifying multicollinearity, all independent variables representing personality traits were simultaneously inputted. Additionally, a stratified analysis was performed by age group. We classified participants into two groups: younger and older adults. Considering the number of participants in each age group, the older group included individuals aged 60 and above.

Regarding the variables, the scoring for each outing was classified into three groups—increase (1=Increased and 2=Somewhat increased), no change (3=No change), and decrease (4=Somewhat decreased and 5=Decreased)—which were used as dependent variables, and “no change” was used as the baseline category for multinomial logistic regression. The responses for personality traits were categorized into two groups—applicable (4=Somewhat agree, 5=Agree, and 6=Strongly agree) or not applicable (1=Strongly disagree, 2=Disagree, and 3=Somewhat disagree)—which were considered as independent variables. The following variables were included as covariates: gender, age, marital status, living alone, living with children (aged younger than 15 years), living with older adults (aged older than 65 years), low household income (less than JPY 4 million (USD 36500) annually), self-reported disease, self-reported health, employment status, and emergency declaration status.

For all statistical analyses, we used STATA 17/SE (StataCorp LLC) with a statistical significance of P<.05.

This study was approved by the Ethics Committee of the National Institute of Public Health, Japan (NIPH-IBRA#12302, approval date: November 17, 2020), in accordance with the Declaration of Helsinki. All participants provided informed consent for data collection and storage. Written informed consent for study participation was obtained upon registration. The web-based questionnaire survey was conducted by an authorized survey company in adherence with personal information protection regulations. Anonymized data were obtained from the company after survey completion.

The analysis revealed two main findings. First, the impact of COVID-19 on outing behavior varied depending on the type of activity. Although most participants reported no change in the frequency of medical institution visits, the majority reported a decrease in the frequency of eating out and traveling. To compare the age groups, a higher percentage of older adults reported no change in medical institution visits and a decrease in eating out and traveling compared with younger adults. Second, the impact of COVID-19 on outing behavior was associated with particular personalities. Participants with a sociable personality were more likely to report a decrease in the frequency of medical institution visits and an increase in the frequency of eating out, while participants with a strong sense of responsibility were more likely to report a decrease in the frequency of traveling. The association between changes in outing behavior and personality differed by age group. Younger adults showed similar trends across all participants; meanwhile, older adults with a strong sense of responsibility were more likely to report a decrease in the frequency of eating out.

For visits to medical institutions, previous studies have revealed that seeking emergency treatments remained unchanged and screening examinations decreased during the pandemic [11,12,36,41]. Examining visits to medical institutions, which may include outpatient treatment, we found that this aspect remained unchanged for most participants during the pandemic. Although telemedicine was introduced in Japan in 2018, its application remained limited even during the pandemic; thus, its impact on medical visits was considered minimal, if any. Regarding eating out, studies in Japan [24], Poland [38], and the United States [23] found a reduction in frequency during lockdowns, consistent with our findings. Similarly, for travel, previous studies reported that the COVID-19 pandemic and infection control measures led to a decrease in long-distance travel [42-45], consistent with our findings. Regarding the variation in results by age group, Harrera et al surveyed during the pandemic in Spain and clarified the association between older age and higher frequency of preventive practices, higher health literacy, and greater concern about COVID-19 [22]. They also reported that increased adherence to preventive practices was associated with being female, having greater concern about COVID-19, and seeking more information among the younger generation; meanwhile, for the older generation, it was associated with greater concern about COVID-19 [22]. Concerns about COVID-19 might also have influenced our result that a higher proportion of older adults reduced the frequency of eating out and traveling.

Importantly, three types of outing behaviors were simultaneously investigated, revealing that medical institution visits remained unchanged, whereas eating out and traveling decreased, with the trend differing among age groups. This suggests that the impact of the pandemic on outing behavior varied depending on necessity.

The findings also indicated that outing behaviors in response to COVID-19 were associated with particular personality traits. Personality traits have been reported to be associated with various behaviors related to the COVID-19 pandemic [25,26,29,30]. For example, Qian and Yahara [26] determined an association between preventive behaviors and personality traits such as extraversion, neuroticism, openness, conscientiousness, and agreeableness, as well as a sense of morality. Hygiene and social distancing behaviors have also been associated with personality traits such as extroversion, honesty, modesty, conscientiousness, and neuroticism [25,30]. In a study conducted in Italy, Costantini et al [30] identified personality traits that predicted compliance with COVID-19 prevention measures, including outing behavior. Their findings revealed that the frequency of outing behavior during the pandemic was negatively associated with modesty and creativity and positively associated with liveliness and sociability. The findings on eating out among younger adults align with those of this study. Although the previous study found a positive association between the frequency of outings and sociable personalities, we revealed that sociable people were more likely to report a decrease in the frequency of one type of outing: medical institution visits. This inconsistency may attributed to the evaluations in previous studies focusing on “unnecessary outings” [30].

Cultural aspects were reported to influence behavior during the pandemic. Abuliezi et al [31] reported the characteristics of preventive behavior among Japanese students compared with those in the United States. They determined that Japanese medical students performed significantly better in preventive behavior, although perceived health competence was lower, which means people do not have the confidence to manage their health through health-related habits [31]]. Furthermore, Japanese people tend to avoid social pressure and negative evaluations from others [33,34]. To sum up these cultural aspects, the finding regarding the association between sociable personality and an increase in the frequency of eating out among younger adults may be attributed to peer pressure to eating out or such a trend among peers.

Furthermore, changes in people’s perceptions and behaviors after the pandemic have been reported [46]. Although a better attitude and higher knowledge levels were observed, adherence to protocols remained inconsistent one year after the outbreak of the pandemic among younger adults aged 16‐30 years in Bangladesh [46]. This survey was also conducted a year after the first emergency declaration; therefore, the results might be influenced by changes in people’s perceptions and attitudes toward the pandemic over time. Considering the results and the survey periods of this study, concerns regarding COVID-19 among older adults might have changed their outing behavior even during the early stages of the pandemic. For a deeper understanding, further analysis would be required to clarify how the experience of the pandemic and repeated emergency declarations affected attitudes and behaviors.

Although this study effectively elucidates the association between personality traits and different types of outings during COVID-19 and differences by age groups, there are certain limitations that are worth mentioning. First, the participants responded to the questions based on their memories of the situation prior to the pandemic, and the actual number of outings before and during the pandemic was not observed; thus, response bias is possible. Furthermore, response bias resulting from the high social desirability of Japanese people could have affected the results [47]. Second, given the equal recruitment across age groups and genders, the participants did not represent the general population structure of Japan. In accordance, it should be noted that descriptive statistics regarding changes in the frequency of outing behavior do not represent the situation in Japan. Finally, we cannot eliminate the possibility of selection bias, as we only included participants registered with an internet panel company who had access to the internet and could respond on the web. Nevertheless, panel-based surveys have been widely used in recent years [48-50]. In addition, the participants received a small cash reward for completing the survey, which could have affected the randomness of the sample. However, because the participants were randomly selected from a large number of people registered with the engaged company across Japan, bias is expected to be minimized.

We examined various behavioral changes observed during the pandemic for different types of outings and their associations with personality traits and differences between age groups. The findings can help in understanding how to effectively communicate and promote appropriate behaviors in public health emergency settings.