Instant and long-term mental health concerns are common responses to the COVID-19 pandemic (8, 9). Two meta-analyses revealed that the pandemic increases the prevalence of stress, depression, anxiety, and insomnia (10, 11). COVID-19 patients, patients with preexisting psychiatric symptoms, and healthcare workers were found to suffer from poorer mental health conditions (12). In general population, female gender, younger age, chronic illnesses, unemployment, student status, and frequent exposure to social media/news concerning COVID-19 were significant correlates linked to mental distress (13). Psychological resources, such as personalized psychological flexibility, resilience, and extraversion were found to act as mediators when explaining the psychological distress among the general population (14, 15). In summary, descriptive epidemiological studies on this topic are predominant, while research on specific psychosocial mechanisms behind mental health problems, especially the interaction between objective and subjective factors among the general population during this pandemic is in its infancy (12, 16).

In China, numerous studies are carried out since the Wuhan outbreak to facilitate mental health treatment and care during the pandemic (17, 18). Despite the various sample size and measuring tools, these studies reached a consensus that mental health symptoms were common among the general population in China during the COVID-19 outbreak. Individuals’ characteristics, such as gender, age, education, occupation, and geographical location were found to be significantly associated with mental wellbeing (19–22). In particular, residents in Hubei Province had increased odds of mental health symptoms such as depression, anxiety, and posttraumatic stress symptoms (23–25). Nonetheless, there are hardly any studies focused on the psychological status of the general population living in Wuhan, who were exposed to a huge risk of infection in the earliest epicenter of the COVID-19 outbreak in China.

COVID-19 is an infectious disease that is easily transmitted and rapidly spreading. That means, exposure to COVID-19, no matter by contacting with an infected object or environment, or by human-to-human approach in a family or community, has a degree of risk of getting infected (26). Existing evidence suggested that exposure to infectious diseases of the respiratory system is closely associated with individuals’ mental health outcomes, among which the frontline healthcare practitioners obtained more academic attention (27). For instance, Wu et al. (28) revealed that hospital employees who worked in high-risk locations or had friends or close relatives who were contracted, were more likely to exhibit posttraumatic stress symptoms during the SARS epidemic in Beijing, China. In a Korean study, medical staff who were engaged in MERS-related tasks had highest risk to exhibit post-traumatic stress disorder symptoms (29).

Jiang et al. (30) conducted a similar study among the general population in China. They cited an official guiding principle of emergent psychological crisis intervention and divided the participants recruited by social media into four groups. The participants ranged from level 1 (patients with severe symptoms of COVID-19, frontline medical workers, CDC researchers, or administrative staff) to level 4 (people in affected areas, susceptible groups, or the general public). They found that exposure level was significantly linked to participants’ post-traumatic stress symptoms (PTS) only in the mildly PTS symptom subgroup, rather than the moderate and high PTS symptoms subgroups.

As suggested in prior studies, exposure to infectious diseases such as COVID-19 is uneven across different populations. Social and environmental factors contribute to shaping people’s exposure risk (31). In addition to the occupational factor mentioned earlier, geographical location is also closely related to the individual’s exposure. In China, Wuhan is the first city that reported confirmed cases of COVID-19 at the end of 2019, most of which had exposure history of Huanan Seafood Wholesale Market. On 20 January, the Wuhan government declared that a headquarters of epidemic prevention and control was set up. On 23 January, the headquarters announced that channels to leave Wuhan by airport or railway stations would be temporarily closed, and the residents should not leave Wuhan if not necessary (32). As of 24:00 on 9 February, there were a total of 35,982 confirmed cases in China, of which 16,902 cases were reported in Wuhan (33). On 10 February, the headquarters proclaimed that closed-off management would be carried out among all of the housing estates in Wuhan to reduce the contact of people and interrupt the transmission of COVID-19 (34). In the meantime, residents were required to stay at home and avoid visiting public places. Residential buildings which had confirmed or suspected cases of COVID-19 were strictly monitored and managed.

Although these strict measures were implemented in succession and the occupational exposure among the general population was far less than that among healthcare workers, residents inevitably experienced other forms of exposure to COVID-19 outbreak events, such as having a family member or friend who was infected. In a study based on the nationwide sample, 1.1% of the respondents had at least one family member or friend who was infected with COVID-19 (25). In an online study conducted among the general population in Hubei Province (Wuhan residents accounted for nearly half of the sample), there were 17.9% of the respondents who were familiar with someone who had COVID-19, of which three quarters revealed that a familiar relationship was friendship (23). Accordingly, we believed that this situation was not uncommon among Wuhan residents, since there were approximately half of the confirmed cases across the country reported in Wuhan at that time, suggesting a necessity to examine its psychological impact.

Thus far, little is known about the mechanism between exposure and psychological status among the general population in Wuhan, the hardest-hit city in China. In addition, exposure to COVID-19 has different levels, which are not fully addressed in the most existing literature. Given that the general population were required to stay at home and strictly keep social distancing during the Wuhan lockdown, exposure could be divided into two categories: exposure within a certain physical distance and exposure within the social network. The former refers to having a family member living together, a neighbor living in the same building, or a resident living in the same housing estate, who contracted COVID-19. The latter includes having family members not living together or people in the social network (such as friends, colleagues, and acquaintances) who contracted COVID-19. In other words, exposure at a certain physical distance means a degree of direct infection risk, while exposure in the social network indicates the possible prevalence of the disease. Because COVID-19 is highly infectious, exposure in or out of a certain physical distance might have different impacts on individuals’ psychological status at the beginning of the outbreak, when the knowledge of the virus and medical materials were extremely limited. To distinguish different levels of exposure among the general population in Wuhan and evaluate their impacts, we propose the first set of hypotheses as follows:

In addition to the objective exposure experiences, people’s attitudes, feelings, and perceptions are also closely related to their psychological outcomes when confronting different kinds of crisis events (35–37). Risk perception refers to people’s subjective judgments about the likelihood of negative occurrences, including injury, illness, disease, and death (38). There is a body of work suggesting that perceived risk is a salient factor linked to mental health outcomes among healthcare workers or the general population during some infectious disease outbreaks. A systematic and thematic review demonstrated that perceived risk was significantly linked to the psychological wellbeing among healthcare employees during the SARS crisis (39). Jalloh et al. (40) revealed that risk perception was independently associated with anxiety-depression and PTSD symptoms in the general population in an African country affected by the Ebola pandemic.

A handful of existing studies examined the relations between the perceived risk of COVID-19 and an individual’s mental health. Yıldırım et al. (41) found that perceived risk positively predicted depression, anxiety, and stress among 204 healthcare professionals who were actively treating patients confirmed with COVID-19 in Turkey. Kim et al. (42) indicated that a higher perceived risk of COVID-19 infection was associated with greater depressive symptoms during the first 6 weeks of quarantine in 221 adults in urban South Africa. Dratva et al. (43) reported that the perceived risk of COVID-19 was significantly associated with general anxiety among Swiss university students.

Similar studies were conducted among the Chinese population, all of which utilized the online questionnaire survey with a sample size ranging from 693 to 2,993 (44–46). The perceived risk of COVID-19 measured by one or more proxies, such as perceived severity, perceived controllability, perceived risk of infection, affective risk perception, or cognitive risk perception, was revealed to significantly relate to individuals’ mental health in existing studies. However, scant attention had been paid to Wuhan residents, who lived in the epicenter of China at the beginning of the pandemic. At that time, the origin of the virus, patterns of spread, therapeutic strategy, and the prognosis of the disease remained unclear. In addition, unprecedented strict measures such as lockdown and closed-off management of intra-city housing estates were implemented consecutively, and more than 30,000 medical practitioners from all over the country were sent to assist the medical system in Wuhan. How Wuhan residents perceive this pandemic in such an uncertain environment and how their perceived risk relates to psychological status have not been fully addressed in previous studies. Therefore, we develop the second hypothesis as follows:

Furthermore, although people confront the same situation during a crisis like the COVID-19 pandemic, they perceive it in different ways due to knowledge, certainty about the risk, or even individuals’ characteristics. Paek and Hove (38) gave an example in the health context, indicating that people are more likely to perceive colon cancer as a highly fatal disease if they have friends or family members who died of it. This point is echoed by a study on the 2003 SARS outbreak in Beijing, indicating that hospital employees who experienced work exposure or any quarantining perceived a greater level of risk, and the perceived risk was significantly associated with PTS symptom level (28). Such mediating effects of individuals’ subjective perception between the objective exposure experience and mental health outcome, demonstrated in the aforementioned study, were rarely examined among the existing studies on COVID-19. Having a family member, relative, friend, or neighbor who was infected with COVID-19 was not infrequent among the general population in Wuhan. Whether such exposure experience shapes residents’ perceived risk and thereby indirectly affects their mental health status needs to be further elucidated. Furthermore, people may have different perceptions of risk when they encounter different levels of exposure. Therefore, drawing lessons from the previous empirical evidence, we propose the third set of hypotheses:

To sum up, this study had three goals: to evaluate the psychological distress among the general population during the Wuhan lockdown; to explore the relations between two levels of exposure to the COVID-19 outbreak and psychological distress; and to examine the mediating effect of perceived risk between exposure and psychological distress. To the best of our knowledge, this study was one of the first attempts to investigate the psychological impacts of COVID-19 based on a relatively large Wuhan sample.

Data were retrieved from the Community Life Survey among General Population in Wuhan during the COVID-19 Outbreak, which was jointly carried out by researchers at the Zhongnan University of Economics and Law, the Hong Kong University of Science and Technology, and the Huazhong University of Science and Technology. The formal survey was conducted from 20 February to 4 March 2020, around a month after Wuhan implementing lockdown. The survey was implemented in three steps. First, we issued online recruitment notices in colleges and universities in Wuhan, setting criteria for investigators: adult and living in Wuhan at that time. More than 300 students submitted the entry form. We interviewed them and finally selected 149 investigators covering all 13 administrative districts in Wuhan, during which we considered the population size of each district. Second, a pilot study was carried out from 10 to 19 February, according to which the questionnaire was refined and finalized. Then, the investigators were strictly trained in the online meeting to ensure that they could offer the necessary assistance to their respondents.

The sampling approach of this online survey differed from most similar studies conducted simultaneously in China, that is, eligible respondents were accessed by investigators proactively. Several inclusion criteria about the respondents were set in advance: lived in Wuhan during the COVID-19 outbreak; one respondent in one household (household was defined by the independent right of property); and gender balance as much as possible. We required investigators to note the age structure of potential respondents. For instance, respondents should cover youth, middle-aged people, and old people. We did not prescribe any limits to the personal relationship between the investigator and his/her interviewees.

An online questionnaire was adopted in this survey. Each investigator was assigned a personal account so that the questionnaire that responded under this account could be easily tracked. To ensure the authenticity of the survey and the data quality, we forbade the investigators to share the questionnaire link on social network platforms (such as WeChat group and WeChat moment which were widely used in China) and required them to communicate with each potential respondent by communication tools (such as telephone and WeChat) before sending the questionnaire. Respondents were informed of the objective and the content of this survey. The website link and password of the questionnaire were sent to the respondents after obtaining their consent. Respondents were asked to directly submit the questionnaire online once completed. Every investigator was required to send out at least 15 questionnaires, but we did not propose an upper limit on the number of questionnaires for every investigator. We terminated the survey when the sample covered all districts in Wuhan, with 4,267 completed questionnaires. To examine the quality of the survey, we randomly selected some respondents and paid a return visit to them by telephone. In the final analyses, a total of 4,234 respondents were included. This study was approved by the Ethics Committee of Zhongnan University of Economics and Law.

Random iterative method (RIM) weighting was utilized to deal with the potential sampling bias. RIM weighting allows researchers to weigh each variable as an individual entity to ensure that each data point is accurately represented while keeping the characteristics proportionate as a whole (49). Given that the investigators in our online survey were students from colleges and universities, the unweighted sample was relatively young and well-educated. Weight factors were generated and applied to the sample such that the weighted sample matched the two independent distributions, that is, age and education status (50). Subsequent analyses were conducted in weighted data.

We used descriptive analyses to summarize the characteristics of the sample. T-tests and ANOVA were adopted to examine whether psychological distress was different across disparate groups in one category. Pairwise comparisons of HSCL-5 mean score between any two groups in one category were conducted using the Scheffe approach (51). Then, we performed the multivariate linear regressions to explore the relations among exposures, perceived risk, and psychological distress. Models 1 and 2 examined the effects of exposure within a close physical distance and within the social network on psychological distress, respectively, and Model 3 simultaneously adopted two kinds of exposures. Finally, we employed the mediating effect analyses to investigate the mediating effect of perceived risk between exposures and psychological distress. Sobel–Goodman tests were utilized to examine whether the mediating effects were significant (52, 53).

The current study focuses on the psychological distress among the general population in Wuhan, who lived in the earliest epicenter of the COVID-19 outbreak and experienced the unprecedented lockdown. According to an online questionnaire survey, we collected valid data from 4,234 respondents. We found that nervousness, fear, and worry were the most common symptoms among Wuhan residents, while depressed feeling was relatively less prevalent. Although confronting an extremely tough situation during the early outbreak, most of the respondents did not feel hopeless about the future.

We also investigated the exposure to the COVID-19 outbreak among the general population in Wuhan, by asking them if there were anyone physically close to them or in their social network who had contracted the virus. In our sample, nearly 70% of the respondents reported that their families who lived together and the residents who lived in the same building or the same housing estate were infected. In addition, there were approximately 40% of the respondents reported that their families who were not living together or their friends/schoolmate/colleagues were infected. Even though the exposures in our study were self-reported, these findings contributed to the current knowledge of the prevalence of COVID-19 in Wuhan from the perspective of local residents. Moreover, although the social and economic consequences of this pandemic are far-reaching, people’s perceptions of the same situation are divergent. In our sample, less than 15% of the respondents were not worried about contracting the coronavirus at all, and nearly 20% of the respondents were extremely worried about getting infected.

We performed multivariate linear regressions to explore the correlates of psychological distress among Wuhan residents. Female respondents were more likely to experience psychological distress, which was consistent with the results of previous studies with a national sample, even though we utilized different measuring tools for mental health outcomes (21, 22). Compared with the respondents who were at work, those who were out of work were more vulnerable to poor mental health, which was in accordance with a previous study (54). Students had better mental health on the 0.1 level, which differed from the finding of Wang et al. (22). In the research of Wang et al. (22), student status was significantly associated with a greater psychological impact of the outbreak and higher levels of stress, anxiety, and depression. The possible reason is that the surveys were conducted at different time points and distinct measuring tools were adopted. We found that higher family SES was a protective factor of people’s mental health status during the pandemic. Those Wuhan residents who regarded their family SES as the underclass were at higher risk of psychological distress. These findings may assist policymakers in accurately identifying people who are more susceptible to psychological disturbance and providing essential mental health services to target populations during and after the pandemic in time.

As Hypotheses 1a and 1b suggested, exposure within a close physical distance and exposure within the social network were significantly associated with psychological distress among Wuhan residents independently. This result was partly in accordance with a national online survey study involving all 34 province-level regions in China, indicating that people who had a relative with confirmed or suspected COVID-19 were more likely to have negative mental health outcomes (25). However, this study did not distinguish the disparate types of exposure and failed to compare their effects consequently.

We included two types of exposures simultaneously in one linear regression model (Model 3), finding that the standard coefficient of exposure within the social network was slightly larger than that of exposure within a close physical distance. An Individual’s social network size is usually larger than the size of his/her kinship (55, 56). To some extent, more exposure within the social network meant that the infectivity and severity of COVID-19 were great and uncontrollable, which might cause more panic and distress among residents directly. Hypotheses 2, 3a, and 3b, focusing on the direct and mediating effects of perceived risk on Wuhan residents’ psychological distress, were all confirmed by empirical data. Respondents who were exposed to COVID-19 were more likely to be worried about being contracted, and those who were more worried exhibited more distress. What is noteworthy is that perceived risk mediated more of the total effect of exposure within a close physical distance on psychological distress than that of the exposure within the social network (25.08 vs. 17.89%) in our study. This finding was consistent with the contagious nature of COVID-19, which can rapidly transmit within the family and the community (57). Infected family members who were living together or infected residents who lived in the same building or the same housing estate enabled respondents to perceive a higher risk of getting infected, and then to express more distress. The mediating effect of the perceived risk between exposure within the social network and psychological distress was relatively weaker.

We argue that demonstrating the interaction between objective exposures to the COVID-19 outbreak and subjective perception of risk on psychological distress among the general population in Wuhan is a distinct contribution of the present study. Perceived risk is associated with a greater likelihood of engagement in preventive behaviors, such as wearing the mask, avoidance of public transportation, frequent handwashing, COVID-19 testing, and vaccination (58, 59). Understanding the complicated mechanisms between inevitable exposures, psychosocial factors, and an individual’s mental wellbeing during the pandemic is beneficial to increase preparedness for the unforeseeable future outbreak and other public health crises. However, given the negative impact of perceived risk on an individual’s psychological outcome, the governmental sector and healthcare institutions have obligation to apply the appropriate health communication strategy, medium, and tool to allay the fears and maintain the perceived risk at a moderate level. Diversity of exposure level and risk perception across different social groups should be taken into consideration when designing and delivering health education and crisis intervention program. Moreover, people’s exposure and risk perception vary across different stages of the pandemic, which should be fully considered by policymakers.