A cross-sectional survey with convenience sampling was conducted in mainland China from February 11 to 16, 2020. A web link and quick response code were distributed via WeChat, a widely used social network platform in China, to collect self-administrated questionnaires with support from the Environmental Health Institute at China Medical University. Participants anonymously completed a questionnaire in 15 to 20 minutes comprising questions based on the Patient Health Questionnaire-9 (PHQ-9) and concerning psychological and behavioral responses to the COVID-19 pandemic. Participants included in this study met the following inclusion criteria: aged ≥18 years; able to read and write Chinese; able to use WeChat to complete the questionnaire independently; willing to participate and provide signed web-based informed consent. Any participant meeting one or more of the following criteria was excluded: receiving treatment for any psychological illness; having significant visual impairment; having a history of drug dependence; having been diagnosed with a disease that would prevent them from completing the questionnaire independently.

This survey recruited a total of 1675 adults from the general public in 3 municipalities and 22 provinces or autonomous regions in China. Each participant was well informed of the aims, funding, and contents of the questionnaire and the commitment to the privacy of the participants. Questionnaires that were answered completely and logically were regarded as valid. A total of 1342/1675 participants provided valid answers to the questionnaire, resulting in a valid response rate of 80.12%. This study conformed to ethical standards and was conducted in accordance with the Helsinki Declaration as revised in 1989. The Ethics Committee of China Medical University approved the protocols of this study.

The demographic characteristics of the participants that were collected included gender (male, female), age (≤35 years, >35 years), marital status (married, other), occupation, education, and monthly income. Age was categorized according to proportionality, population distribution, and previous studies on the relationship between age and use of WeChat [49,50]. Occupation was categorized as government worker/civil servant/village committee worker/enterprise employee, health care worker, student, teacher/lawyer/journalist, and other. Education was categorized as below junior college, junior college, and bachelor’s degree and above. Monthly income (RMB) was classified as ≤¥5000 (≤US $725.19), ¥5001 to ¥10,000 (US $725.34 to $1,450.39), and >¥10,000 (>US $1,450.39).

Depression was assessed using the 9-item PHQ-9, which is one of the most widely used tools to assess depression [51]. Each item asked about the situation in the past two weeks using a 4-point Likert-type scale with choices of “not at all,” “a few days,” “more than half of the two weeks,” and “nearly every day,” giving a total score between 0 and 21. A score above 10 is regarded to be indicative of depression [52-55]. The Cronbach α coefficient of the PHQ-9 in this study was .91.

Psychological responses were measured by self-developed questions. Psychological responses included “feeling stressed,” “feeling helpless,” “persistently being worried even with support,” “being calm,” and “being optimistic.” Answers were grouped into three categories of “disagree,” “not sure,” and “agree” to assess the participants’ psychological responses during the past two weeks.

Behavioral responses were measured by self-developed questions. Options for reflecting behavioral responses for this study included “never feeling clean after disinfecting; scrubbing hands and items repeatedly” (answers were “no,” “sometimes,” and “always”), “hoarding food, medicine, or daily supplies,” “social avoidance to avoid infection,” and “being distracted from work or study.” Answers were grouped into 3 categories of “disagree,” “not sure,” and “agree” to evaluate the participants’ behavioral responses during the past two weeks.

Social support was assessed by a “yes or no” question that asked whether people had received social support in the past two weeks.

All analyses were performed using SPSS version 23.0 for Windows (IBM Corporation). A two-tailed probability value <.05 was considered statistically significant. We used t tests and one-way analysis of variance to compare differences in depression among categorical variables. Hierarchical multiple regression (HMR) analysis was conducted to test incremental variance using the following independent variables: Step 1: demographic characteristics of the general public; Step 2: psychological responses to the COVID-19 pandemic; Step 3: behavioral responses to the COVID-19 pandemic; and Step 4: social support. The depression scores were continuous in HMR and used as the dependent variables. Standardized parameter estimates (standardized β) were used to compare the magnitude of associations of the independent variables. The depression scores in the multivariate logistic regression analysis were binary (depression or no depression) with a cutoff score of 10. Multivariate logistic regression analysis was performed to explore risk factors associated with depression using odds ratios (ORs). Among all categorically independent variables, items for which more than 95% of individuals had the same response were not included in the data analysis. A two-tailed P value <.05 was considered to be statistically significant.

A total of 1342 subjects participated in this study, giving a valid response rate of 1342/1675 (80.12%). The distribution of the demographic characteristics of the participants, prevalence of depression, and univariate analysis of depression are listed in Table 1. The prevalence of depression among the general public was 182/1342 (13.60%).

The 1342 participants included 500 men (37.26%) and 842 women (62.74%). Approximately one-third of the participants (467/1342, 35.77%) had a bachelor’s degree or higher degree. Depression scores for the participants who were aged <35 years were significantly higher than those of the other participants (P=.001). Participants who were married tended to have lower depression scores than participants who were not (P<.001). Government workers, civil servants, village committee workers, enterprise employees, and health care workers had lower depression scores than those engaging in other occupations (P<.001).

The distribution of depression scores according to psychological and behavioral responses to the COVID-19 pandemic are shown in Table 1. The participants who disagreed that they felt stressed or helpless tended to have lower depression scores than those who agreed or were not sure (P<.001). Depression scores were lower among participants who were persistently worried even with support (P<.001). The depression scores for the individuals who agreed that they were calm or optimistic about the COVID-19 pandemic were significantly lower (P<.001).

The participants who never felt clean after disinfecting and who scrubbed their hands and items repeatedly had significantly higher depression scores (P<.001). Participants who agreed that they hoarded food, medicine, or daily supplies tended to have significantly higher depression scores (P<.001). Scores were higher for the participants who agreed that they were distracted from work or study (P<.001).

The results also showed that participants who reported that they did not receive any support from others when they had difficulties during the COVID-19 pandemic had a significantly higher score of depression than those who received social support (P<.001).

Table 2 shows the final results of the HMR models of depression among the general public. A total of 42.1% of the variance was explained by the final model. The R² changes indicated that the incremental variances explained by each block of variables were 3.1%, 28.6%, 10.1%, and 0.3% for demographic characteristics, psychological responses, behavioral responses, and social support, respectively. In the final model of the HMR and the forest plot (Figure 1), the risk factors of depression included feeling stressed; feeling helpless; persistently being worried, even with support; never feeling clean after disinfecting, and scrubbing hands and items repeatedly; hoarding food, medicine, or daily supplies; and being distracted from work or study (all P<.01). However, being calm and social support were negatively associated with depression.

Table 3 presents the results of the multivariate logistic regression analysis. Feeling helpless (OR 2.341, 95% CI 1.367-4.009), persistently being worried even with support (OR 3.315, 95% CI 1.696-6.479), never feeling clean after disinfecting and scrubbing hands and items repeatedly (OR 1.941, 95% CI 1.153-3.267), hoarding food, medicine, or daily supplies (OR 1.822, 95% CI 1.012-3.279), and being distracted from work or study (OR 27.225, 95% CI 8.243-89.918) increased the risk of depression. However, being calm (OR 0.344, 95% CI 0.186-0.635) and social support (OR 0.529, 95% CI 0.308-0.908) decreased the chance of suffering from depression. Additionally, compared with other occupations, teachers, lawyers, and journalists (OR 2.053, 95% CI 1.117-3.776) had higher chances of depression. In the final model of the multivariate logistic regression and the forest plot (Figure 2), the risk factors of depression included occupation of teacher, lawyer, or journalist; junior college education; feeling helpless; persistently being worried, even with support; never feeling clean after disinfecting and scrubbing hands and items repeatedly; hoarding food, medicine, or daily supplies; and being distracted from work or study (all P<.01). However, being calm and social support were protective factors of depression.

This study revealed that 13.6% of the general public suffered from depression during the early stages of the pandemic of COVID-19 in China. The prevalence of depression was slightly lower than the prevalence before the outbreak of COVID-19 from previous studies [56-58] (13.8% vs 18.8%). The results also indicated that the prevalence of depression during the COVID-19 pandemic is almost identical to that during the SARS epidemic (11.9% vs 18.0%) among the general public in China [59,60]. Most of the research on mental disorders during public health emergencies in China focuses on the psychological status of patients or health care workers; depression in the general public is neglected and untreated [61-64]. The general public are extremely vulnerable to symptoms of depression during the COVID-19 pandemic. People with mental symptoms of depression should be identified and managed in a timely manner to improve their psychological health status [65].

Depression is closely associated with the psychological and behavioral responses of the general public during the COVID-19 pandemic. This study found that psychological responses contributed the most (28.6%) to the variance of depression. Feeling stressed, feeling helpless, and feeling persistently worried even with support were positively associated with depression, which is in agreement with previous studies indicating that the increase of perceived susceptibility to the epidemic is associated with the prevalence of mental disorders [66]. Perception of the risks of infection is a possible reason for negative psychological responses to pandemics and is highly correlated with distress in individuals [67-69]. Additionally, poor self-rated health status can be causative and can result in psychological burdens and excessive worries about the pandemic [70,71]. In addition, individuals who are worried even when they have support may be more prone to behavior problems, which can exacerbate the symptoms of depression [72,73].

Notably, being calm and optimistic are negatively associated with depression. Calmness and optimism can help people cope with problems rationally and may mediate the negative effects of perception of epidemic-related stress and indirectly attenuate the symptoms of depression [74]. On the other hand, positive cognition of public health emergencies has close correlations with stress and social functions [75]. Additionally, positive psychological capabilities such as calmness and optimism can ameliorate depression, in particular for people who experience higher levels of helplessness [76]. Lessened psychological capabilities can result in high susceptibility to anxiety and depression during negative emotional experiences or periods of adversity such as the COVID-19 pandemic. Thus, psychological management and regulation of the responses to a public health emergency may benefit from consciously controlling stress, helplessness, and worry and from maintaining calm and optimism.

This study showed that depression was significantly linked to behavioral responses. More than half of the participants (799/1342, 59.54%) reported that they never felt clean after they disinfected and scrubbed hands and items repeatedly, and 1067/1342 (79.51%) of these members of the general public practiced social avoidance to avoid the infection of COVID-19. Preventive behaviors are key to reducing the spread and impact of a pandemic. However, excessively preventive and avoidant behavioral responses, which possibly resulted from the perceived stress induced by the COVID-19 pandemic, were positively associated with depression. Recent studies have reported that psychological impacts, including depression, anxiety, and stress, are linked with the adoption of precautionary measures to prevent the spread of COVID-19 [16]. For example, using medical preventive equipment and washing after touching contaminated surfaces can provide potential psychological benefits by offering a sense of security and comport for the general public and can prevent the spread of COVID-19. However, these precautionary measures can also have negative effects on individuals’ mental health when they become excessive [44,77]. People may feel more susceptible to becoming infected during pandemics. Negative psychological impacts revealed by these behaviors can be easily popularized among the general public. This finding is consistent with other studies showing that protective behaviors such as avoiding sharing of utensils during meals and washing hands frequently increase the likelihood of stress, anxiety, and depression [78]. Health professionals should pay more attention to these behaviors, which can have adverse impacts on mental health, with a view to realizing early warning signs and conducting psychological interventions to protect vulnerable groups [79].

Specifically, social support appeared to be a protective factor for depression. Social support can enhance self-esteem and positive adaptation to stress and combat the detrimental effects of stress or worry on the development of depression [44]. Social support such as positive communications and entertainment may help individuals to recover quickly the from the COVID-19 outbreak [80]. In contrast, lack of social support can result in hostility and uncertainty in social life, thus increasing susceptibility to serious mental disorders, particularly depression [16]. Therefore, adequate and appropriate social support should be provided to prevent the development of mental disorders in the general public.

Several limitations should be acknowledged when presenting our findings. First, this survey was conducted with convenience sampling, which limits our ability to generalize the results of this study to the overall population. Second, this study was based on a self-administrated questionnaire completed via smartphone, which limited the diversity of the sample and the authenticity of the answers. Depression and the associated factors revealed in this study were limited to the early stages of the COVID-19 pandemic.

The general public in China suffered from high levels of depression during the early stages of the COVID-19 pandemic. Being stressed, feeling helpless or worried, and being distracted from work or study increased the risks of depression. People who experience these adverse psychological responses should be provided with psychological interventions to reduce the symptoms of depression. Moreover, positive emotions and psychological responses such as calmness and optimism decreased the risks of depression and should be encouraged by health professionals and educators. Behavioral responses, including repeated cleaning, hoarding supplies, and distraction from work or study, were not only influencing factors of depression but also expressions of depression. Therefore, early detection and psychological and behavioral interventions should be developed to help people cope rationally with pandemic-related stress and improve the mental health of the general public.