Psychological Status and Associated Factors During the Lockdown Period of the COVID-19 Epidemic in China: A Web-Based Survey

Background: The aim of this study was to survey the general public in China to better understand their levels of psychological state and its inuencing factors after the Wuhan shutdown on 23 Jan. Methods: A survey was conducted on Feb 20-24, using an online self-administrated questionnaire among 4071 participants. Data on subjective indicators of daily-life change was collected, and individual scores on changes in anxiety, depression, and stress were generated by 8-item, 11-item, and 6-item questions. After bivariate analyses, multiple linear regression analyses were conducted to investigate independent associations between socio-demographic variables, subjective indicators of changes in daily life (cid:0) and summary scores including anxiety, depression, and stress scores. Results: Information from 3803 participants was available for analysis. Multivariable regression analyses showed that the anxiety (B=-1.27, 95%CI=-1.71 to -0.82), depression (B=-1.47, 95%CI=-2.06 to -0.88), and stress (B=-0.79, 95%CI=-1.13 to -0.46) scores of people in rural areas are lower than those in urban areas. People living in the other regions except Hubei, higher education were independent correlates of less negative emotions, while people with relatively high incomes had poor psychological status in anxiety (B=0.73, 95%CI=0.08 to 1.38),depression (B=1.45, 95%CI=0.60 to 2.30) and stress (B=0.65, 95%CI=0.17 to 1.13). Married people were less anxious (B=-0.67, 95%CI=-1.30 to -0.05), depressed (B=-1.14, 95%CI=-1.96 to -0.33), and stressed (B=-0.47, 95%CI=-0.93 to 0.00) than single people. The level of attention, self-assessed infection risk, impact of the daily life and mental-health help-seeking tended to be positively associated with the scores of anxiety, depression, and stress (p<0.001). Conclusions: Usual residence, education,


Background
Being highly contagious, COVID-19 has arisen a national pandemic and a rapid transmission globally to a lot of countries worldwide [1]. In the wake of this global health crisis, stringent public health measures have been implemented to curtail the spread of COVID- 19. In China, in order to prevent the further spread, a lockdown was imposed on Wuhan on 23 January, with travel restrictions, followed by the entire Hubei Province a day later. In addition, the government has taken some measures to prevent further dispersal, including closing entertainment venues, cancelling the party, extending the Chinese New Year holidays, forcing people to wear masks in public as well as limiting the number and frequency of outings per household. The Chinese experience actively treats infected patients, protects susceptible populations, and cuts off transmission routes, which are proved to be effective and stop at least 700,000 cases of COVID-19 [2].
However, the outbreak itself and the measures taken to combat the epidemic could lead to widespread fear and panic, which may escalate into further negative psychological reactions including adjustment disorder and depression. With the closure of schools and business close, the negative emotions experienced by individuals become more complicated [3]. At the same time, as most residents are restricted to their homes, they tend to face too much negative news at home every day, which may lead to a psychological crisis.
Previous evidence showed that quarantine and isolation of patients led to widespread fear and panic, resulting in negative psychological reactions including adjustment disorder and depression [4][5][6][7][8]. And one recent study pointed to an increase in psychological problems in this epidemic, including anxiety, depression, and stress [9]. Two studies had noted an increase of psychological problems during the epidemic, and emphasized that we should pay attention to the mental health of speci c groups such as children, old adults, patients, and medical staff [10,11]. However, no data is available examining the psychological impact of COVID-19 on the general population in China one month after the Wuhan shut down.
Therefore, we conducted the survey to investigate the residents' change of life and psychological conditions in one month after the Wuhan lockdown. This may assist government agencies and healthcare professionals in safeguarding the psychological wellbeing of the community in the face of COVID-19 outbreak expansion in China and different parts of the world.

Study participants
We conducted an online survey in one month (Feb 20 to 24) later after the shutdown of Wuhan (Jan 23) and Hubei province (Jan 25) against the COVID-19 spread, and we received 4071 anonymous questionnaires in the investigation, covering 33 Chinese provinces and autonomous regions except Taiwan.
Inclusion Criteria:1. Male or female ages 15-85 years; 2. Participants must have capacity to understand the study and provide informed consent; 3. Participants must be uent in Chinese; 4. Participants currently live in China.
Exclusion Criteria: 1. Serious neurological (speci c or focal) disorders preventing full participation in the protocol; 2. The illogical case in the questionnaire. Examples include: Select the same option consecutively; The results of similar choices vary widely.
After eliminating the invalid samples, 3803 (93.42%) valid questionnaires were nally obtained. This study was approved by the research ethics committees of Wuhan University. All participants provided informed consent. Based on the investigation of the psychological state after the disaster in China and compiled after the discussion of experts, the self-administrated questionnaires were mainly divided into three parts.

Changes in psychological status
Changes of psychological status: In the study, 21 feeling items were used to measure the changes of psychologic status, including sorrow, fear, tired, irritability, loneliness, sleep condition, self-perceived uselessness, irritability and loneliness, weight, appetite, chest tightness, disturbed, muscle ache and others (see details in the Supplementary Materials). We rated these items in a 5-point response format. ("-2 = signi cantly decrease", "-1 = decrease", "0 = unchanged", "1 = increase", or "2 = signi cantly increase"). Sorting the items and calculating the scores can analyze the changes in the residents' psychological status in a more targeted manner. Therefore, literature review and experts interview methods were used to construct the index system rstly. According to the literatures [12][13][14], the 21 feeling items in the self-made questionnaires were classi ed into three categories: anxiety, depression, and stress. The total scores were calculated by simple addition based on the extent of the feeling. A negative score indicated that the negative emotions of the participants decreased compared to the previous week; otherwise, a positive score indicated that the negative emotions increased. The higher the score, the worse the psychological condition. An additional movie le shows the questionnaire in more detail [see Additional le 1]. The reliability of the questionnaire checked using Cronbach's Alpha and reliability coe cient was 0.958.

Subjective indicators of changes in daily life
The status of daily life of residents after the Wuhan shutdown is composed of level of attention, self-assessed infection risk, impact of the daily life, self-perceived health status, mental health help-seeking and satisfaction with community work. The rst 3 items were rated as "1 = decreased", "2 = unchanged", "3 = increased"; while the self-perceived health status was rated as " 1 = good/very good", "2 = average ", "3 = bad/very bad", mental health help-seeking was related as "1 = found and tried", "2 = found but not tried", "3 = not found yet", "4 = not looked for", "5 = no need to adjust", and the satisfaction with community work was rated as "1 = satis ed", "2 = general", "3 = unsatis ed".

Statistical analysis
Data were double-entered and cross-checked using Excel version 2019 (Microsoft Corp.; Redmond, USA), R3.6.2 was used for data cleaning, SPSS 25.0 was used to conduct corresponding statistical analysis, and a two-sided p value less that 0.05 was considered statistically signi cant.
To identify the determinants of participants' psychological feelings, we rst examined the effects of their characteristics on changes of anxiety, depression, and stress scores with one-way analysis of variance (ANOVA) or the nonparametric Kruskall-Wallis test for categorical variables, depending on the distribution of the variables. The statistically signi cant variables were then allowed to enter the multiple linear regression model, and dummy variables were created when appropriate. Dichotomous variables were explored for the regression analysis in order to simplify the relationships. A series of multiple linear regression analyses (stepwise method) were explored to investigate the independent associations between socio-demographics, subjective indicators of changes in daily life and summary scores including anxiety, depression, and stress scores after checking the assumptions of distribution and independence of the residuals as well as multicollinearity. Normality was assessed by visual inspection of the P-P plot. Linearity and homoscedasticity was investigated by visual inspection of the plot of the predicted values and standardized residuals. A variance in ation factor (VIF) of greater than 10 was used to identify possible multicollinearity among independent variables.  Univariate analysis showed that participants with different place of living, usual residence, monthly income, and whether there were diagnosed patients in the relationship network had signi cant differences in the scores of anxiety, depression and stress (Table 1). In addition, the differences in the six indicators of self-perception factors, that is, level of attention, self-assessed infection risk, impact of the daily life, self-perceived health status, mental health help-seeking and satisfaction with community work, also had signi cant differences in the scores of the three psychological conditions.

Results
The assumptions for linear regression were met for our data. Linearity, homoscedasticity and normal distribution of residuals were validated in the models. The VIFs were less than 10, indicating multicollinearity was not observed in the models.
Multivariable analyses were then performed to identify these variables with a signi cantly independent impact on the changes in psychological status ( Table 2). The scores of anxiety, depression, and stress were dependent variables and independent variables of the models were age (≥ 50 as reference), gender (male as reference), place of living (urban as reference), usual residence (other areas in China as reference), education (middle school or below as reference), marital status (single as reference), occupation (medical staff as reference), monthly income (< 2,000 yuan as reference), number of cohabitants (0 as reference), quarantine or not (yes as reference), con rmed infected in personal network (yes as reference), level of attention (increased as reference), self-assessed infection risk (increased as reference), mental health help-seeking (not found yet as reference) and satisfaction with community work (dissatis ed as reference).  Compared to people with middle school education and below, the mental health status of people with high education is better. Married persons had lower anxiety, depression, and stress scores than single persons, that is, their mental health was relatively better. People with monthly incomes above 10,000 have higher anxiety, depression, and stress scores than those with monthly incomes below 2,000. The scores of anxiety and stress increased as the frequency of attention to the epidemic increased. Anxiety, depression, and stress were higher among those who thought they had an increased risk of contracting COVID-19 in recent days and those who believed that the epidemic's impact on their lives had increased. Taking people who had sought mental health help but not yet as a reference, people who had tried to adjust their mental state in some way and those who thought they did not need to adjust their mental state had lower scores of anxiety, depression, and stress. In addition, people who had received psychological help but had not yet tried had lower anxiety and stress scores. Gender, age, occupation, con rmed infected in personal network and satisfaction with community work appeared not to be signi cant correlates of the anxiety, depression, stress scores.

Disccusion
Since COVID-19 outbreak sparked a global public health crisis by spreading across China and other countries, various mandatory precautions have been taken by governments and individuals [15]. To the best of our knowledge, the present investigation is the rst study to characterize people's psychological status in anxiety, depression and stress one month after the Wuhan lockdown, when the government's initiatives have achieved initial success. Studying the residents' psychological conditions at this time point can re ect the effectiveness of the anti-epidemic work of the government, communities and all walks of life from the side.
In multivariable analyses, we found that urban residents were more likely to gain anxiety, depression and stress than rural counterparts. In densely populated urban areas, well-planned e cient public transportation systems can facilitate residents' travel [16]. The disruption of daily life and absence of entertainment or recreation made it impossible to release the excess inner pressure of urban residents. A more important reason is that due to the high density of urban population and the greater mobility of people than rural areas, the risk of disease infection is greater. High population density increases people's exposure to infectious diseases [17], which may lead to increased negative emotions among urban residents. People living in Hubei province reported signi cantly higher anxiety and stress scores, perhaps because of Hubei in the center of the epidemic. Higher educated participants possibly have a better understanding of the epidemic and look for appropriate care for their condition, which might lead to lower levels of negative emotions and better coping strategies. People who were married reported greater mental health status, this might partly be attributable to the fact that married people can share the burden of negative emotions and obtain psychological support from their families [18,19], indicating that family support is of great importance. It is well documented that low-income groups are more likely to suffer from depression and anxiety [20,21]. However, our study showed that the epidemic had a greater impact on high-income groups, people with monthly incomes above 10,000 have higher anxiety, depression, and stress scores than those with monthly incomes below 2,000, their concerns about delays in working hours and subsequent deprivation of expected income may explain the high level of stress [22].
Another noteworthy nding of this study is that the subjective indicators of changes in daily life played an important role in the scores of people's anxiety, depression and stress. Our study found the greater the level of attention to COVID-19, the greater the negative emotions, which is in agreement with previous research [23]. In addition to the Wuhan lockdown, relevant actions including the urgent establishment of two quarantine hospitals (Huoshenshan Hospital and Leishenshan Hospital) within a 10-day span in response to the outbreak [24,25]. Other protective measures have been enacted such as building a series of cabin hospital to receive people who have tested positive for the coronavirus but show no severe symptoms [26].Our investigation was carried out in mid-February, and there was no sign of improvement at the time. During this period, domestic and foreign media rush to report on the incident. Since people cannot differentiate true and false news, the more attention to COVID-19, the more unclear information may be received, which negatively affects respondents' psychological status. Therefore, the content of health information provided during the epidemic needs to be based on evidence to avoid adverse psychological reactions.
Our ndings also revealed that the level of self-assessed infection risk also in uenced participants' mental state. Anxiety, depression and stress outcomes were elevated with the increase of self-assessed infection risk. It may be resulted from actual conditions, respondents received signals from the surrounding environment and is supposed to make corresponding assessment of their risk of infection. The respondents who felt severely affected by the lockdown exhibited more obvious anxiety, depression, and stress than the rest of them. The respondents who were seriously affected by the quarantine exhibited more obvious anxiety, depression and stress than the rest of them. This gives an indication that guarantee day to day lives for the residents will be bene cial for mental health [27].
In addition, people who had tried to adjust their mental state in some way and those who thought they did not need to adjust their mental state had lower scores of anxiety, depression, and stress. And people who had received psychological help but had not yet tried had lower anxiety and stress scores. It re ected from the side that when one nds a problem with one's mental state, actively seeking a solution can effectively relieve negative emotions. During the COVID-19 epidemic, online mental health services have become the mainstream way of mental health services, including online cognitive behavioral therapy for depression, anxiety, and insomnia (e.g., on WeChat) [28]. So, for people with psychological problems, it is also a good choice to seek help from professionals on the Internet.
This study not only a supplementary of the psychological status of residents during the lockdown period, but also helps to better understand which group of people are more likely to produce negative emotions when the disease is epidemic, which makes a lot of sense to China and other countries.

Limitations
There are some limitations to the study. On the one hand, during the process of data collection, sources of bias include potential selection bias of respondents, as respondents were asked if they were willing to participate in the survey, resulting in volunteer bias and may not be truly representative of the general population. On the other hand, although we have su cient respondents, the sampling method may have non-response bias by two surveys [29].

Conclusion
In conclusion, the life and psychological state of the urban population had produced negative changes after the Wuhan shutdown in 23 Jan. Usual residence, education, marital status, monthly income, the level of attention, self-assessed infection risk, impact of the daily life and mental-health help-seeking are important correlates of the scores of anxiety, depression, and stress. At present, China has achieved great success in the ght against epidemics, but the epidemic situation in some parts of the world has not improved. Most residents are still in a state of quarantine at home. Awareness of these relevant factors could help the government and related personnel to prevent more severe psychological trauma in the later period.

Declarations
Ethics approval and consent to participate The project was conducted through an online survey and we obtained the informed consent of the participants before doing the survey. Due to the epidemic situation, we cannot return to school for the time being. The statement of the Ethics Committee will be added after returning to school.