![ORCID Icon]({"type":"image","src":"/templates/jsp/_style2/_tandf/images/orcid.svg"} "ORCID Icon"){kind=small}
https://orcid.org/0000-0001-5070-2922
Abstract
?
![MathJax Logo]({"type":"image","src":"/templates/jsp/_style2/_tandf/pb2/images/math-jax.gif"})
?Abstract
Objective
The outbreak of the coronavirus disease (COVID-19) poses an unprecedented threat to public health. Current measures to control the spread include social distancing and quarantine, which may trigger mental health problems.
Design and Main Outcome Measures
The sample (N = 1160) constituted three groups: people quarantined in an affected area, unaffected areas, and people not in quarantine. The Center for Epidemiological Studies Depression Scale (CES-D-20) and the Goldberg Depression and Anxiety Scale (GAD-7) were administered as measures of depression and anxiety, respectively. The multi-variant logistic and multiple linear regression identified factors associated with depression and anxiety.
Results
Probable depression and anxiety were reported by 26.47% and 70.78% of all respondents, respectively. After adjusting for demographic and community variables, quarantined respondents reported a higher likelihood to exhibit symptoms of depression and anxiety than those not quarantined. Respondents living in communities where screening for COVID-19 was required were less likely to report depression and anxiety symptoms.
Conclusion
The incidence of depression and anxiety among quarantined respondents was significantly higher than that of respondents not quarantined, and twice as common among quarantined respondents in unaffected areas as those in affected areas. Appropriate community screening may reduce the risk of depression and anxiety during an epidemic.
Introduction
The outbreak of the coronavirus disease (COVID-19) in Wuhan has received global attention. COVID-19, a highly transmissible disease, has been reported to cause respiratory problems (Lu et al., 2020). The most common way to curb the spread of the virus is community-wide containment involving isolating and restricting the movement of people who have been potentially exposed to the virus to ascertain if they develop symptoms, thus reducing their risk of infecting others (Centers for Disease Control & Prevention, 2020; Hellewell et al., 2020). Quarantine was imposed in China on 23 January 2020, and the Chinese government adopted rigorous emergency response measures, including the lockdown of cities, travel bans, and concentrated quarantine. Several thousands of people returning home from affected areas were instructed to self-isolate at home or in a dedicated quarantine facility (China Daily, 2020a). To some extent, these measures have helped in controlling the outbreak, but at the same time created an unpleasant experience. Separation from family, and uncertainty regarding contracting the disease makes quarantine a depressive and anxious experience. In particular, the social exclusion of confirmed patients, suspected cases, and their close contacts typically lead to public anxiety and panic (Barbisch et al., 2015).
There are two forms of quarantine situations, namely quarantine in affected areas and quarantine in unaffected areas. People quarantined in the former have to bear the risk of both infection and isolation, which may pose a huge burden on their psychological wellbeing. Fortunately, national management has unceasingly provided support to affected areas in the form of medical and living supplies and moral support (China Daily, 2020a). However, an unfavourable phenomenon emerged in the unaffected areas. An official report stated that nearly 5 million people travelled from the affected areas due to the Spring Festival and the outbreak (China Daily, 2020b). Since the estimated incubation period of the coronavirus disease is 1–14 days, potential carriers may infect others without exhibiting symptoms. This invisible contagion caused public worry and panic (World Health Organization, 2020). Moreover, due to the misinformation spread by the media and strict prevention and control measures in communities, these 5 million residents became the target of rejection and abuse for people in other areas overnight. Labels, such as ‘deserter’ or ‘all infected with the novel coronavirus’, may lead to stigmatization, discrimination, and other mental health challenges (Bao et al., 2020).
Previous studies investigated the impact of quarantine during the Ebola, SARS, and MERS on mental health (Bai et al., 2004; Barbisch et al., 2015; Hawryluck et al., 2004; Joo et al., 2019; Shultz et al., 2015). People in quarantine were significantly more likely to report exhaustion, anxiety, irritability, insomnia, and indecisiveness (Blendon et al., 2004; Samantha et al., 2020). Healthcare workers who were quarantined or working in a high-risk area reported significant symptoms of alcohol abuse or dependency after the SARS outbreak (Bai et al., 2004). We found one study compared the mental health of quarantined undergraduates from those not quarantined, and found no significant difference in mental health (Wang et al., 2011). Another study showed that 66% of those quarantined for close contact with persons at risk of SARS reported various negative reactions, with only few reporting positive feelings (Reynolds et al., 2008). Healthcare workers working in quarantine centres were more likely to experience acute stress symptoms post quarantine (Liu et al., 2012). Moreover, several studies suggest that participant characteristics and socio-demographics predict the psychological impact of quarantine (Hawryluck et al., 2004; Taylor et al., 2008). Similarly, a study reported that younger age, lower educational qualifications, and female gender were associated with negative psychological states (Hawryluck et al., 2004). Conversely, Taylor et al. (2008) noted that demographic factors such as marital status, age, educational qualifications, living with family, and the number of children in a family were not associated with mental illness. Further, a rapid review of the psychological impact of quarantine identified common stressors during quarantine, including duration of quarantine, fear of infection, inadequate supplies, and inadequate information (Samantha et al., 2020).
Overall, most quantitative studies analysed data regarding quarantined respondents and reported a high prevalence of negative psychological symptoms. Nevertheless, the mental health of people quarantined in unaffected areas is ignored. In addition, stigmatization and discrimination during community surveillance, especially after carpet check and being reported by relatives, and subsequent social exclusion of suspected and confirmed cases may lead to public anxiety and panic. In addition to the influence of demographic characteristics on the mental health of those in quarantine, the capacity of community screening to control disease outbreaks should be taken into consideration.
At the time of study, thirteen days after the lockdown on January 23, 2020, the outbreak is entering its the rapid growth period (Figure 1), the number of new infections was rapidly increasing. Many provinces have taken a series of prevention and control measures and activated the highest-level emergency response. Thus, our study was conducted during 5 February to 7 February, 2020, we explored people quarantined in affected areas, unaffected areas, and people not in quarantine for symptoms of anxiety and depression related to the COVID-19 outbreak and the relative risk factors involving participant characteristics and sociodemographic characteristics. We also assessed subsequent community measures in the early weeks of the outbreak.
Published online:
22 June 2020Figure. 1. Trajectory of the pandemic in China with the number of deaths and confirmed cases.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/gpsh20/2021/gpsh20.v036.i02/08870446.2020.1782410/20210501/images/medium/gpsh_a_1782410_f0001_c.jpg"})
Figure. 1. Trajectory of the pandemic in China with the number of deaths and confirmed cases.
Methods
Study population
During the first two weeks of the lockdown of Wuhan, we conducted an internet-based cross-sectional survey to examine the potential psychosocial crisis associated with mass-quarantine. We investigated three groups constituting people quarantined in the affected area, unaffected areas, and people who were not quarantined.
People quarantined in the affected area
Wuhan was the first city to be hit by the outbreak in the beginning of 2020 and the government, subsequently, announced a complete lockdown of the city on 23 January 2020. Thus, people living in Wuhan since the lockdown were defined as people quarantined in the affected area.
People quarantined in unaffected areas
Unaffected areas include all areas except Wuhan. People quarantined in unaffected areas constituted people with a travel history to and/or from Wuhan who were instructed to quarantine.
People not in quarantine
This group constitutes people who had not visited the affected area from 1 December 2019 to 25 January 2020. Although cities in China were placed under mass quarantine, people who had not visited an affected area were not required to undergo screening or quarantine.
Participant recruitment
In this urgent circumstance, most of provinces have taken the joint prevention (further travel bans and some executive orders to quarantine suspected cases) and control measures to reduce the risk of infection in China, field investigation was not possible and not allowed. Moreover, people quarantined in the affected and unaffected areas which could be referred to as ‘hidden populations’ because of the quarantine and the related stigma. Snowball sampling, where existing study subjects recruit further subjects from among their acquaintances, is a popular approach when the objective is sampling hidden populations (Salganik & Heckathorn, 2004). Thus, we used a combination of stratified and snowball sampling methods, with financial incentives. We instructed 10 initial respondents, called ‘seeds’, to list their social contacts. The initial respondents were stratified according to their geographical location. To prevent selection bias, the second and third seeds were generated to reproduce target population variability and links to an online survey were distributed through individual friendship networks. Each new participant recruited similar individuals, people quarantined in affected areas and unaffected areas, from their social networks in a multistage process. A web page (http://www.survey.163.com) was designed to provide detailed information (questionnaire) on the survey and to checking IP (Internet Protocol) address to avoid re-sampling. We finally applied effect size to check sample representativeness and weighted the sample data to reduce the evident selection bias (see Appendix, Table A1), which shows that the sample used for analysis was slightly younger and relatively better educated comparable to a recognized random sample in China. Post-stratification weighting was then applied using raking (Lumley, 2013). The Institutional Review Board of Wuhan University approved the study and a written informed consent was obtained from all participants. Eligible participants were 18 or older.
A total of 1389 participants from 31 provinces of mainland China and Hong Kong responded to the questionnaires, of which 1160 questionnaires were complete and correctly filled. We removed nonconforming questionnaires (answer time less than 200 seconds, IP address is the same for multiple times). A statistical power of 0.93 with a two-tailed test was obtained in this study. Of these, 172 were quarantined in the affected area, 219 were quarantined in an unaffected area, and 769 were not quarantined. The socio-demographic profile of the respondents quarantined in the affected area revealed that 71 (41.28%) respondents were aged 18–29 years, 65(37.79%) were males, and 86 (50%) had senior high school level qualifications, while 49 (28.49%) had college and above qualifications. Of 172 respondents, 76 (44.19%) had an income of over 5000 Ren Min Bi (RMB) per month and 52 (30.23%) were 2000–4999 RMB per month, and 71(41.28%) were married. The socio-demographic profiles of respondents quarantined in unaffected areas revealed that 138 respondents (63.01%) were female, with 140 (63.93%) aged 18–29 years, and 52 (23.74%) aged 30–49 years. Of 219 respondents, 124 (56.62%) had senior high school level qualifications, 39 (17.81%) had college and above qualifications, while 78 (35.62%) were 2000–4999 RMB per month, 74 (33.79%) earned over 5000 RMB per month, and 138 (63.01%) were married.
Psychological outcomes
We focused on depressive symptoms and anxiety, which are most commonly observed mental health outcomes during natural adversities (Steel et al., 2009). Depression symptomatology was assessed using the 20-item Centre for Epidemiological Studies Depression Scale (CES-D-20) (Andresen et al., 1994). The CES-D-20 is a 20-item self-report questionnaire designed to measure the frequency of depressive symptoms in the past 2 weeks. Items assess four dimensions: depressed mood (8 items), positive mood (4 items), physical symptoms (6 items) and interpersonal relationships (2 items). The total score of 20 items ranged from 0 to 60, and a score of ≥15 indicates depressive symptoms (0= normal, 1= depression). Item responses are weighted based on symptom frequency and range from 0 to 3, with 0 reflecting little or no (<1 day), 1 reflecting not much (1–2 day), 2 reflecting sometimes or half the time (3–4 day) and 3 reflecting most of the time (5–7day). Items about positive mood (the fourth, eighth, twentieth and sixteenth items) are scored in reverse. CES-D-20 has demonstrated a good reliability (Cronbach’s α = 0.87).
Anxiety symptoms were assessed using the 7-item Generalized Anxiety Disorder Scale (GAD-7) (Lowe et al., 2008). The GAD-7 asks participants if they had anxiety symptoms in the past 2 weeks. Items are rated from 0 (not at all) to 3 (nearly every day) indicating symptom frequency. In this study, GAD-7 with a score of ≥5 refers to a categorical cut-off for risk of anxiety (0= normal, 1= anxiety), GAD-7 has demonstrated a good reliability (Cronbach’s α = 0.89).
Independent variable and control variables
In order to compare the psychological problems of different groups during the onset of the epidemic, we divided residents into people quarantined in affected area, quarantined in unaffected area, and not quarantined. Place of current residence was surveyed by asking: ‘Are you living in Wuhan currently?’ and ‘Have you been to the affected areas from December 1, 2019 to January 25, 2020’. Participants had ‘Yes’ or ‘No’ response options and were categorized into the three groups based on their respective responses. Additionally, individual, household and community characteristics were included. Individual demographic variables included gender, age (three groups: 18–29, 30–49 and ≥50), education (groups: junior high school or below, senior high school, and college degree or above), marital status (single vs. married or formerly married) and income (three groups: <2000 RMB, 2000–4999 RMB and ≥5000 RMB) per month. The family variable assessed whether we live alone or not and whether the respondents exercised at home or not. Further, the community variable assessed the measures undertaken by the community to screen people with a travel history to and from Wuhan or screen people with corresponding symptoms.
Statistical analysis
We used four logistic regression models to analyse risk factors of depression and anxiety. In the first model, we only added the independent variable. From the second model to the fourth model, demographic variables, family variables and community variables were entered into the models in sequence. Odds ratio (OR) and 95% confidence interval (CI) were obtained from logistic regression. For all comparisons, differences were tested by two-tailed tests and p values < 0.05 were considered statistically significant. All statistical procedures were performed using the Statistical Package for Social Sciences 20.0 (SPSS Inc., Chicago, IL, USA) for Windows and R version 3.5.2.
Results
Table 1 displays the demographic characteristics of the sample (N = 1160) and the three groups. Of 1160 participants, 60.00% were female, 54.14% were less than 30 years old, 2.59% lived alone, and 14.57% exercised at home. A total of 65.26% of the participants quarantined in the unaffected area were screened for COVID-19.
Table 1. Descriptive statistics of variables during the COVID-19 outbreak (N = 1160).
Moreover, only 26.47% (95% CI: 23.93–29.00) of the respondents reported symptoms of depression, and respondents quarantined in unaffected areas [55.25% (95% CI:48.40–61.95)] reported a significantly higher rate of depression than those quarantined in affected areas [36.05% (95%CI:28.87–43.22)] and those not quarantined [16.12% (95% CI:13.53–18.72)] (
=143.610, p < 0.001). A total of 70.78% (95% CI:68.16–73.39) of the respondents reported symptoms of anxiety, and respondents quarantined in unaffected areas [84.02% (95%CI:79.17–88.87)] reported significantly higher rates of anxiety than those quarantined in the affected area [76.74% (95%CI:70.43–83.06)] and those not quarantined [65.67% (95%CI: 62.31–69.03)] (
=31.223, p < 0.001) (Figure 2).
Published online:
22 June 2020![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/gpsh20/2021/gpsh20.v036.i02/08870446.2020.1782410/20210501/images/medium/gpsh_a_1782410_f0002_c.jpg"})
Figure. 2. Percentage (95% CI) on depression and anxiety for different groups.
The univariable regression analysis (Model 1) revealed that the odds of reporting symptoms of depression were 2.932 higher in respondents quarantined in the affected area (95%CI: 2.034-4.227, p < 0.001) and 6.422 higher among respondents quarantined in unaffected areas (95%CI: 4.624-8.920, p < 0.001) (Table 2). Respondents quarantined in unaffected areas reported significantly higher scores on all four dimensions of depression than those quarantined in the affected area or not quarantined (Appendix, Figure A1). The odds of reporting symptoms of anxiety were 1.725 among respondents quarantined in the affected area (95% CI: 1.175-2.532, p < 0.001) and 2.748 in respondents quarantined in unaffected areas (95% CI: 1.859-4.063, p < 0.001). Further, the multivariable regression analysis (Models 2 to 4) revealed that the quarantined respondents remained significantly associated with reporting symptoms of depression and anxiety, compared to respondents not in quarantine (Tables 2–3).
Table 2. Risk factors of probable depression associated with the COVID-19 outbreak in China.
Table 3. Risk factors of anxiety associated with the COVID-19 outbreak.
Additionally, respondents who were males (OR = 1.466, 95% CI: 1.078-1.994), had a college degree and above (OR = 1.975, 95% CI:1.237-3.154), lived alone (OR = 2.353, 95% CI:1.003-5.524)), and exercised with family at home (OR = 1.600, 95% CI:1.074-2.384) reported a higher prevalence of probable depression. However, respondents aged 50 years or older (OR = 0.411, 95% CI: 0.233-0.725), having a higher income (2000–4999RMB: OR = 0.662, 95% CI: 0.459-0.956; ≥5000RMB: OR = 0.221, 95% CI:0.149-0.327), and who were screened in the community (OR = 0.667, 95CI% CI: 0.479-0.930) reported a lower prevalence of probable depression. Further, respondents who had a college degree and above (OR = 2.054, 95% CI: 1.298-3.251) and were married (OR = 2.273, 95% CI:1.427-3.621) reported a higher prevalence of anxiety, while male respondents (OR = 0.585, 95% CI:0.440-0.777), aged 50 years or older (OR = 0.329, 95% CI: 0.190-0.570), and having a higher income (2000–4999RMB: OR = 0.581, 95% CI: 0.362-0.932; > =5000RMB: OR = 0.187, 95% CI:0.121-0.288) reported a lower prevalence of anxiety.
Discussion
To the best of our knowledge, this study was the first to assess symptoms of depression and anxiety among people quarantined amidst the COVID-19 outbreak in China. Our findings show that the risk of mental illness was higher among quarantined people than those not quarantined. Moreover, the prevalence of depression and anxiety was two times higher among people quarantined in unaffected areas than those quarantined in affected areas.
The existing literature on mental health is mostly based on well-defined events, such as wars (refugee), physical and natural disasters, social unrest, and interpersonal violence (Cornwell & Laumann, 2015; Goldmann & Galea, 2014; Laugharne et al., 2011; Ni et al., 2020; Udomratn, 2008), as opposed to major public health emergencies. COVID-19 is infectious, deadly, and unpredictable, thus resulting in individual or mass hysteria and fear. In these stressful circumstances, individuals show a variety of psychological and behavioural reactions influenced by public opinion, mass media, and the impact of the outbreak in terms of restrictions on mobility and shortages of daily necessities. Some others may also experience psychological crises of varying degrees, especially those directly affected by the outbreak. COVID-19 spread to the whole country and subsequently to other countries because the outbreak occurred at the time of the Spring Festival during which billions of people travel throughout the country. Therefore, the outbreak affected not only people quarantined in hard-hit areas but also those quarantined in unaffected areas especially when unaffected areas carried out a strict screening for people who have been to Wuhan at that time, causing high prevalence of depression and anxiety. Although residents of hard-hit areas may be at a higher risk of experiencing fear and panic, those with a travel history to and from the epicentre may also be touched by the outbreak associated stigma, discrimination, and isolation (Samantha et al., 2020). Moreover, social disconnectedness, panic, and stigma may subsequently lead to depression (Bao et al., 2020; Rogers et al., 2020).The uncertainty and unpredictability regarding the risk, treatment, and control of the epidemic coupled with misinformation can instigate stress and psychological ill health, especially among people in quarantine as they face a lack of face-to-face contact, and loneliness.
Many communities required people to report if they had visited an affected area, especially Wuhan, and accordingly be tested for COVID-19 and quarantined. In this study, respondents living in communities where screening for COVID-19 was required were less likely to report depression and anxiety symptoms, which may implicate that people would get a sense of security in the community where screening for COVID-19 was required. Feeling powerless and lack of family support in the face of stress or dilemma may lead to depression, particularly in people who live alone (Talley et al., 2010), as family support acts as a buffer against stress. Furthermore, exercising at home was associated with depression, which is inconsistent with other study (Kandola et al., 2020). Perhaps in certain circumstances, exercise, leisure, and entertainment may prevent depression (Wang et al., 2019).
We considered socio-demographic variables as specific risk factors for depression and anxiety, and found that males were more likely to suffer from depression than anxiety as compared to females. Additionally, older adults reported a lower risk of depression and anxiety, as family reunion during the Spring Festival is the most enjoyable part of the year for elders. People with higher educational qualifications have a tendency to be more curious about the outbreak. However, the longer they focus on negative information, the greater the risk of mental illness. Thus, appropriate communication via social networks could be important for regulating public emotional response during an epidemic (Liao et al., 2019). Further, our findings suggest that a higher monthly income was associated with a lower risk of depression and anxiety, which may be due to less worry regarding the financial impact of the epidemic and ability to afford protective equipment, such as masks and disinfectants. However, low-income people are more likely to face unemployment in this outbreak, and rising unemployment is associated with an increased number of suicides (Wolfram & Carlos, 2020).
Our study, which was conducted two weeks after a complete lockdown was imposed in Wuhan, highlights two important findings. First, the mental health was worse in the quarantined area than in unquarantined areas, and deteriorated most in the quarantined unaffected areas, depression and anxiety were common mental health problems among people in quarantine. Therefore, it is crucial to include psychological interventions as an integral part of the prevention and control measures during public health crises, while also prioritizing high-risk groups such as quarantined people. Second, wide spread screening may play a positive role in protecting individuals against stigma, discrimination, and related negative psychological experiences. Hence, encouraging people to cooperate with community screening could serve as a way of fighting stigma.
There are some limitations to this study. First, although the snowball sampling was suitable method during the mass-quarantine, its nonprobability sampling technique, particularly the ‘seeds’ may bring several subjective biases and limit sample representativeness. For example, the sample used for analyses was slightly younger and better educated compared with the adult population of China, and individuals with wider social networks have greater probabilities of being selected and have more access to epidemic information, while public media reports often employed an alarming tone when the epidemic intensified (Smith, 2006), thus, the results of depression and anxiety could be potentially over-estimates. Further, there are some heterogeneity, our questionnaire did not assess the intensity of respondents’ activities in the affected area, which could ignore the residual confounding; Moreover, a cross-sectional survey cannot explore the trajectories of public psycho-behavioural responses, because our research design examined associations and predictive factors rather than causes of depression and anxiety at the rapid growth period. Further, the psychological burden of population quarantined in unaffected area will decrease, while population in other regions will feel an increased risk of infection and psychological stress as second - and third-generation confirmed cases emerge, our further studies would focus on these.
Despite these limitations, a rapid assessment of symptoms of depression and anxiety in three groups of people during the rising epidemic of COVID-19 was conducted. Our findings reported mental health problems associated with the epidemic, especially among people in quarantine. Health and social services need to be vigilant in recognizing possible signs and symptoms of mental health problems during and after COVID-19, which will require a substantial increase in psychological counselling and social support services. Moreover, enhancing the capability of community screening and encouraging people to cooperate with community screening are important, as appropriate community screening may reduce the risk of depression and anxiety during an epidemic.
| Variables | All | Quarantined in affected area | Quarantined in unaffected area | Without quarantined | p value |
|---|---|---|---|---|---|
| Depression | <.0001 | ||||
| Yes | 307(26.47%) | 62(36.05%) | 121(55.25%) | 124(16.12%) | |
| No | 853(73.53%) | 110(63.95%) | 98(44.75%) | 645(83.88%) | |
| Anxiety | <.0001 | ||||
| Yes | 821(70.78%) | 132(76.74%) | 184(84.02%) | 505(65.67%) | |
| No | 339(29.22%) | 40(23.26%) | 35(15.98%) | 264(34.33%) | |
| Gender | 0.414 | ||||
| Male | 464(40.00%) | 65(37.79%) | 81(36.99%) | 218(41.35%) | |
| Female | 696(60.00% | 107(62.21%) | 138(63.01%) | 451(58.65%) | |
| Age | <.0001 | ||||
| 18–29 | 628(54.14%) | 71(41.28%) | 140(63.93%) | 417(54.23%) | |
| 30–49 | 291(25.09%) | 40(23.26%) | 52(23.74%) | 199(25.88%) | |
| > =50 | 241(20.78%) | 61(35.47%) | 27(12.33%) | 153(19.90%) | |
| Education | <.0001 | ||||
| Junior high school and below | 332(28.62%) | 37(21.51%) | 56(25.57%) | 239(31.08%) | |
| Senior high school | 637(54.91%) | 86(50.00%) | 124(56.62%) | 427(55.53%) | |
| College and above | 191(16.47%) | 49(28.49%) | 39(17.81%) | 103(13.39%) | |
| Income/per month | <.005 | ||||
| <2000 RMB | 258(22.24%) | 44(25.58%) | 67(30.59%) | 147(19.12%) | |
| 2000–4999 RMB | 388(33.45%) | 52(30.23%) | 78(35.62%) | 258(33.55%) | |
| > =5000 RMB | 514(44.31%) | 76(44.19%) | 74(33.79%) | 364(47.33%) | |
| Marriage | <.0001 | ||||
| Married or formerly married | 599(51.64%) | 71(41.28%) | 138(63.01%) | 390(50.72%) | |
| Single | 561(48.36%) | 101(58.72%) | 81(36.99%) | 379(49.28%) | |
| Living alone | 0.155 | ||||
| Yes | 30(2.59%) | 7(4.07%) | 8(3.65%) | 15(1.95%) | |
| No | 1130(97.41%) | 165(95.93%) | 211(96.35%) | 754(98.05%) | |
| Exercise with family | 0.457 | ||||
| Yes | 169(14.57%) | 30(17.44%) | 33(15.07%) | 106(13.78%) | |
| No | 991(85.43%) | 142(82.56%) | 186(84.93%) | 663(86.22%) | |
| Community screening | <.0001 | ||||
| Yes | 757(65.26%) | 34(19.77%) | 149(68.04%) | 574(74.64%) | |
| No | 403(34.74%) | 138(80.23%) | 70(31.96%) | 195(25.36%) | |
| N | 1160 | 172 | 219 | 769 |
| Model 1 OR (95% CI) | Model 2 OR (95% CI) | Model 3 OR (95% CI) | Model 4 OR (95% CI) | |
|---|---|---|---|---|
| Group | ||||
| Quarantined in affected area | 2.932(2.034-4.227)* | 3.158(2.121-4.704)* | 3.128(2.095-4.760)* | 2.527(1.634-3.909)* |
| Quarantined in unaffected area | 6.422(4.624-8.920)* | 5.834(4.117-8.267)* | 5.866(4.132-8.329)* | 5.794 (4.076-8.234)* |
| Without quarantined | 1 | 1 | 1 | 1 |
| Gender | ||||
| Male | 1.423(1.049-1.930)* | 1.455(1.071-1.977)* | 1.466(1.078-1.994)* | |
| Female | 1 | 1 | 1 | |
| Age | ||||
| 18–29 | 1 | 1 | 1 | |
| 30–49 | 1.018(0.646-1.603) | 0.970(0.615-1.528) | 0.938(0.593-1.484) | |
| > =50 | 0.472(0.270-0.825)* | 0.433(0.247-0.759)* | 0.411(0.233-0.725)* | |
| Education | ||||
| Junior high school and below | 1 | 1 | 1 | |
| Senior high school | 1.296(0.907-1.851) | 1.283(0.896-1.837) | 1.325(0.924-1901) | |
| College and above | 1.883(1.184-2.993)* | 1.929(1.211-3.071)* | 1.975(1.237-3.154)* | |
| Income/per month | ||||
| <2000 RMB | 1 | 1 | 1 | |
| 2000–4999 RMB | 0.663(0.461-0.995)* | 0.660(0.458-0.951)* | 0.662(0.459-0.956)* | |
| > =5000 RMB | 0.220(0.149-0.325)* | 0.216(0.146-0.320)* | 0.221(0.149-0.327)* | |
| Marriage | ||||
| Married or formerly married | 0.866(0.559-1.342) | 0.955(0.615-1.484) | 0.989(0.634-1.543) | |
| Single | 1 | 1 | 1 | |
| Living alone | ||||
| Yes | 2.353(1.010-5.483)* | 2.353(1.003-5.524)* | ||
| No | 1 | 1 | ||
| Exercise with family | ||||
| Yes | 1.569(1.054-2.337)* | 1.600(1.074-2.384)* | ||
| No | 1 | 1 | ||
| Community screening | ||||
| Yes | 0.667(0.479-0.930)* | |||
| No | 1 |
Note: Odds ratios are obtained through multivariable analysis. Only the independent variable was added in Model 1. In Model 2, demographic variables (gender, age, education, income and marriage) were adjusted based on Model 1. In Model 3, family variables (living alone and exercise with family) were adjusted based on Model 2. In Model 4 (full model), community screening (whether the community has taken measures to screen people with a travel history to and from Wuhan or screen people with corresponding symptoms) was adjusted based on Model 3.
* p < 0.05.
| Model 1 OR (95% CI) | Model 2 OR (95% CI) | Model 3 OR (95% CI) | Model 4 OR (95% CI) | |
|---|---|---|---|---|
| Group | ||||
| Quarantined in affected area | 1.725(1.175-2.532)* | 1.794(1.173-2.743)* | 1.770(1.156-2.709)* | 1.796(1.137-2.836)* |
| Quarantined in unaffected area | 2.748(1.859-4.063)* | 2.334(1.537-3.544)* | 2.318(1.526-3.522)* | 2.322(1.528-3.528)* |
| Without quarantined | 1 | 1 | 1 | 1 |
| Gender | ||||
| Male | 0.582(0.438-0.772)* | 0.584(0.440-0.777)* | 0.585(0.440-0.777)* | |
| Female | 1 | 1 | 1 | |
| Age | ||||
| 18–29 | 1 | 1 | 1 | |
| 30–49 | 0.656(0.412-1.047) | 0.646(0.403-1.033) | 0.647(0.404-1.035) | |
| > =50 | 0.335(0.194-0.578)* | 0.329(0.190-0.569)* | 0.329(0.190-0.570)* | |
| Education | ||||
| Junior high school and below | 1 | 1 | 1 | |
| Senior high school | 1.993(1.464-2.713)* | 1.987(1.459-2.706) | 1.983(1.455-2.702) | |
| College and above | 2.045(1.293-3.236)* | 2.056(1.299-3.254)* | 2.054(1.298-3.251)* | |
| Income/per month | ||||
| <2000 RMB | 1 | 1 | 1 | |
| 2000–4999 RMB | 0.583(0.364-0.936)* | 0.582(0.363-0.933)* | 0.581(0.362-0.932)* | |
| > =5000 RMB | 0.188(0.121-0.290)* | 0.187(0.121-0.289)* | 0.187(0.121-0.288)* | |
| Marriage | ||||
| Married or formerly married | 2.230(1.407-3.533)* | 2.278(1.431-3.626)* | 2.273(1.427-3.621)* | |
| Single | 1 | 1 | 1 | |
| Living alone | ||||
| Yes | 1.358(0.563-3.274) | 1.357(0.563-3.271) | ||
| No | 1 | 1 | ||
| Exercise with family | ||||
| Yes | 1.072(0.718-1.600) | 1.071(0.718-1.599) | ||
| No | 1 | 1 | ||
| Community screening | ||||
| Yes | 1.028(0.748-1.415) | |||
| No | 1 |
Note: Odds ratios are obtained through multivariable analysis. Only the independent variable was added in Model 1. In Model 2, demographic variables (gender, age, education, income and marriage) were adjusted based on Model 1. In Model 3, family variables (living alone and exercise with family) were adjusted based on Model 2. In Model 4 (full model), community screening (whether the community has taken measures to screen people with a travel history to and from Wuhan or screen people with corresponding symptoms) was adjusted based on Model 3.
* p < 0.05.
| Demographic composition of sample compared to 2018 CFPS* | |||||
|---|---|---|---|---|---|
| Demographics | CFPS 2018* | Sample (Unweighted) | Effect size | Sample (Weighted) | Effect size |
| N (%) | N (%) | CFPS 2018 VS Sample (Unweighted) | N (%) | CFPS 2018 VS Sample (Weighted) | |
| Gender | |||||
| Male | 15,556(47.61%) | 464(40.00%) | 0.15 | 552(47.59%) | <0.01 |
| Female | 17,113(52.39%) | 696(60.00%) | 608(52.41%) | ||
| Age | |||||
| 18–29 | 8,768(26.84%) | 628(54.14%) | 0.62 | 311(26.84%) | <0.01 |
| 30–49 | 11,140(34.1%) | 291(25.09%) | 396(34.10%) | ||
| > =50 | 12,760(39.06%) | 241(20.78%) | 451(38.89%) | ||
| Education | |||||
| Junior high school and below | 13763(42.13%) | 332(28.62%) | 0.39 | 489(42.13%) | <0.01 |
| Senior high school | 11515(35.25%) | 637(54.91%) | 409(35.25%) | ||
| College and above | 7391(22.62%) | 191(16.47%) | 262(22.62%) | ||
| Income/month | |||||
| <2000 RMB | 11,270(34.5%) | 258(22.24%) | 0.34 | 400(34.5%) | <0.01 |
| 2000–4999 RMB | 11,796(36.11%) | 388(33.45%) | 419(36.11%) | ||
| > =5000 RMB | 9603(29.39%) | 514(44.31%) | 341(29.39%) | ||
| Marriage | |||||
| Married or formerly married | 21,891(67.01%) | 599(51.64%) | 0.32 | 777(67.01%) | <0.01 |
| Single | 10,778(32.99%) | 561(48.36%) | 383(32.99%) | ||
Cohen's w effect size: small 0.1; medium 0.3; large 0.5.
2018 CFPS*: The China Family Panel Studies, launched by Peking University, is a nearly nationwide, comprehensive, longitudinal, recognized random sample social survey. Which promises to provide the academic community with the most comprehensive and highest-quality survey data on contemporary China. http://www.isss.pku.edu.cn/cfps/gycfps/cfpsjj/index.htm.
Acknowledgements
The authors thank http://www.survey.163.com for providing technical support and all the respondents.
Disclosure statement
No potential conflict of interest was reported by the authors. No conflict of interest exits in the submission of this manuscript, and manuscript is approved by all authors for publication. This work was original research that has not been published previously, and not under consideration for publication elsewhere. The authors listed have approved the manuscript that is enclosed.
Data sharing
Data obtained for the study will not be made available to others.
Published online:
22 June 2020Figure A1. Scores (95% CI) on CES-D, depressed mood, positive mood, physical symptom and interpersonal relationship for different groups.
![]({"type":"image","src":"/na101/home/literatum/publisher/tandf/journals/content/gpsh20/2021/gpsh20.v036.i02/08870446.2020.1782410/20210501/images/medium/gpsh_a_1782410_a0001_c.jpg"})
Figure A1. Scores (95% CI) on CES-D, depressed mood, positive mood, physical symptom and interpersonal relationship for different groups.
Table A1. Demographic composition of sample compared to 2018 CFPS*.