The target participants in this survey were medical workers working in medical institutions. The medical institutions encompass hospitals, primary healthcare centers, centres for disease management and prevention, as well as maternity and child health care hospitals and sanatoriums. According to their duties, medical workers were divided into three occupations, including doctors, nurses, and paramedical workers (PMW). Paramedical workers included medical technologists, administrators, logisticians, and researchers.

The sample was acquired using a non-probability sampling methodology, and an approach similar to quota sampling was used to calculate the sample size, guaranteeing an adequate representation of respondents from each occupation. Based on a survey using SDS and SAS during the Severe Acute Respiratory Syndrome (SARS) epidemic, the incidence of depression and anxiety in nurse was 21.42 and 25%, respectively (21). So, the lowest prevalence of mental health problems was assumed to be 20% for each occupation. The minimum required sample size (n) was calculated as n = μ_α² p (1 - p)/δ², where μ_α = one-sided magnitude of confidence (μ_α = 1.96 for α = 0.05), p = expected proportion of the outcome of interest. δ = error range (22). To achieve a relative error of 15%, and considering 5% non-response (23), the minimum effective sample size for each occupation would be about 718. It would require about 2,154 participants in total.

A cross-sectional epidemiological research was done from February 17 to 24, 2020, to assess mental health problems among medical professionals working at medical institutes in mainland China (23). An online survey utilizing the Questionnaire Star program (Wenjuanxing), a professional online questionnaire editing program, to edit the questionnaire. ¹ After completing the editing, a QR code-a bar code readable by mobile devices that contains the website link to the questionnaire utilized in this research-was generated, and researchers sent a QR code poster to medical workers via WeChat messages. Organizations use WeChat groups to post work announcements; our researchers initially sent the QR code to different WeChat work groups consisting of medical personnel. In addition, 60 department directors were asked to introduce the survey and QR code poster to their medical colleagues to increase the number of participants. The participants can scan the code to access the Questionnaire Star website and complete the questionnaire electronically. All medical personnel who received a QR code poster were eligible to take part in the survey.

Twenty-three small WeChat groups (of 50–60 members each) and five large WeChat groups (of 500 members each) were provided with the QR code. Additionally, the QR code was distributed to 60 department directors from 60 hospitals to increase the number of participants, and these directors were asked to share the code among their staff members. Once the questionnaire was submitted, the data would be stored on the server of Questionnaire Star and could only be downloaded by an authorized data manager.

The Ethical Committee of Beijing Ditan Hospital Capital Medical University (BJDTH) has granted approval for this survey (JINGDILUNKE (2020)-(012)-01). Writing informed consent was not required of any participant in this web-based survey, as our participation was entirely anonymous and voluntary. Participants were permitted to exit or interrupt the survey without providing an explanation for their actions, and the objectives and methods of the survey were stipulated in a clear manner at the beginning of the survey. And after completion of information collection and collation, the electronic version of the database was password-set for the sole use of the research team to strictly protect the security of the data. All investigational procedures were conducted in accordance with the 1964 Helsinki Declaration and its subsequent amendments, as well as the pertinent ethical regulations of the national and internal research committees.

The prevalence of mental health problems was estimated for SRQ-20, SAS, and SDS-positive. Weighting was not considered during analysis as the sample design of the survey was non-probability. The χ² test was used to test the prevalence between three occupations. The Poisson regression was used to analyze the prevalence ratios (PR) by controlling age, gender, education, marital status, and profession (32). A participant was defined as having any mental problem if he or she received one positive score on the three scales. In each occupational category, univariate logistic regressions were conducted to analyze the relationships between having any mental problem and correlates, followed by multivariable logistic regression that showed significant univariate associations. PR values in Poisson regression and OR values in logistic regression and their 95% Confidence Intervals (95%CI) were reported. In this study, a p value less than 0.05 was considered significant. All statistical analyses were conducted using the Statistical Package for Social Sciences (SPSS, version 22.0; SPSS Inc., Chicago, IL, USA).

A sample of 5,495 medical workers submitted the questionnaire, nine invalid responses were excluded. Of the 5,486 medical workers from 471 medical institutions in 31 provinces, 5,380 (98.1%) worked in hospitals, 72 (1.3%) in primary care institutes, 32 (0.6%) in maternal and child health care hospitals and sanatoriums, and 2 at the CDC. About 44.9% of the sample was from Beijing, the capital of China, and only 2.5% was from Hubei, where the largest numbers of COVID-19 patients were located. The sample involved 1,916 (35.0%) doctors, 2,857 (52.0%) nurses, and 713 (13.0%) PMWs. The average ages of the total sample, doctors, nurses, and PMWs were 36.2 ± 9.0 years old, 40.3 ± 8.6 years old, 33.1 ± 8.1 years old, and 37.6 ± 8.9 years old, respectively.

The characteristics of the three occupations are shown in Table 1. There were significant differences among the three occupations in most variables. Nurses had a higher female percentage than doctors and PMWs. More than 94% of participants had experience receiving training about infectious diseases and training of COVID-19. More than 30% of doctors, 23% of nurses, and 28% of PMWs were overworked. More than 25% of medical workers did not have enough time to rest during the COVID-19 outbreak. More than 37% of medical workers did not get enough PPE. More than 36% of medical workers reported they did not receive enough care from their leaders, and about 11% of medical workers did not get enough support from colleagues. About 6% of medical workers could not receive enough support from their families. There were no significant differences among three occupations on the variables about the previous training in treating infectious disease, training of COVID-19, and previous physical health.

The prevalence of mental health problems in each occupation is shown in Table 2. The prevalence of psychological pain in three occupations was 16.7% in doctors, 15.2% in nurses, and 13.5% in PMWs, respectively. Nurses had the highest prevalence of anxiety (39.1%; 95%CI: 37.3–30.9%) than those doctors (27.9%; 95%CI: 25.9–29.9%) and PMWs (30.0%; 95%CI: 27.3–34.1%) respectively. The prevalence of depression in nurses (17.4%; 95%CI: 16.0–18.8%) was higher than that in doctors (13.8%; 95%CI: 12.3–15.4%) and PMWs (13.7%; 95%CI: 11.2–16.3%). Regarding the prevalence of any mental health problems, nurses had the highest percentage (43.6%), followed by PMWs (34.6%; 95%CI: 31.1–38.1%), and doctors (32.9%; 95%CI: 30.8–35.0%). There were significant differences among the three occupations in the prevalence of anxiety, depression, and any mental health problems.

The prevalence ratios (PR) between three occupations were illustrated in Table 3. Doctors had significant higher psychological pain than PMWs (PR = 1.30; 95%CI: 1.02–1.65). Nurses had significant higher prevalence of anxiety and depression than doctors and PMWs. Moreover, nurses also had significant higher prevalence of any mental health problems than doctors (PR = 1.23; 95%CI: 1.08–1.41) and PMWs (PR = 1.25; 95%CI: 1.09–1.45).

The results of significant associations in univariate logistic regressions and multivariable logistic regression in three occupations were shown in Table 4. Training experience was not included because the sample size of workers without training experience was too small. For all three occupations, the mental health problems were not associated with gender, marital status, having children younger than 18 years old, professional titles or previous experience in treating infectious diseases.

In the multivariable regression, doctors aged 35–44 years (OR = 1.31; 95%CI: 1.01–1.70), with a college certificate or lower (OR = 2.56; 95%CI: 1.54–4.24) or bachelor’s degree (OR = 1.28; 95%CI: 1.01–1.62), not living with family members (OR = 1.63; 95%CI:1.26–2.12), being overworked (OR = 1.45; 95%CI:1.06–1.98), not having enough time to rest (OR = 1.73; 95%CI:1.26–2.36), not always having support from colleagues (most of the time: OR = 1.32; 95%CI:1.01–1.74; hard to say or no: OR = 2.34; 95%CI:1.55–3.54), less family’s support (partly support: OR = 1.45; 95%CI:1.13–1.86; no support: OR = 3.21; 95%CI:2.01–5.12), poor previous mental health status (OR = 3.36; 95%CI:2.28–4.95) and poor previous physical health status (OR = 1.61; 95%CI:1.10–2.36) were more likely to have more mental health problems.

Nurses working in a designated institution for COVID-19 treatment (OR = 1.22; 95%CI:1.02–1.45), positive COVID-19 event exposures (OR = 1.34; 95%CI:1.01–1.83), being overworked (OR = 1.65; 95%CI:1.31–2.07), not having enough time to rest (OR = 2.00; 95%CI:1.61–2.48), less support from colleagues (OR = 1.22; 95%CI:1.02–1.45), less family support (OR = 1.74; 95%CI:1.23–2.46), and poor previous mental health status (OR = 3.86; 95%CI:2.82–5.28) were likely to have a higher prevalence of mental health problems than the contrast group.

PMWs with a college certificate or lower (OR = 2.80; 95%CI:1.40–5.61) or bachelor’s degree (OR = 2.07; 95%CI:1.10–3.90), working in a high-risk department (OR = 1.22; 95%CI:1.02–1.45), being overworked (OR = 1.67; 95%CI:1.07–2.61), not having enough time to rest (OR = 2.04; 95%CI:1.31–3.14), not having enough care from leaders (OR = 1.54; 95%CI:1.02–2.34), less support from colleagues (OR = 2.18; 95%CI:1.14–4.15), and poor previous mental health status (OR = 2.99; 95%CI:1.54–5.78) were likely to have a higher prevalence of mental health problems than the contrast group.

A previous study found that women reported more severe symptoms of depression, anxiety, and distress (7). However, there was no correlation with gender across the three occupations in our study. This may result from the number of female medical workers involved. There were 76.7% females in the previous, while in our study, 80.8% of participants were women. In addition, the previous study mainly consisted of physicians and nurses (7, 11); we included paramedical workers, which may be another potential reason. The study found a significant correlation between educational level and mental health in doctors and PMWs, indicating that individuals with lower educational levels were prone to experience mental health problems. In this study, doctors had the highest percentage of individuals with master’s or PhD degrees, followed by PMWs and then nurses. Even though more than 98% of medical workers were trained with COVID-19, individuals with higher education were less likely to be influenced by the outbreak, which may be attributed to the potential impact of higher education on enhancing individuals’ understanding of psychological distress issues. Consequently, individuals with higher education may be more likely to take positive measures to prevent related symptoms (38). The correlation between age and mental problems was shown in doctors, indicating that younger doctors had more emotional problems than older people. This was in line with the results of previous studies that have shown that younger medical workers had higher stress and depressive symptoms than older workers (18, 39). However, this relationship did not exist in both nurses and PMWs.

Working status during the outbreak was an essential in impacting the mental health of medical workers (9, 10). Being overworked and not having enough time to rest were related to mental health problems in all three occupations, which was in line with previous studies indicating that due to the heavy workloads and hazardous working environments, healthcare workers were at high risk of psychological stress (6, 10, 39, 40). Therefore, it is necessary to provide more healthy and safe working conditions to healthcare workers. Although wearing personal protective equipment was the most effective and common way to protect medical workers from being infected in the context of lacking any clear treatments, our results indicated that personal protective equipment was not the influencing factor of mental health problems in all three occupations, which was inconsistent with previous studies which indicated that inadequate PPE was associated with psychological stress and depressive symptoms (41). The greater accessibility of PPE compared with general people may explain this difference. A study showed that the general public who have less access to PPE could contribute to a higher prevalence of anxiety and depression than that of frontline medical workers (35). Another unexpected result was that the working department was not related to mental health problems in doctors, nurses, and PMWs. This result disagreed with a previous study conducted during the SARS (15); the difference may be due to the following reason. The average incubation period for COVID-19 was 14 days; it was uncertain whether the patients they treated in low-risk departments were infected with COVID-19 during the incubation period (11). Consequently, medical workers in high-risk or low-risk departments had to be worried about the infectious risks.

Social support is also a key for medical workers to cope with psychological symptoms (15). The findings presented in the current study indicated that individuals having less support from others had a high prevalence of mental health problems. For doctors and nurses, support from colleagues and family, but not from their supervisor or head of a department, could reduce mental health problems. This finding aligned with the study conducted in Singapore (42), which revealed that doctors and nurses who encountered psychiatric symptoms found the support from their supervisor or head of the department unhelpful. However, they reported less emotional distress when they received support from their colleagues. For PMWs, individuals who did not receive enough support from their supervisor or head of a department reported more mental problems than those who received. It may be that the work of PMWs is more relevant to the authorities than other occupations during the outbreak. Those results were consistent with prior research, which showed that receiving support from family, friends, and colleagues may serve as a protective factor for mental health during infectious disease outbreaks (43, 44). Furthermore, our study provided more evidence that medical workers in different occupations may need support from different sources.

For nurses but not for doctors and PMWs, their mental health status was associated with their experience of COVID-19 event exposures. In this sample, quarantining was discovered to be a strong predictor of a high level of mental health issues, which is consistent with earlier studies examining the psychological impact of quarantining on both the general population and front-line healthcare staff involved in combating an outbreak (45, 46). Overall, researchers point to a wide-ranging, substantial, and prolonged psychological impact of quarantine (47). Furthermore, another study found a robust correlation between the experience of being in quarantine and the severity of depressive symptoms, even 3 years after the isolation measures were terminated (48).

Previous mental health status was relevant to mental health during the outbreak in all three vocations examined in this study. The finding was line with a prior study, which indicated that health workers with poor or fair health conditions had higher stress and mental problems (49). Poor previous physical health status was a risk factor for mental health in doctors but not in nurses and PMWs.