We used cross-sectional data from a multicenter, nationally representative survey of clinical therapists in China, which was collected anonymously through WeChat from January 2021 to March 2021. We adopted whole-group sampling to investigate alcohol use and misuse among clinical therapists and their risk factors during the early Covid-19 pandemic. All clinical therapists (N = 445) from 41 tertiary psychiatric hospitals in 29 provinces in China were invited to participate in the survey. Excluding the unfinished questionnaires and the questionnaires with logical problems, a total of 396 questionnaires were finally included in the statistical analysis (the response rate was 89.0%).

In this cross-sectional study, Socio-demographic characteristics (gender, age, education levels, marital status, income), health-related behaviors (smoking or not, insomnia, regular physical exercise), alcohol use/misuse, and occupational burnout were collected with the online questionnaire. AUDIT-C was used to investigate the patterns of alcohol use, and Maslach Burnout Inventory Human-Services-Survey (MBI-HSS) was used to investigate their occupational burnout.

We developed the electronic questionnaire based on existing literature and prior studies (16, 21). AUDIT-C is a concise version of the AUDIT and was used to assess patterns of alcohol use among clinical therapists. A higher total score (ranging from 0 to 12) is associated with higher severity of alcohol use problems. Through the AUDIT-C scores, the participants were divided into probable alcohol misuse [a total score of ≥3 (women) and ≥ 4 (men)] and low-risk users of alcohol [a total score of <3 (women) and < 4 (men)] (22, 23).

Occupational burnout was assessed using the MBI-HSS, which is a psychological assessment instrument including 22 items about occupational burnout that includes three subscales: emotional exhaustion (EE, which assesses the experience of being emotionally exhausted by the demands of work), depersonalization (DP, refers to the degree to which each one recognizes attitudes of unfeeling and impersonal response from the recipients of their care) and personal accomplishment (PA, refers to the feelings of self-efficacy and accomplishment at work). We considered the cut-off of EE ≥ 27 points or DP ≥ 10 points to define the presence of occupational burnout following previous studies (24).

Data analysis was performed using IBM’s Statistical Package of Social Sciences (SPSS version 22.0). The sample distribution was conducted using frequency for categorical variables and mean ± standard deviation for continuous variables. For the statistical analysis, the Chi-square test was utilized to assess the variables which were not in the normal distribution. The independent correlates of current alcohol use and alcohol misuse were examined by multiple logistic regression, and alcohol use or alcohol misuse were the dependent variables (yes = 1, no = 0). Age groups (≤34 years old, 35–49 years old, or ≥ 50 years old), educational level (associate degree or less, college degree, master’s degree or more), gender (male and female); marital status (single, married, divorced/widowed); smoking habit (yes/no); and other studied variables were entered as the independent variables. The level of statistical significance was set at value of ps of 0.05 (two-tailed).

The socio-demographic characteristics of our sample were shown in Table 1. The mean age of participants was 33.8 years old. 307 subjects (77.5%) were female, more than two-thirds (67.2%) were younger than 35, and 62.4% were married. 155 participants (39.1%) self-reported as current alcohol users. Significant differences were found between participants who did and did not report alcohol with respect to the demographic (gender, education, marital status) and health-related behaviors (smoking habit, regular exercise). Significant differences were found between those with and without alcohol misuse concerning the demographic (gender, education, smoking habit) and burnout factors (depersonalization) (all p< 0.05) (Table 1). Nearly one-fifth (19.9%) of responding clinical therapists reported symptoms of burnout with high emotional exhaustion in 46 (11.6%), and high depersonalization in 61 (15.4%). Among 155 current alcohol users, 41 (26.5%) reported alcohol use less than before, 103 participants (66.5%) reported no change, and 11 (7.1%) reported an increase in alcohol use than before. Participation in the front-line work of treating COVID-19 was not significantly associated with alcohol use and misuse among Chinese clinical therapists.

Less than one-third (30.3%) of all participants reported consuming alcohol once a month or less, 6.3% reported at least twice a month, and 2.5% reported twice a week or more. More than two-thirds (68.4%) of all current alcohol users were in the age group of ≤34 years old, and 53.8% of all alcohol misusers were ≤ 34 years old. The prevalence of alcohol misuse was 6.6% overall and there was a significant gender difference (16.9% in males and 3.6% in females, p < 0.001). The overall prevalence of alcohol use among the participants was 39.1% with a significant gender difference (70.8% prevalence in males and 30% in females, p < 0.01. Compared to those with Master’s degrees or more (39.1%), those with Bachelor’s degrees or less (60.9%) had significantly higher rates of alcohol use and misuse (both p < 0.01). The smoking rate among alcohol misusers was more than 5 times higher than those who did not report alcohol misuse (30.8% vs. 4.6%, p < 0.001). The smoking rate among alcohol users was higher than that of non-alcohol misusers (13.5% vs. 1.7%, p < 0.01). The rate of alcohol use was significantly higher in people with regular exercise (defined as exercising at least three times per week in the past month according to the recommendations of the National Fitness Guideline) than in those without regular exercise (44.5% vs. 22.7%, p < 0.01). Alcohol misuse was higher among clinical therapists who reported high job burnout (depersonalization) than those who did not report such burnout (14.8% vs. 5.1%, p < 0.01) (Table 1).

We divided the samples (N = 396) into current alcohol users (those who drank alcohol in the past 12 months; N = 155) and non-alcohol users (N = 241) and performed multiple logistic regression analyses to examine the associations between alcohol use and other factors. The references of the categorical variables were defined as shown in Table 2. Alcohol use among clinical therapists was associated with male gender (OR = 4.392; 95% CI =2.443–7.894), marital status (single, OR = 1.652; 95% CI =0.970–2.814), smoking habit (OR = 3.847; 95%CI =1.160–12.758) and regular exercise (OR = 2.719; 95%CI =1.490–4.963).

We also used AUDIT-C cut-off scores (alcohol misuse: the AUDIT-C score of ≥3 (women) and ≥ 4 (men)) to divide participants into those with probable alcohol misuse (N = 26) and those without (N = 370). Multiple logistic regression was used to examine the association between probable alcohol misuse and other variables (Table 3). A higher risk of alcohol misuse was associated with male gender (OR = 3.367; 95% CI =1.174–9.655), low education level (OR = 3.788; 95%CI =1.009–14.224), smoking habit (OR = 4.626; 95%CI =1.277–16.754) and high burnout (depersonalization, OR = 4.848; 95%CI =1.433–16.406).

Some common variables were significantly associated with both alcohol use and alcohol misuse among clinical therapists, such as smoking habits (OR = 3.847 and 4.626 in current alcohol users and alcohol misusers, respectively) and male gender (OR = 4.392 and 3.367 in current alcohol users and alcohol misusers, respectively). Lower education level (bachelor’s degree or less) and occupational burnout (depersonalization) were significantly associated only with alcohol misuse, while marital status (single) and regular exercise were significantly associated only with alcohol use.

The results of this study reported rates of alcohol use and misuse that were significantly higher in males (70.8 and 16.9%, respectively) compared with females (30.0 and 3.6%, respectively). In the multivariate analysis, male clinical therapists were more likely to be alcohol users (OR = 4.392; 95% CI =2.443–7.894) and misusers (OR = 3.367; 95% CI =1.174–9.655) than female therapists. Consistent with previous studies (29–32), male gender was a significant risk factor for alcohol use and alcohol misuse. A recent study of mental health professionals in 41 hospitals in China (3,479 males and 10,501 females) also found significant differences between male and female medical workers in alcohol misuse with nearly two-thirds (63.3%) of alcohol misusers being male (21).

We also found a significant association between smoking habits and alcohol consumption. Cigarette use was significantly associated with reported alcohol misuse (OR = 4.626; 95%CI =1.277–16.754) and alcohol use (OR = 3.847; 95%CI =1.160–12.758). Our finding about the association between alcohol use/misuse and cigarette use among clinical therapists is consistent with findings from several other studies conducted in China (33, 34) and other countries (29, 35–37). The combined use of alcohol and tobacco have a multiplicative effect on the risk of health problems (38–40). Tverdal et al. (41) found that an association between higher levels of alcohol consumption (more than one alcohol glass/day) and an increased risk of pancreatic cancer could be explained by tobacco use. Some studies showed that helping patients quit and reduce smoking may reduce their alcohol use/misuse (35, 42). Therefore, addressing tobacco use could also be a way to decrease the risk of alcohol use/misuse among clinical therapists.

Our survey found that single marital status and regular physical exercise were risk factors for alcohol use among clinical therapists. 75.5% of clinical therapists in our sample reported exercising regularly. A positive association between exercise and alcohol use is also found in prior studies of non-clinical individuals across the world, potentially explained by social factors, personality factors, and shared reward response pathways in the mesocorticolimbic region. The positive relationship between single marital status and drinking may be explained by increased engagement in social drinking (e.g., going to bars; drinking after regular exercise) among unmarried participants (43). Consistent with prior studies, a stable marital relationship may also reduce the risk of medical staff’s alcohol use and alcohol misuse (21). Compared to never-married individuals, divorced/separated individuals were significantly more likely to report binge drinking in the past year (44). However, the results of our survey do not suggest a significant association between divorced or widowed and alcohol use and misuse, which may be due in part to the specific characteristics of clinical therapists such as their ability to cope with and process through challenging situations.

A low education level (bachelor’s degree or less) was a risk factor for alcohol misuse among clinical therapists. Studies from other countries also suggest that education level may affect perception and use of alcohol. Individuals that attained lower education levels may consume more alcohol than those that attained higher education levels (45). Previous studies showed that healthcare workers with higher education levels might have a higher risk awareness regarding alcohol use/misuse and are less likely to misuse alcohol. However, two studies found the opposite. One study in Austria involving 400 office-based physicians showed that physicians with higher education admitted to drinking alcohol more frequently (46). Another recent study of 13,980 medical professionals in China found that those with high education levels drink more heavily than those with lower education levels, especially among doctors (21). One possible explanation is that a higher level of education usually means more disposable money and a possibly broader social network, which may in turn contribute to drinking (21, 47).

Similar to the results of previous studies (9, 48–51), occupational burnout among healthcare workers was at a relatively high level (19.9%). In our study, occupational burnout (depersonalization) was associated with an increased risk of alcohol misuse. Alcohol consumption may be a coping strategy to deal with the occupational burnout related to the COVID-19 pandemic (11). Similarly, a cross-sectional survey of 4,000 randomly selected physicians in Danish found that the association between risky alcohol consumption and alexithymia was partially mediated through depersonalization (15). As risky alcohol use and burnout may independently affect patient safety, there is also a need to stratify which risk factors are most likely to lead to occupational burnout to develop appropriate interventions during the epidemic (14, 50). Recent literature explores the impact of the COVID-19 pandemic on burnout among healthcare workers (9, 17), even among clinical therapists (17), and how it has significantly impacted the mental well-being of healthcare workers overall (7, 52). Healthcare workers have reported high levels of burnout, further exacerbated by effects from the COVID-19 pandemic (14). Stress, emotional distress, individual and psychosocial factors, and personality traits are found to be potential risk factors for burnout (53, 54). The management and prevention of occupational burnout in clinical therapists remain important issues as they may lead to turnover and affect workforce sustainability.

Three limitations should be considered when interpreting the findings of the current study. First, the heterogeneity of tools used to measure occupational burnout and AUDIT-C limited the ability to compare the rate of occupational burnout and alcohol misuse across different studies. Second, as a cross-section survey, we cannot infer the causal relationship between alcohol misuse/use and other factors. Third, outcomes are based on self-reports of the amount of drinking/problematic behaviors as opposed to clinical measures.