Gender differences in alcohol abuse/dependence among medical undergraduates during the post-COVID‑19 pandemic period (October 20, 2020-April 5, 2021) in China

A cross-sectional web-based survey was conducted between October 20, 2020, and April 5, 2021 in China. During this period, the COVID-19 pandemic was subsiding in most areas of China and medical undergraduates were returning to campus. This study is a part of a large online study on the mental health of medical students and professionals in China [47, 48]. We used the snowballing sampling technique to recruit participants. The inclusion criteria included 1). undergraduate medical students; 2). Age greater than or equal to 18 years old. The excluded criteria were as follows:1). non-completion of the questionnaire; 2). invalid questionnaires. The researcher predetermined the criteria for determining an invalid questionnaire: (1) wrong answers for the general knowledge question; (2) all the answers are the same throughout the questionnaire; and (3) there is inconsistency in the logical consistency check.

A sample size of 3008 produces a two-sided 95% confidence interval with a width equal to 0.020 when the sample proportion is 0.083. Sample proportion based on alcohol abuse/dependence reported by 8.25% of Chinese medical students in previous study [9]. The sample size was calculated using PASS 15.0.5. Given the possibility of invalid questionnaires, the sample size should be increased by 10%. Therefore, this study requires at least 3,308 participants.

In order to recruit a large number of participants as conveniently as possible, we used the snowballing sampling technique. The data were collected via an online survey platform (http://​www.​wjx.​cn) and a social media platform (WeChat) in China. First, 40 medical students and medical university teachers were selected as the “original delivers”, all of whom were members of this research group. They shared this survey link through WeChat group and Moments to invite their classmates or students to participate in this survey. These delivers and potential participants are WeChat friends or in the same WeChat group. By clicking on the survey link to participate in the study. In addition, participants were also encouraged to share this survey link via their WeChat group and Moments. Only participants who responded to all the questions were able to submit the questionnaire, thus, there were no missing data in this study.

The online questionnaire was written in Chinese. This survey included items on alcohol abuse/dependence, mental distresses (learning burnout, depression symptoms, anxiety symptoms, excessive daytime sleepiness, and history of mental disorders), as well as basic demographic information (age, gender, relationship status, monthly income, and years of school). Before starting the final questionnaire dissemination, it was piloted on 20 medical students. These piloted medical students were not included in the final analysis.

Firstly, each IP address can only fill in the questionnaire once to avoid repeated data collection. Secondly, the criteria for invalid questionnaires were set in advance by the researcher and the sample size was higher than the minimum required sample size. Therefore, the effect of deleting the invalid questionnaire on the results of the study was negligible.

This study was approved by the ethics committee of the Second Xiangya Hospital of Central South University (LYF2020075). All the participants were voluntary, with informed consent provided through an online survey platform (http://​www.​wjx.​cn). Immediately after completing the survey, participants will receive a report on their mental health status and recommendations to promote their mental health status. In addition, participants will be given the contact information of our team of mental health professionals to enable them to contact us for help.

A total of 3,566 medical students participated in this online survey. Through data screening, 154 respondents were deemed invalid and excluded from the analysis. Of the 154 excluded participants, 32 had incorrect answers to the general knowledge questions and 41 had the same answers throughout the questionnaire; 81 participants had inconsistencies in the logical consistency check. Thus, 3,412 participants were included in the final analysis, with the efficiency rate of 95.7%. The flow chart for participant recruitment is shown in Fig. 1. The participant characteristics (overall and by gender) are listed in Table 1. In this study, there were 1067 (31.3%) male and 1,027 (68.7%) female. Their age range was mainly 18–20 years (72.3%), relationship status was mainly single (78.4%), year in school range mainly 1–3 years (87%), and monthly income was mainly < 615 CNY (47.7%). In the whole studied population, a total of 350 (10.3%) medical undergraduates met the criteria for alcohol abuse/dependence, and the prevalence of learning burnout, depression symptoms, anxiety symptoms, excessive daytime sleep, having a history of mental disorders was 38.3%, 17.9%, 8.9%, 42.9%, and 2.9%, respectively.

Compared with females, male students showed a higher prevalence of alcohol abuse/dependence (16.6% vs 7.4%, p < 0.001). This difference remained significant after controlling for characteristics including age, relationship status, monthly income, anxiety symptoms, depression symptoms, learning burnout, excessive daytime sleepiness, and history of mental disorders (B = 1.009, Wald statistic = 68.804, p < 0.001, odds ratio [OR]: 2.742, 95% confidence interval [95%CI]: 2.161–3.481). In addition, the male students showed a higher prevalence of learning burnout (40.9% vs 37.2%, p = 0.041) and a lower prevalence of excessive daytime sleepiness (33.6% vs 47.1%, p < 0.001) than their female counterparts. Gender differences in the proportion of years of school, and monthly income were also significant (both p < 0.001, Table 1). There were no significant differences in the proportion of age, relationship status, depression symptoms, anxiety symptoms, or history of mental disorders between female and male students (all p > 0.05, Table 1).

The univariate analyses of male medical students with alcohol abuse/dependence demonstrated that alcohol abuse/dependence was associated with learning burnout, depression symptoms, anxiety symptoms, excessive daytime sleepiness and a history of mental disorders (all p < 0.05, Table 2). In addition, inter-group differences were also significant in the proportion of age, relationship status, and monthly income (all p < 0.05). We then performed a multivariate logistic regression and variables with p < 0.05 in the univariate analysis were included in this model (Table 3). The results of the Hosmer–Lemeshow goodness-of-fit test showed that χ2 = 5.816, P = 0.561, indicating that the model fit was good. Multivariate logistic regression indicated that alcohol abuse/dependence was significantly associated with excessive daytime sleepiness (OR: 2.133, 95%CI: 1.508–3.018, p < 0.001), anxiety symptoms (OR: 2.766, 95%CI: 1.668–4.587, p < 0.001), and a history of mental disorders (OR: 2.527, 95%CI: 1.084–5.887, p = 0.032). Demographic variables independently associated with alcohol abuse/dependence included the age of 21–23 years (vs. 18–20 years, OR: 1.672, 95%CI: 1.155–2.422, p = 0.007) and higher monthly income (1,311–2,086 Chinese yuan [CNY] vs. under 615 CNY, OR: 2.367, 95%CI: 1.569–3.570, P < 0.001; > 2,086 CNY vs. 615 CNY, OR: 2.318, 95%CI: 1.048–5.127, P = 0.038).

Female medical students with alcohol abuse/dependence in our study showed higher prevalence of learning burnout, depression symptoms, anxiety symptoms, excessive daytime sleepiness and a history of mental disorders, as compared with those without alcohol abuse/dependence (all p < 0.05, Table 2). Inter-group differences were also significant in the proportion of monthly income and relationship status (all p < 0.05). Then a multivariate logistic regression was performed (Table 3). The results of the Hosmer–Lemeshow goodness-of-fit test showed that χ2 = 0.949, P = 0.813, indicating that the model fit was good. Multivariate logistic regression indicated that alcohol abuse/dependence was associated with learning burnout (OR: 2.168, 95%CI: 1.552–3.030, p < 0.001), anxiety symptoms (OR: 2.481, 95%CI: 1.645–3.741, p < 0.001), and a history of mental disorders (OR: 2.912, 95%CI: 1.607–5.277, p < 0.001) (Table 3). Demographic variables independently associated with alcohol abuse/dependence included having a partner (vs. single, OR: 1.788, 95%CI: 1.271–2.516, p = 0.001), and a monthly income over 615 CNY (616–1,310 CNY vs. under 615 CNY, OR: 1.656, 95%CI: 1.135–2.414, p = 0.009; 1,311–2,086 CNY vs. under 615 CNY, OR: 1.982, 95%CI: 1.273–3.086, p = 0.002; > 2,086 CNY vs. 615 CNY, OR: 3.357, 95%CI: 1.753–6.430, p < 0.001).

To our knowledge, this is the first study on the prevalence and gender-specific risk factors of alcohol abuse/dependence among Chinese medical students during the post-COVID‑19 pandemic period, which revealed that alcohol abuse/dependence was common among Chinese medical undergraduates. We also found that male medical undergraduates were more likely to report alcohol abuse/dependence than female students. Age and excessive daytime sleepiness were associated with alcohol abuse/dependence only among male students, whereas learning burnout and having a partner were associated with alcohol abuse/dependence only among female students. The above findings indicated the need for gender-specific screening and interventions for medical undergraduates to prevent alcohol abuse/dependence and its adverse consequences.

Approximately 10.3% of the medical undergraduates reported alcohol abuse/dependence, slightly higher than of previously reported in China (4.57%-8.25%) [9, 10]. Some possible reasons may contribute to these differences. On the one hand, the difference might have been partially attributed to the different methods used to assess alcohol abuse/dependence. The two previous studies used the Alcohol Use Disorders Identification Test (AUDIT) to detect alcohol abuse/dependence, whereas our study used the AUDIT-C. On the other hand, the two previous studies were conducted before the COVID-19 outbreak, while our investigation was conducted after the outbreak of this pandemic. Previous studies suggested the prevalence of alcohol-related illnesses in the general population has increased after the pandemic outbreak [62, 63]. One of the possible reasons for this phenomenon is the subjective stress and mental distress caused by the COVID-19 pandemic, which was associated with the increase of alcohol consumption [64, 65]. However, to our knowledge, the present study is the first to investigate alcohol abuse/dependence among Chinese medical students after the outbreak of the COVID-19. Therefore, further research is needed to determine whether pandemics may have a negative impact on alcohol use and related disorders among medical students. Finally, although only 10.3% of the medical students in this study reported alcohol abuse/dependence, Chinese undergraduate medical enrolment in 2020 is already over 2 million [66], which means that a large number of students are potential victims of alcohol abuse/dependence. Therefore, timely screening for alcohol use and prevention of alcohol abuse/dependence is warranted, especially after the outbreak of the pandemic.

The present study demonstrated that the prevalence of alcohol abuse/dependence is significantly higher in males than in females, which was similar to the findings reported in other countries [5, 67]. The gender differences in the prevalence of alcohol abuse/dependence may be driven by biological and socio-cultural (sex-related) factors. On the one hand, previous studies have found that men release more dopamine than women when exposed to similar levels of alcohol, which is strongly associated with alcohol abuse/addiction [68, 69]. On the other hand, traditional gender roles have led to a tolerant attitude towards male alcohol consumption, i.e., heavy alcohol consumption by men is often viewed as a higher level of masculinity but alcohol use by women is regarded as indiscretion [70‐72]. Although emerging evidence have shown that gender differences in problematic alcohol use have narrowed in the general population as gender roles have become more equal [24], our study still showed that male medical undergraduates are nearly 3 times more likely to engage in alcohol abuse/dependence than their female counterparts. Therefore, it is important to prospectively track the prevalence of alcohol abuse/dependence in different genders.

A better understanding of gender-specific risk factors for alcohol abuse/dependence helps promote the wellbeing of medical students. Our findings suggest that older age and excessive daytime sleepiness were strongly associated with alcohol abuse/dependence in male medical students. Some other studies also suggested that the frequency and amount of alcohol consumption increased with age [38, 73, 74]. The possible reason for this is that as they get older, females and males are under increasing pressure to conform to culturally sanctioned gender roles [75]. In China, traditional gender roles tend to discourage or restrict female drinking, but accept and even expect male drinking [71]. Therefore, males may consume more alcohol as they age, thereby increasing the risk of alcohol abuse/dependence, suggesting that more attention needs to be paid to alcohol abuse/dependence among older medical students, particularly male students. In addition, after controlling for demographic variables, only excessive daytime sleepiness was independently associated with alcohol abuse/dependence among male medical students but not their female counterparts in our study. One possible reason for this is that men tend to self-medicate with alcohol more often than women when they have sleep-related problems [76]. Both this study and some previous studies have shown that over one third of medical students have experienced excessive daytime sleepiness [77‐79]. Therefore, excessive daytime sleepiness among medical undergraduates needs more attention, and the improvement of sleep quality may help them reduce their problematic alcohol use, especially for male students.

The present study also showed that having a partner and learning burnout were related to alcohol abuse/dependence among female medical students. Similarly, previous study indicated that alcohol abuse/dependence was more likely to occur in physicians who had a partner [18]. However, this association was not found among male medical students in our study. Previous studies suggested that female surgeons were more vulnerable to work-family conflicts, which appeared to be associated with their depression and alcohol abuse [80]. Female medical students with a partner might also be more vulnerable to the imbalance between study and relationship, which might be the reason for their alcohol abuse/dependence. Furthermore, our study found that only female medical undergraduates showed strong association between alcohol abuse/dependence and learning burnout. This was consistent with findings of other studies, which showed a significant association between the habit of drinking once or twice a week and burnout in female medical students only [81]. Notably, although a significantly higher prevalence of learning burnout was observed in male rather than female medical undergraduates, which was consistent with other studies [81‐83], no significant association was found between alcohol abuse/dependence and learning burnout among male students. This might be due to the gender difference in motivations for drinking, e.g., women were more likely to drink due to coping issues while men were more likely to drink for social and recreational reasons [84, 85]. Thus, female medical students may be more likely to use alcohol as a coping strategy for learning burnout. Based on our findings, reducing learning burnout and improving the ability to cope with unpleasant emotions might be helpful to address the issue of alcohol abuse/dependence among medical students, especially for females. As we did not directly assess the motivations of alcohol drinking, we could only speculate that motivations of alcohol drinking might be the source of gender differences in alcohol abuse/dependence and deteriorating mental health. Nevertheless, the causality of this association can only be determined by further longitudinal studies.

There are several limitations to our study. First, this is a cross-sectional survey, which has limited the elucidation of the causality. Thus, future longitudinal studies are warranted. Second, this survey used convenience sampling and all the data were self-reported, which might have led to report bias; however, this might have been mitigated by the anonymity of the participants, as studies have shown that data from responders can be valid and reliable when their privacy is protected [86]. Third, this study lacked comparable pre-pandemic alcohol abuse/dependence among medical students to make direct comparisons. Finally, we used self-reported scales instead of standardized diagnostic tools to measure alcohol abuse/dependence and other mental illness, which may have created a potential bias that needs to be improved in future studies.