Are doctors feeling burnout? – Multicenter cross-sectional study on burnout syndrome and its determinants during the ongoing pandemic

Abstract

Background: 

Though the concept of burnout has been around for long, its significance is increasing nowadays owing to the demanding nature of jobs. The latest ICD-11 also provides a detailed description of Burnout syndrome. Physicians are at high risk for experiencing burnout and this becomes especially relevant in the ongoing COVID-19 pandemic.

Aim: 

To determine the risk of burnout among medical faculty and its predictors, if any.

Materials and Methods: 

This was a multicentric cross-sectional study that included medical faculty from four tertiary care government teaching hospitals in north India. A survey was conducted during the current COVID-19 pandemic to assess burnout using a structured online questionnaire based on Burnout Assessment Tool. The questionnaire also included relevant socio-demographic, professional, health, and lifestyle-related details. Descriptive statistics, Mann-Whitney U Test/Kruskal Wallis Test, and Kendall’s tau-b Test were used for statistical analysis.

Results: 

A total of 244 medical faculty completed the survey. 27.87% were at risk of burnout, out of which 11.89% were at a very high risk of burnout. Dissatisfaction with the job and dissatisfaction with sleep (P < 0.01 for both) were associated with greater burnout scores and a greater risk of burnout.

Conclusion: 

Faculty members are at high risk of burnout, regardless of sociodemographic and work-related factors.

INTRODUCTION

Burnout is included by World Health Organisation (WHO) in the 11^th Revision of the International Classification of Diseases (ICD-11) as an occupational phenomenon and is conceptualized as a syndrome resulting from chronic workplace stress that has not been successfully managed. It is characterized by - feelings of energy depletion or exhaustion; increased mental distance from one’s job, feelings of negativism or cynicism related to one’s job; and reduced professional efficacy.^[1] Burnout is rampant across professions, work settings, and geographical areas. Although recently gaining popularity, its history began long back when Herbert Freudenberger introduced the term burnout in the 1970s, ushering research in the field.^[2]

The stressful, draining, and resource-constrained nature of work puts healthcare professionals at extreme risk of experiencing burnout.^[3] Burnout is known to be more common among physicians compared to the general population.^[4,5] Rotenstein and colleagues in their systematic review published in 2018, noted that the prevalence of burnout in physicians ranged anywhere from 0% to 80.5% in the 122 studies which had reported burnout prevalence.^[6] A more recent systematic review including 15 studies and 3845 healthcare professionals from India reported pooled prevalence of burnout between 23-27%.^[7] The COVID-19 pandemic has tested healthcare systems throughout the world and India is no exception. Consequently, physician burnout has reportedly increased during the COVID-19 pandemic, with various pandemic-related factors playing a mediating role.^[8–10]

In the Indian health care system, apart from serving as prime centers of patient care, medical colleges impart training to future doctors. Therefore, burnout symptoms in medical faculty can additionally compromise the training of young doctors ultimately affecting the society they serve. Surprisingly, there are very few studies evaluating burnout symptoms in this discrete population. With this backdrop and considering the paucity of data from India, this study was undertaken to determine the risk of burnout among medical faculty and its predictors.

MATERIALS AND METHODS

This was a multicentric cross-sectional study that included medical faculty from four tertiary care government teaching hospitals in north India and was conducted between January 2022 to May 2022, after the country had experienced three prominent COVID-19 waves. The study population comprised medical faculty members who had completed their postgraduate training and were employed as faculty members (Assistant Professors, or higher) in different departments of four tertiary care government teaching hospitals in north India. Non-faculty and other health personnel like residents, nursing staff, paramedics, and auxiliary staff were excluded from the study.

A pre-tested structured questionnaire in the English language was prepared as a “Google Form”. The study questionnaire included socio-demographic, professional, and lifestyle-related queries as well as questions from the Burnout Assessment Tool (BAT).^[11] The survey questionnaire took approximately 15 minutes to complete. Google Forms is an internet-based service that allows data collection via personalized surveys without levying any charges. The form can be shared by sending the link through e-mail or WhatsApp to respondents, who can fill the form by logging into their e-mail ids on a computer, laptop, tablet, or smartphone. The collected information is stored on the Google Drive of the form owner.

The questionnaire contained a statement of consent and anonymity, information about the study, and instructions on how to complete the survey at the beginning. It was further divided into three sections. Section one consisted of the socio-demographic and profession-related queries. Section two contained questions relating to health, lifestyle, COVID-19 exposure, consequent hospitalization of the respondent or a family member, and the impact of COVID-19 on working hours and workload. Section three contained questions from BAT.^[11]

BAT is the latest tool for assessing burnout that addresses most shortcomings of earlier tools while maintaining reliability and validity with other burnout measures like the Maslach Burnout Inventory (MBI).^[12,13] The shorter work-related version of BAT with 12 questions was used in this study (BAT-12). It has four subscales - exhaustion, mental distance, cognitive impairment, and emotional impairment, that can be interpreted separately or together as a composite score. Compared to other measures of burnout, it provides clear cut-offs for those at no risk of burnout, those at risk for burnout, and those at very high risk for burnout, overall as well as in each of the subscales. The answers to questions are scored between 1 to 5. Three categories are identified for the composite score: no risk of burnout (1.00 – 2.53), at risk for burnout (2.54 – 2.95), and at very high risk of burnout (2.96 – 5.00) based on the total scores obtained.^[11]

A message with the Google form link was circulated via different social platforms to medical faculty members. It was also recirculated at regular intervals to prompt response. The filing of the form was voluntary. The study tool was so designed that only one response was possible per participant. Once the time frame of the study was reached, the Google form link was deactivated.

A similar study conducted exclusively on medical faculty members in India reported the prevalence of burnout in at least two domains out of the total three domains of MBI as 19%.^[14] Since our scale provides a composite score after measuring burnout in four subscales, we used a more balanced approach by taking the prevalence of burnout in at least two of the three domains of MBI rather than all three domains for sample size estimation and also because MBI doesn’t provide a composite score of the three domains. Using the formula Z_1-α/2² p (1-p)/d² for sample size estimation with an alpha error of 5% and an absolute precision of 5%, the calculated minimum sample size required was 237. With an additional 10% allowance for invalid/incomplete entries, the sample size was estimated as 264.

The automatically filled dataset was verified, validated, and analyzed on SPSS v29.0. Descriptive statistics were applied and categorical data were presented as percentages. Mann-Whitney U Test/Kruskal Wallis Test was used to evaluate differences between groups. Kendall’s tau-b test was utilized to find the correlation between burnout scores and other variables under study. A P value less than 0.05 was considered to be statistically significant.

The study was approved by the Institutional ethics committee (IEC No – SNMC/IEC/2021/56).

RESULTS

We received a total of 270 responses and after removing invalid/incomplete entries a total of 244 responses were considered for statistical analysis. Half (n = 122, 50%) of the faculty members were 41-50 years of age, followed by those aged less than 40 years (n = 84, 34.43%). More than half (n = 140, 57.38%) of the respondents were males. The majority (n = 231, 94.67%) of the respondents were married. The highest educational qualification attained was MD/MS or equivalent for the majority of participants (n = 216, 88.52%). There was nearly an equal distribution between the number of respondents from medicine and allied specialties (n = 93, 38.11%), paraclinical/preclinical departments (n = 77, 31.56%), and surgery and allied specialties (n = 74, 30.33%). More than three-fourths of the participants were permanently employed (n = 192, 78.69%). Most were employed for either <10 years (n = 131, 53.69%) or for 10-20 years (n = 96, 39.34%) in their current jobs. About one-third of faculty members were Assistant Professors (n = 93, 38.11%) or Professors (n = 75, 30.74%), followed by Associate Professors (n = 47, 19.26%). Also, 29 (11.89%) were serving as Head of Department/Principal/Dean. About two-thirds (n = 161, 65.98%) of faculty members worked 8-12 hours/day, while a quarter worked less than 8 hours/day (n = 62, 25.41%). Strikingly, 21 (8.61%) of the faculty members worked 12 hours or more in a day. The spouses of the majority of respondents (n = 194, 79.51%) were working, out of whom most (n = 171, 70.08%) were medical professionals. More than half (n = 144, 59.02%) of respondents had 2 children and were living in a nuclear family (n = 159, 65.16%). 43 (17.62%) respondents had hypertension, 26 (10.66%) had diabetes, and 28 (11.48%) had a thyroid disorder [Table 1].

Table 1: Sociodemographic characteristics of the study population

The majority of the respondents (n = 194, 79.51%) had direct exposure to COVID-19 cases. Nearly an equivalent percentage (n = 181, 74.18%) had been infected at least once with COVID-19. Interestingly, a quarter (n = 63, 25.82%) of faculty members had never been infected. Only 13 (5.33%) of the respondents required hospitalization for COVID-19 but 93 (38.11%) of the respondents had some family member hospitalized for COVID-19. A large proportion of faculty members had their workload increased (n = 218, 89.34%), and even more had their work hours affected (n = 226, 92.62%) during the pandemic. 48 (19.67%) faculty members reported dissatisfaction with their job whereas 196 (80.33%) were satisfied or very satisfied with their job. One out of every four respondents (n = 61, 25%) was not satisfied with sleep (sleep duration of more than or equal to 7 hours and feeling fresh after sleep). Only half of the respondents were exercising regularly (at least 5 times/week for 30 minutes) (n = 120, 49.18%) or taking sufficient portions of fruits/vegetables (at least 5 servings/day) (n = 128, 52.46%) [Table 2].

Table 2: COVID-19-related, work-related, and lifestyle-related characteristics of the study population

More than a quarter of faculty members (n = 68, 27.87%) were at risk of burnout. Furthermore, 29 (11.89%) were at a very high risk of burnout. About a quarter (n = 58, 23.77%) reported exhaustion, half (n = 120, 49.18%) reported mental distance, and one-third (n = 81, 33.20%) reported emotional impairment. Only 33 (13.53%) of faculty members reported cognitive impairment [Table 3].

Table 3: Burnout-related characteristics of the study population

After applying Mann-Whitney/Kruskal Wallis, statistically significant differences in burnout scores were noted for all three categories of job satisfaction (P < 0.01) and satisfaction with sleep (P < 0.01). The median BAT scores were highest in those who reported being “not satisfied” with their job, followed by those who reported being “satisfied”. The lowest median BAT scores were found in those who were “very satisfied” with their job. Similarly, higher median BAT scores were noted in those reporting not being “satisfied with sleep” compared to those reporting satisfaction with sleep. A statistically significant correlation was also noted between burnout scores and satisfaction with job and satisfaction with sleep using Kendall’s Tau-b test. Dissatisfaction with the job and dissatisfaction with sleep resulted in higher burnout scores. No statistically significant differences were noted in terms of burnout scores for age, gender, marital status, education, sociodemographic, institution or employment-related factors, work-related factors, COVID-19-related factors, and lifestyle-related factors [Tables 4 and 5].

Table 4: Mann-Whitney U Test/Kruskal Wallis Test to compare the burnout scores against various variables

Table 5: Kendall Tau-b test for correlation between BAT scores and different variables

DISCUSSION

This study was novel in its attempt to determine burnout risk in a large exclusive sample of faculty members of four tertiary care teaching hospitals in north India. Unlike most other studies, our study included only those doctors who were working as medical faculty in teaching hospitals. In addition, most studies have either not used a standardized tool/assessment method or have used a tool that doesn’t provide specific cut-off scores for identifying burnout or its risk. Keeping these lacunae in mind the researchers felt that it is important to assess burnout risk in the medical faculty.

We found more than a quarter of faculty members (27.87%) to be at risk of burnout, out of whom 11.89% were at a very high risk of burnout. Wide variations exist in the reported prevalence of burnout symptoms. Glisch et al. and Glasheen et al. using a single self-reported question, reported burnout in 62% and 23% of academic hospitalists, respectively^[15,16] while Hoff et al. found that 13% of medical faculty had burnout and an additional 25% were at risk of burnout.^[17] Hinami et al. using the single-item measure of burnout noted burnout in 30% of medical faculty.^[18] A study by Khan et al. in Canada during the current pandemic noted the prevalence of burnout in physicians was 68%.^[19] Studies from India have also shown a high prevalence of burnout in medical faculty. Krishna et al. reported that 25.5% of academic hospitalists had burnout symptoms in any one dimension of MBI, whereas Chichra et al. found that 19.1% of the faculty members scored high on two or more domains of the MBI.^[14,20] In our study, using BAT we noted that about a quarter (23.77%) of faculty members reported exhaustion, half (49.18%) reported mental distance (49.18%), and one-third (33.20%) reported emotional impairment. Only 13.52% of faculty members reported cognitive impairment. In the study by Hashmi et al. 38.9% of respondents scored high on emotional exhaustion and 31.5% scored high on depersonalization.^[21] Chichra et al. reported that 29% of faculty scored high on the emotional exhaustion subscale and 20.8% on the depersonalization subscale while Khan et al. reported the prevalence of emotional exhaustion as 63% and depersonalization as 39%^[14,19]. Hence, when similar studies in the literature are examined, the rates of overall burnout and its subdomains appear to be similar to our study.

The obvious high variability in the reported prevalence of burnout among different studies reflects not only differing workplace environments but variations in study designs and assessment methods as well. This was also concluded by Rotenstein et al. in their systematic review who were unable to provide a pooled prevalence for burnout due to differing methods of burnout evaluation.^[6] By and large, the high rates of burnout in faculty members can be explained by the high degree of human interaction needed- with patients, trainees, other health care staff, and administrators; high expectations; being overworked with poor work-life balance; lack of time for self, leisure or recreational activities; less than expected remuneration; non-recognition of efforts; lack of support; being responsible for the health and wellbeing of others, patients’ and attendants’ hostile behaviors and complaints, and dealing with ailments, suffering, and death. Somewhat lower rates of burnout in our study may be because the faculty members in our centers typically have a uniform pre-decided schedule, and don’t have to deal with health records and insurance/reimbursement. The prevalence of burnout in our study is similar to studies completed before the pandemic in similar settings suggesting only a minimal effect of the COVID-19 pandemic. Probable justification is that those who choose the medical profession and work in a teaching hospital, are already aware and ready for the hardships of their job and become adapted to them. Another possible reason is the rigorous and demanding training that most faculty members undergo during their early career, especially in residency provides a resilient frame against burnout. Another explanation for getting rates of burnout similar to pre-COVID studies might also simply indicate that by the time of data collection most of the workplace changes forced by the pandemic had already reverted or the workplace changes were being better managed and organized. In line with this, a recent longitudinal study on the Finnish population noted no significant change in burnout scores over the pandemic and its multiple waves.^[22]

In our study, 80.33% of faculty members were satisfied or very satisfied with their job. Glisch et al. and Hoff et al. have reported 74% and 91.8% of academic hospitalists are satisfied with their jobs.^[15,17] Chichra et al. reported high job satisfaction in 71.4% of medical faculty while Chaudhuri et al. in their study conducted during the current pandemic reported that 85.58% of doctors in a medical college had high to average job satisfaction.^[14,23] The satisfaction with the job appears to be similar across studies and roughly corresponds to the burnout rates in most studies as well as in our study which is expected as burnout is a syndrome related to workplace stress and job satisfaction. About a quarter of respondents (25.41%) in our study were not satisfied with their sleep. Batra et al. in their metanalysis calculated the pooled prevalence of insomnia in doctors as 39.1% and the overall prevalence of impaired sleep quality as 64.3%.^[24] The high prevalence of sleep problems likely suggests a possible relationship with burnout.

Statistically significant relationships with BAT scores were found for satisfaction with the job (P < 0.01) and satisfaction with sleep (P < 0.01). This does not imply any cause-effect relationship as this is a cross-sectional study rather this simply reflects a statistically significant relationship between these variables and burnout risk. In line with findings from our study, Chichra et al. found low job satisfaction to be related to higher levels of burnout in medical faculty.^[14] Glisch et al. found no significant difference in burnout prevalence with gender, career length as a hospitalist, or job satisfaction as a hospitalist.^[15] Krishna et al. reported that burnout was significantly higher in those ≤35 years, but found no association with gender, medical or surgical branches, working hours/day, work experience, and marital status.^[20] Predictors of burnout in the study by Glasheen et al. were low satisfaction with the amount of personal/family time and low satisfaction with control over the work schedule.^[16] Khan et al. in their study during the current pandemic noted the highest burnout in those 36-50 years of age which became lower with increasing age. They also found longer clinical hours to be associated with burnout but noted no association of burnout with caring for patients having COVID-19, being female, and the number of children.^[19] In a review of studies from China, it was suggested that burnout was associated with low job satisfaction at the individual doctors’ level.^[25] The lack of a significant relationship between the BAT scores and sociodemographic, professional, and COVID-19-related variables in our study perhaps implies much more complex interactions between these factors that are not readily explained by simple assumptions. This complex relationship is further substantiated by the observation that various authors have found different sociodemographic, professional, and pandemic-related factors to be associated with burnout, and often these don’t correspond across studies.^{[9,10,25–28]} The positive relationship between job satisfaction and burnout hints at a reciprocal relationship between them further strengthening and validating the findings of this study. This also validates that the exhaustion, mental distance, cognitive impairment, and emotional impairment-related symptoms enquired by the BAT tool used in the study truly represent work-related burnout symptoms. Thus, we infer that our study has been able to capture the burnout rates effectively while eliminating potential confounders. Moreover, the interdependent relationship between sleep and burnout found in our study suggests a bidirectional link between them, where each may compound the effects of the other ultimately leading to a vicious cycle. Burnout and sleep-related problems are likely to co-occur, and sleep disturbances are harbingers of various mental disorders.

Therefore, we conclude that medical faculty is at risk of burnout, and acknowledging this will help to tackle this problem effectively. Burnout and related mental health problems impact the medical care provided by physicians as well as their career-related decisions.^[29–31] Consequences of burnout include medical errors, decreased patient satisfaction, absenteeism, decreased job satisfaction, turnover, and cynicism. It may also result in anxiety, depression, substance abuse, and relationship problems.^[32] Hence, considering the profound implications of burnout on mental health and the practices of teaching and treating, creating awareness and addressing burnout-related issues at both the organizational and individual levels is essential. The specific organizational strategies practiced and promoted by the Mayo Clinic are – acknowledging and assessing the problem, harnessing the power of leadership, developing and implementing targeted interventions, cultivating community at work, using rewards and incentives wisely, aligning values and strengthening culture, promoting flexibility and work-life integration, providing resources to promote resilience and self-care, and facilitating and funding organizational science.^[33] Simple practices like promoting autonomy, increasing physicians’ participation in organizational decisions, giving physicians the ability to influence their work environment and control their time schedules, providing adequate office resources and support staff, and, facilitating a collegial work environment can go a long way in decreasing burnout. Individual-focused interventions include self-care workshops, stress management skills, communication skills training, a balanced diet, sleep hygiene, yoga, mindfulness, and meditation.^[34] With these measures, sleep satisfaction and job satisfaction are likely to improve along with the decrease in burnout risk. Often, the best interventions are probably those that consider specific factors in a particular setting and take into account individual, interpersonal, and organizational factors.

The study has some inherent limitations. Being a subjective, self-reported online survey there is the issue of non-disclosure or selective disclosure, objectivity, reproducibility, and response bias. The authors also acknowledge the subjectivity in answering various questions including questions about job satisfaction, sleep satisfaction, practicing exercise/yoga, and fruit/vegetable intake. Not all individuals have responded to the questionnaire, resulting in some degree of sampling bias. The survey is cross-sectional and responses can be affected by transient day-to-day stress and events. The cross-sectional design also precludes any evaluation for causality. Other mental illnesses that ICD-11 itself mentions as exclusions - Adjustment disorders, Disorders specifically associated with stress, Anxiety, or fear-related disorders, and Mood disorders weren’t ruled out. Different definitions and methods for assessing burnout have been used by different authors and this may at times translate to non-comparability. It would also have been better if certain other factors like choice of specialty, work-life balance, satisfaction with remuneration, and intent to leave were assessed.