Study of burnout and depressive symptoms in doctors at a central level, state hospital

Background Doctors are at high risk of burnout, which has far-reaching consequences on an individual and organisational level. Several studies have shown an association between burnout and depression. Aim This study aimed to determine the rate of burnout and depressive symptoms among doctors, as well as factors associated with both conditions. Setting Charlotte Maxeke Johannesburg Academic Hospital. Methods Burnout was measured using the Maslach Burnout Inventory–Human Services Survey and defined as the total score of high emotional exhaustion (≥ 27 points) + high depersonalisation (≥ 13 points). Individual subscales were analysed separately. Depressive symptoms were screened using the Patient-Health Questionnaire-9 (PHQ-9) and a score of ≥ 8 was deemed indicative of depression. Results Of the respondents (n = 327 for burnout and n = 335 for depression), 46.2% screened positive for burnout, whilst 53.73% screened positive for depression. Factors associated with increased burnout risk were younger age; Caucasian race; internship and/or registrarship; the discipline of emergency medicine; and having a prior psychiatric diagnosis of depressive and/or anxiety disorder. Factors associated with increased risk of depressive symptoms were females; younger age; being an intern, medical officer or registrar; disciplines of anaesthetics and obstetrics and gynaecology; having a prior psychiatric diagnosis of depressive and/or anxiety disorder; and family history of psychiatric disorder. Conclusion A high rate of burnout and depressive symptoms was determined. Although there is an overlap between the two conditions in terms of both symptomatology and risk factors, specific risk factors were determined for each in this population. Contribution This study highlighted the rate of burnout and depressive symptoms experienced by doctors at the state level hospital necessitating individual and institutional interventions to address this.


Introduction
The term 'burnout', first used by Herbert Freudenberger in 1974, 1 described strain arising from prolonged exposure to chronic, job-related stress. In addition, Maslach and Schaufeli supported this in their findings. 2 Maslach and Jackson 3 redefined the concept as a syndrome consisting of three dimensions: • High emotional exhaustion (EE) -chronic fatigue and low motivation • High depersonalisation (DP) -a feeling of distance from one's job, cynicism and negativism • Low personal accomplishment (PA) Further research by Maslach led to the development of the Maslach Burnout Inventory-Human Services Survey (MBI-HSS), which is considered the gold standard measure for assessing burnout. 4 Healthcare professionals are considered vulnerable to burnout with rates varying between 21% and 100% amongst South African doctors. 5,6,7,8,9,10,11 Internationally, the 2019 Medscape National Physician Burnout, Depression and Suicide Report quotes burnout prevalence rates of 44% 12 especially in disciplines of urology, neurology, internal medicine, surgery and anaesthetics. Similarly, a European-based systematic review found that certain surgical specialities and oncology personnel were more likely to experience burnout. 13 Bureaucratic tasks, followed by long working hours, were the largest contributors resulting in higher dissatisfaction at work. 14 Many studies explored the risk factors contributing to burnout, which may be divided into three domains: 13,14,15,16,17 • Occupational factors -discipline, job position (registrarship in particular), clerical work, working hours, high patient load and risk of litigation • Physicians' attributes -younger age, female sex, certain personality traits (e.g., neuroticism) and maladaptive coping skills • Organisational factors -poor leadership, hostile work environment and lack of peer support The consequences of burnout can generally be divided into four domains: 13,18,19,20 • Physical consequences -coronary heart disease, chronic fatigue and increased risk of early mortality • Psychological consequences -depressive symptoms, substance abuse and increased risk of suicide • Professional consequences -absenteeism, medical errors, less productivity and resignations • Organisational consequences -financial losses incurred in training and recruitment because of clinicians leaving the organisation, as well as malpractice costs Burnout is often associated with depression with considerable overlap of symptoms 20 ; however, burnout is directly linked to work-related stressors, whereas depression may occur in the absence of work stress. 19 As a result of the overlap, correlations between measures of the two constructs 16 may be artificially inflated. The overlap may be more profound and reflect burnout as a depressive condition. 21 According to most researchers, burnout and depression are independent entities, differentiated by symptoms specific to depression, such as anhedonia, inappropriate guilt and definitive mood change. 22 An international systematic review estimated the prevalence of depressive symptoms among registrars as 28.8% 20 compared with South African rates of 30% -40.7%. 21,22,23 According to the American Foundation for Suicide Prevention, training physicians are at an increased risk of suicide, with a fourfold increased risk of suicidal ideation during the first three months of internship. 24 Twenty-eight per cent of registrars experience a major depressive episode during training, compared with the general population rate of 7% -8%. According to Medscape's 2019 report, 14% of physicians admitted to previous suicidal ideation and 1% had attempted suicide. 12 Despite prevalence rates of 44% for burnout and 4% for clinical depression among physicians, only 13% were seeking professional help.
Considering these alarming figures, this study aimed to determine the rate and associated factors for burnout and depressive symptoms among doctors employed at Charlotte Maxeke Johannesburg Academic Hospital (CMJAH). We hypothesised that the burnout and depression rate would be elevated (± 50% of doctors) with an overlap of risk factors for each condition.

Study design
This cross-sectional analysis of burnout and depression among doctors using the MBI-HSS for Medical Personnel and Patient-Health Questionnaire-9 (PHQ-9) was conducted between 22 November 2019 and 29 February 2020.

Study setting
Charlotte Maxeke Johannesburg Academic Hospital is a state central level academic institution in Johannesburg, Gauteng, South Africa employing doctors across multiple disciplines and levels of training. All doctors in full-time employment (interns, medical officers, registrars and specialists) were approached to participate in the study.

Measures
Once informed consent was obtained, paper-based questionnaires (MBI-HSS and PHQ-9) and a data collection form were self-reported and deposited into sealed, secure collection boxes that were collected by the primary investigator. Data were then recorded in an Excel spreadsheet. The analysed variables included sex, age, ethnicity, relationship status, job position, discipline, number of hours of overtime per week, number of hours remunerated work outside the public service (RWOPS) per week, academic and/or study time per week, hours per week spent on clerical work, psychiatric diagnosis, family history of psychiatric disorder, psychotherapy attendance, psychotropic medication and/or substance use in the past month and lifetime medication and/or substance use. An instrument was not scored if three or more items were missing; therefore, 327 MBI questionnaires and 335 PHQ-9 surveys were scored. Categories with n < 15 were not analysed.

Data analysis
Burnout is defined as a high range score (≥ 27 points) on the EE subscale, a high range score (≥ 13 points) on the DP subscale and a low range score (≤ 31 points) on the PA subscale. 4 Further definitions offered by Brenninkmeijer and VanYperen 25 include the 'Emotional Exhaustion + 1' principle which states that a person can be diagnosed as being clinically burnt out with a high EE score plus either a high DP score or a low PA score. This definition encompasses a psychological state of sufficient severity to support a diagnosis of burnout.
For the purpose of this study, the 'Emotional Exhaustion + 1' principle 25 was applied in defining burnout. Burnout was, therefore, defined as high EE (≥ 27 points) + high DP (≥ 13 points). The authors favoured this definition considering the participants different levels of experience. Furthermore, this was carried out noting that EE has been identified as the category that mostly indicates burnout. Individual measures of high EE, high DP and low PA (≤ 31 points) were also examined. A score of 27 and above in the EE subscale was considered high, 19-26 moderate and less than 19 low. A score of 10 and above in the DP subscale was considered high, 6-9 moderate and less than 6 low. A score of 40 and above in the PA subscale was considered high, 34-39 moderate and less than 34 low. 26 The PHQ-9 has acceptable diagnostic properties at a range of cut-off scores (8)(9)(10)(11). 27 A score of ≥ 8 on the PHQ-9 was deemed indicative of depression. 27 The association between risk factors and overall burnout and depressive symptoms was assessed by univariate binomial regression analysis. Factors with p < 0.20 were combined into a multivariable model, after examining each pair of variables for possible confounding using the chisquared test (or Fisher's exact test for 2×2 tables); a value of Cramer's V (or the phi coefficient for Fisher's exact test) > 0.50 was regarded as too strong an association to include both variables in the multivariable model. Non-significant variables were removed sequentially from the multivariable model to derive a final model. Data analysis was carried out using SAS version 9.4 for Windows. A 5% significance level was used.

Ethical considerations
The study was approved by the University of the Witwatersrand, Human Research Ethics Committee. A distress protocol, detailing contact information of available resources for psychological and psychiatric care, was provided to all participants.

Findings
Doctors in full-time employment at CMJAH over the study period were 694; 337 doctors for whom either the burnout or depression scale(s) could be scored participated   Table 2 continues on te next page→ in the study (overall response rate of 48.6%; margin of error 3.8%). The response rate for interns and registrars was significantly higher than other groups. A total of 58.8% (n = 198) of participants were female and 41.2% (n = 139) male ( Table 1). Characteristics of the study participants are further described in Table 2. Variables were analysed for their association with burnout and depression (Tables 3 and 4).

Burnout
The total number of doctors classified as experiencing burnout (high EE + high DP) was 151 (46.2%). Associations of statistical significance included age, job position, discipline, psychiatric diagnosis, the use of medication and/or substances and not performing RWOPS.
Risk factors for burnout can be divided into three sections: • Individual factors -Female sex was not associated with overall burnout, but they had a higher risk of high DP and low PA; younger doctors (20-39 years) displayed a higher risk of burnout (87.4% of the study population) and high DP; Caucasians showed an increased risk of burnout and high DP. • Occupational factors -Participants who screened positive for burnout constituted registrars (49.7%) and interns (19.9%); with interns, registrars and medical officers being at higher risk of high EE and high DP but not low PA. Burnout risk was increased, although not statistically significant, in emergency medicine (RR 1.65; CI 1.14-2.40, p-value 0.01) and pathology (RR 1.38; CI 0.97-1.97, p-value 0.07); high DP was associated with emergency medicine, pathology and obstetrics and gynaecology and low PA was associated with emergency medicine, public health, pathology and anaesthetics. Overtime was associated with high EE values but with an increased risk of overall burnout. Those who did RWOPS (mainly specialists) displayed lower risks of burnout (RR 0.64; CI 0.42-0.97, p-value 0.03) with a low/moderate EE and DP. Although academic and/or study hours were associated with high DP, it was associated with an increased risk of overall burnout. Longer clerical hours were associated with higher EE but not with an increased risk for burnout. • Psychiatric factors -Psychiatric diagnosis was associated with an increased risk of burnout (RR 1.  Table 3 shows the variables examined and their relative risk of high EE, high DP and low PA.

Depression
The number of participants who screened positive for depressive symptoms was 180 (53.73%) ( Table 5).

Burnout
Of the participants, 46.2% of doctors employed at CMJAH had burnout. This figure is in keeping with international rates 12 , but it is lower than rates found in South African studies. 8,9,21 This could be attributed to a larger sample size in this study, with a broader range of disciplines and ranks of doctors assessed. We also utilised a definition of burnout that is a combination of the EE and DP subscales, whereas the majority of other SA studies considered individual subscales of burnout.  Those aged 20-39 were more likely to suffer burnout. Internship and registrarship in South Africa mostly commence within this age group and interns were found to be three times more likely to be burnt out than specialists. Common psychiatric disorders also tend to have their onset in this age group 28 ; therefore, there may be an overlap in risk factors between burnout and mental illness. Factors contributing to results in this group may be organisational factors, clinical and clerical work demands and medical training insufficiently preparing doctors for the work environment, thus contributing to dysfunctional adjustment. Studies have indicated registrar training as a vulnerable time in which burnout is more likely to develop, 13,16,20,29 possibly because of stressors of the academic training requirements, frequent rotations between training institutions and transitioning into more senior roles in the work environment.
Emergency medicine and pathology doctors were shown to be at a higher risk of burnout, whereas international data show that urology, neurology, internal medicine, surgery and anaesthetics had the highest burnout rates. 12 The increased risk in emergency medicine in this setting may be because of the high output, community serviced by the hospital, crisis management service inherent to the discipline or because of specific departmental factors that need further exploration. Other studies have demonstrated that paramedic and emergency medicine personnel have high rates of burnout. 9,30,31 Those who performed RWOPS were at a lower risk of burnout. In South Africa (SA), only specialists and medical officers are legally allowed to perform RWOPS. Although RWOPS performance increases work hours in this study, it was not found to increase the risk of burnout. This finding, however, may be confounded by a lower risk of burnout in specialists because of other factors, such as older age and less study time. One could also argue that while RWOPS increases time spent at work, remuneration for this work may decrease financial stress. Institutional factors may need to be examined in greater depth to account for this, for example, a lack of human and infrastructural resources may be more significant in the public versus the private healthcare sector.
Doctors with a psychiatric diagnosis (depression and anxiety) displayed a higher risk of burnout with the risk increasing during and after internship. This could be explained by various risk factors for mental illness during the period of Radiology and nuclear medicine -

Clerical hours
More Depression and/or anxiety diagnosis -

Recent medication use
Antidepressant and/or anxiolytic

Past medication use
Antidepressant and/or anxiolytic RWOPS, remunerated work outside the public service; EE, emotional exhaustion; DP, depersonalisation PA, personal accomplishment.
internship, as outlined here. Whilst we observe the association between mental illness and burnout, it is difficult to statistically surmise causality from this finding, that is, if mental illness is an independent risk factor for burnout or if burnout predisposes one to develop mental illness.
Over 60% of doctors assessed as burnt out had never attended psychotherapy. There was a significant association between burnout and having past or present psychotherapy. Doctors may engage with psychotherapy for the management of depressive or anxiety disorders or to improve coping skills related to innate personality factors. It should be observed that the association strengthens with present psychotherapy. We can postulate that the current working environment exposes a need to engage in help-seeking behaviour, whether it is because of burnout itself or other psychiatric illness.
There was a significant association between burnout and the use of recent antidepressants and/or anxiolytics, benzodiazepines and stimulants, as well as for past use of antidepressants and/or anxiolytics and benzodiazepines. This may be because of either: (1) prescribed medication for a depressive and/or anxiety disorder or (2) self-medicating symptoms of either burnout or depression. The latter may be more likely as the association of benzodiazepine use and stimulant use versus antidepressant use is higher.
Benzodiazepines are known for their anxiolytic and hypnotic properties, whereas stimulants may be utilised to manage symptoms of fatigue or improve alertness. This finding also implies that stimulant use began or increased after internship. This is an interesting finding, as the number of doctors who, according to our survey, had a diagnosis of 'other' was 9 out of 337 (2.7%) (Attention deficit hyperactive disorder [ADHD], the most common condition, for which stimulants such as methylphenidate are prescribed, would be included here).
There was no significant association between burnout and sex, relationship status, ethnicity, overtime per week, academic hours per week, clerical hours per week and family history of psychiatric disorder. Some of these findings differ from what has been reported in the literature; it has been reported that females, clerical work and academic or study time have also been cited as risk factors of burnout. This study showed that females are at increased risk of depressive symptoms (in keeping with accepted risk factors for depression 32,33 ) but not burnout. This implies that, although burnout and depression may have some common risk factors, sex may not be one of them in this population and it is in keeping with findings from a 2012 systematic review. 13 Multivariate analysis revealed that burnout was associated with Caucasian race. There are little data regarding ethnicity as a risk factor for burnout. South Africa has a unique history regarding racial divides, but it is uncertain whether this is a factor that needs to be considered in the light of this finding.

Individual measures of burnout
Three different measures of burnout were analysed individually to determine if particular associations differed from the definition of burnout utilised here (EE + DP). A total of 60.9% of doctors scored high on EE; 59.9% scored high on DP and 55.4% scored low on PA. Global rates of EE range between 43% and 80%, which was comparable to what we found. 34 Global rates for high DP are 35% -61% (similar to our rates) and 32% -44% for low PA (lower than our rates). Longer overtime and clerical hours were high EE risk factors, which could be explained by increased emotional and physical fatigue. Having a psychiatric diagnosis, the use of medications and engagement in past or present psychotherapy were all associated with high EE. These three categories are closely related to each other. Having a pre-existing mental illness may be a risk factor for or a result of high EE. Symptoms of depression and anxiety may be the same as those found in EE. We hypothesise that medications could be used by doctors to either manage EE or a mental illness. Interestingly, EE was associated more with psychiatric diagnosis after internship, further enmeshing the symptomatology between these two parameters.
Depersonalisation represents the interpersonal element of burnout and it appears that certain disciplines (such as emergency medicine and obstetrics and gynaecology) may be particularly vulnerable in terms of this aspect of burnout but not EE or PA. High levels of DP were found in Caucasians and it is uncertain why this may be a factor in our setting. Those who performed less than 10 h per week of academic duties had the highest risk; this may be linked to interns, medical officers and specialists when compared with registrars (who are enrolled in a training programme), engaging in less academic activities. It also implies that high DP is associated more with other risk factors rather than the academic aspects of medicine. Psychiatric diagnosis made before internship was associated with increased risk of high DP; it can, therefore, be surmised that pre-existing mental illness will dictate higher levels of DP. It is the only subscale that shows this particular risk pattern, as PA was only associated with increased risk after internship, and EE was associated with an increased risk both before and after internship (the risk being higher after internship).
Females were found to be at higher risk of low PA. One explanation is that females often face more discrimination in the workplace compared with their male counterparts. 35,36 The disciplines of pathology, public health, emergency medicine and anaesthetics were found to have low PA scores and this will require exploration in a larger scale study. While the discipline of emergency medicine was significantly associated with high overall burnout, anaesthetics was associated with higher depressive symptoms.

Depression
Doctors have higher rates of depression and suicide when compared with the general population. 14,37 Of the participants, 53.31% at CMJAH screened positive for depressive symptoms http://www.sajpsychiatry.org Open Access higher than the reported rates in previous studies (30% -40.7%). 21,22 A systematic review on registrars indicated a prevalence of 1% -56% for depression. 37 This may be linked to using a cut-off score of eight on the PHQ-9 in order not to miss milder forms of depression, which may have been understated in other clinician populations.
In terms of individual and psychiatric factors, our findings are in keeping with most known risk factors for depression viz. female sex, age of onset (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), a past episode of depression and positive family history. 38 Relationship status was, however, not associated with depressive symptoms in this study; this differs from the literature with those who are single, widowed or divorced are at a higher risk of depression. 39 Ethnicity was also shown not to be a risk factor, which is also in keeping with other findings in the SA context. 40,41 While there are global variances in rates of depression between medical disciplines, in this study, the disciplines of anaesthetics and obstetrics and gynaecology had a significantly elevated risk for depressive symptoms but not for overall burnout.
Anaesthetists are more likely to complete suicidal acts and more likely to die by substance-related suicide when compared with the general population. 42 This concerning finding warrants further investigation and application of appropriate preventative and other interventions.
The current use of medications such as antidepressants, benzodiazepines and stimulants was associated with depression and anxiety. While we can conclude that the use of antidepressants is likely to be for the treatment of depression, we cannot conclusively say the same for the latter two groups of medication. We know that doctors are less likely to seek medical help and are more likely to selfmedicate. 12 We also know that that long-term benzodiazepine use may be linked to depressive symptoms. 43,44,45 The bidirectional relationship between burnout and depression makes it difficult to determine causality over the course of both conditions. A recent systematic review and meta-analysis concluded that job strain may precipitate clinical depression. 46 Although there is an overlap of both symptoms and risk factors, it appears, from our findings, that unique risk factors for depression amongst doctors are female sex, family history of depression and working in the fields of anaesthetics and obstetrics and gynaecology. For burnout, the only standalone factors of note are Caucasian race and working in the discipline of emergency medicine. If one examines the individual burnout subscales, other factors may contribute. Younger age is a common risk factor for both conditions and junior doctors were more likely to be both depressed and burnt out.

Conclusion and recommendations
High burnout and depression rates were found in this study. There appears to be a bidirectional relationship between the two conditions. Although they have shared risk factors, specific risk factors may be determined for each in this population and confirmed with a longitudinal study looking at the causality between burnout and mental illness. Further studies with semi-structured interviews within the institution may determine organisational and departmental factors, impacting the high rate of burnout and depression, especially within disciplines with significant findings, for example, anaesthetics, emergency medicine and obstetrics and gynaecology.
While this study briefly reflected on the use of medications, the study population is unique because doctors have easier access to medication when compared with the general population. In this regard, it may be interesting to note how many of the physicians self-prescribed medication to manage burnout and depressive symptoms.
It is important to determine the incidence of burnout and factors associated with it to tailor interventions appropriately. 47 Organisation-based interventions target workload reductions, enhancing teamwork, structured schedules and supervising job demands. 47 Individual-based interventions involve psychological therapies, such as cognitive behavioural therapy and mindfulness, as well as strategies at improving coping and communication. 48 There is an emphasis on convening physicians in social settings (a shared space), improving communication and decreasing stress. However, there are currently no definitive guidelines or interventions that are considered applicable in all settings for all professionals. Given the prevalence and risk burnout poses to both physicians and patients, the development of guidelines on how to manage this condition would be useful.
Further research in this regard should be encouraged and should focus on establishing a cohesive strategy between individual and organisation interventions to decrease burnout rates.

Limitations
There are several limitations to this study. Firstly, selfreported questionnaires were utilised and these are dependent on participants, who may over-or under-report symptoms. The response rate for interns and registrars was higher, which may indicate responder bias, showing that these groups were more likely to screen positive for burnout and depression. Additionally, findings at a tertiary-level academic institution in an urban setting in South Africa may not be generalisable to other areas of healthcare. This study utilised the PHQ-9, which is a screening tool, and therefore, a formal clinician review would be needed to make a diagnosis of depression. Finally, this study determined associations between risk and certain variables, and therefore, we cannot comment on causality.