Article Text

Download PDFPDF

Prevalence of burnout syndrome among medical residents: experience of a developing country
  1. Khalil Ashkar1,
  2. Maya Romani1,
  3. Umayya Musharrafieh1,
  4. Monique Chaaya2
  1. 1Department of Family Medicine, American University of Beirut, Beirut, Lebanon
  2. 2Department of Epidemiology and Population Health, American University of Beirut, Beirut, Lebanon
  1. Correspondence to Monique Chaaya, Associate Professor, Department of Epidemiology and Population Health, Faculty of Health Sciences, American University of Beirut, Beirut 11-2036, Lebanon; mchaaya{at}aub.edu.lb

Abstract

Background It has been well established that burnout, defined as emotional exhaustion (EE), de-personalisation (DP), and a ‘decreased sense of personal accomplishment (PA) due to work related stress’, is prevalent to a great extent among medical residents. This has been implicated in the delivery of suboptimal patient care. Residents in developing countries work under difficult working and social conditions and therefore examining burnout is important when addressing quality of care and working conditions.

Objectives The objectives of the study were to assess the point prevalence of burnout among residents in Lebanon and investigate some of the work and non-work related factors associated with burnout.

Methods A cross sectional study of all medical residents was carried out in two major hospitals in the capital city. A total of 155 residents responded, representing all specialities. Maslach Burnout Inventory for Health Services Workers (MBI-HSS) was used. Burnout in three domains was dichotomised into high versus not high burnout.

Results The prevalence of burnout was high in all the domains with the highest in the domain of emotional exhaustion (67.7%). 80% of the sample had a high level of burnout in at least one domain. Being female, experiencing a major stress, working for more than 80 h per week, and having more than eight calls per month increased the odds of burnout in at least one domain.

Conclusions The high burnout level calls for action. This could be addressed by re-examining workload and other working conditions of residents as well as attending to their psychological wellbeing. The limitations of the study are also discussed.

  • Burnout
  • medical residents
  • stress
  • work-related
  • developing country
  • education and training (see medical education and training)
  • human resource management
  • quality in healthcare

Statistics from Altmetric.com

Introduction

Burnout is a syndrome defined by the three principal components of emotional exhaustion (EE), depersonalisation (DP), and diminished feelings of personal accomplishment (PA). Unlike major depressive disorder, which pervades all aspects of a patient's life, burnout is a distinct work related syndrome.1 It occurs in jobs that require extensive care of other people such as medical doctors. Several small non-randomised studies found a higher incidence of depression,2 anger and hostility3 in medical residents than in the general population. These findings were partly attributed to chronic sleep deprivation and fatigue.4 5 The consequences of burnout may be serious. Studies have correlated medical errors and various measures of distress; those who scored high on burnout measures were twice as likely to report an error in the next 3 months. A similar trend towards increased future errors for physicians with symptoms of depression was also noted.6 Despite the fact that residency programmes are implementing procedures to identify, manage, and reduce resident stress and impairment,7 new sources of stress continue to emerge. Besides physical exhaustion, residents are required to master a rapidly expanding scientific and clinical knowledge base, cope with the managed care systems' policies and rules, deal with sicker patients, discharge patients faster from inpatient services, and adapt to computerised systems and increased amount of paper work. All of these factors add to the resident's workload and increase stress.

In developing countries, stresses are present and even more amplified. Residents work under suboptimal conditions, are rewarded less financially, and have to cope with different sources of stress such as war, absence of infrastructure for healthcare and substandard economic conditions.

It is important to note that Lebanon ranked second among the countries from where physicians in the USA graduated after adjusting for country population size. Forty-one per cent of Lebanese medical graduates over the last 25 years are currently active physicians in the USA. There has been a consistent upward trend in the number of Lebanese medical graduates in the USA since the late 1970s. An alarming percentage of students from Lebanese medical schools intend to migrate for postgraduate training, mainly to the USA, and only a minority intends to return directly to Lebanon after finishing training abroad. This is mainly due to the lack of financial incentives, political instability, or deficiencies in leading residency programmes.8 9

Given the magnitude and historical trends of migration of Lebanese medical graduates, exploration of its causes and impact are warranted. To our knowledge this is the first study on the prevalence of burnout syndrome among residents in the Middle East. In Lebanon the need to assess the prevalence of burnout among residents becomes important because of the additional burdens of political instability and difficult daily living conditions. The data obtained may reveal the predisposing factors and suggest possible solutions for burnout in this population of residents.

The specific objectives of the study were to assess the point prevalence of burnout among Lebanese medical residents and to identify which factors—demographic, work related or psychosocial—were associated with high burnout.

Subjects and methods

A cross sectional study was conducted in two large academic hospitals in the capital city, Beirut. The first was a private hospital accommodating 300 beds, while the second was a public academic centre serving 550 beds. The study period extended from 20 May 2008 to 20 June 2008. The study was approved by the ethical committees in the two concerned hospitals.

All residents from both hospitals were asked to complete a self administered questionnaire. Out of a total of 206 residents in both hospitals, 155 participated in the survey (81 private and 74 public) with a response rate of 75%. The survey questionnaire consisted of two parts. The first part covered information on sociodemographic characteristics (age, gender, marital status, and number of children), work related questions, and psychosocial variables. The second part of the questionnaire was the Maslach Burnout Inventory for Health Services Workers (MBI-HSS). MBI consists of 22 items related to emotional exhaustion (EE), depersonalisation (DP), and diminished feelings of personal accomplishment (PA): nine items investigated EE, five for DP and eight for PA. Items comprising the scale are measured on a seven point Likert scale (0= never, 1=a few times or less, 2=once a month or less, 3=few times a month, 4=once per week, 5=a few times a week, 6=every day). MBI is widely used in assessing burnout and has a proven reliability and validity.1 The internal reliability in this study measured by Cronbach α was 0.74 for the total scale and 0.86, 0.68, and 0.66 for the three dimensions EE, PA and DP, respectively. Permission was obtained to reproduce 500 items from the publisher, and the English version was used with no translation.

Questionnaires were mailed to the residents with instructions for returning them to the secretary of their respective departments. Residents who did not respond in 10 days were sent reminders.

Measures

Three separate scores were computed to assess the three domains of burnout. These were categorised as follows: for EE: low burnout <13, average 14–26, high >27; for DP: low burnout <5, average 6–9, high >9; and for PA: high burnout <33, average 34–39, low >40. The three burnout scores were further dichotomised into high burnout versus no high burnout.

Age was categorised into ≤30 years and >30 years. The work related factors consisted of six main variables: speciality grouped into three main categories—medical (including medicine and paediatric specialities), surgical (including surgery, obstetrics and anaesthesia specialities), and services (including laboratory and radiology specialities); postgraduate training year (1 to 5); continuous working hours (<30 h, ≥30); number of calls per month (<8 vs ≥8); exceeding 80 h of hospital/clinic work per week (yes/no); availability of adequate supervision (agree, neutral, disagree); and type of hospital (private, public). Residents were also asked whether or not they experienced a major stressful life event 6 months before the study. The four items on coping were grouped into two categories: those who reported drinking alcohol/smoking a lot or doing nothing to cope with stress at work were grouped under ‘negative coping’; the remaining residents who reported trying to find solutions, going out with friends and speaking about their problems were labelled as having ‘positive coping’.

Analysis

Descriptive statistics were calculated for the total sample (means and standard deviations of the total and sub-scores when measured as continuous and frequency distributions for all other categorical variables). Bivariate analyses were performed to compare proportions with high burnouts among different groups of the sociodemographic and work related variables. Only variables significant at p<0.1 were considered for multivariable analyses. Logistic regressions were performed to calculate the independent contribution of each factor to predict high burnout in the three domains. Adjusted odds ratios (OR) and their 95% confidence intervals (CI) were computed.

Results

The distribution of the residents according to demographics and work related variables is displayed in table 1. The sample included a larger proportion of males with an overwhelming majority younger than 30 years. A little less than one fourth were married with only one divorced and the remaining single. The sample included residents from 10 specialities and the five residency years. The majority reported working consecutively for more than 30 h.

Table 1

Distribution of residents by basic demographic and work related variables

The percentage distribution of the sample according to the three different burnout domains is presented in table 2. The largest percentage that exhibited high burnout was in the domain of emotional exhaustion (67.7%). Eighty per cent of the sample had a high level of burnout in at least one domain. In the bivariate analyses, results demonstrated that female residents have significantly higher proportions of high burnout in EE and PA. Age and marital status were not related to any burnout domain. With respect to work related factors, number of calls, type of speciality, postgraduate years, and number of continuous working hours were statistically associated with high burnout. Residents who had more than eight calls per month, and those working continuously for more than 30 h, were significantly more likely to have high burnout than their counterpart for EE and DP domains. Working for more than 80 h per week increased the likelihood of high burnout only for the DP domains. Residents specialising in medicine were more likely than the two other speciality groups to have high burnout in EE; however, the differences among the groups were not statistically significant. Surgery residents were statistically significantly more likely to have high burnout in DP while residents in services showed higher burnout in PA. There was no statistically significant association with postgraduate year, availability of adequate supervision, and type of hospital. Exposure to a major stress 6 months before the study was significantly associated with high burnout in the two domains of EE and DP. Those who reported coping with stress, by either drinking alcohol/smoking a lot or doing nothing, were more likely to have high burnout in DP. (table 3)

Table 2

Proportion of residents with different burnout levels in the three domains

Table 3

Proportion of residents with high burnout scores in each of the domains by selected demographic and work related variables

Predisposing factors for burnout

Table 4 shows the results of the different regression models for the association between the different dimensions of burnout and the sociodemographic, work factors, and psychosocial variables that were statistically significant at the bivariate level (p≤0.1). Multivariable analysis showed that gender, working hours and major stress were significant independent determinants of EE. Being female, working more than 30 h, and experiencing major stress in the past 6 months increased significantly the odds of high EE burnout by 2.54, 2.91, and 3.38, respectively. Regarding DP, high burnout was significantly associated with presence of a major stress, and having more than eight calls per month. The odds of high burnout among those who have more than eight calls were 3.47 times the odds of those who have fewer than eight. Experiencing major stress was associated with an odds ratio of 2.04. None of the investigated variables was statistically associated with high burnout in PA.

Table 4

Results of the logistic regression analysis models for the association between high burnout in the different domains (EE, DP, PA) and selected sociodemographic and work related variables

Discussion

This study represents one of the first attempts to assess the magnitude of burnout among medical residents specialising in different areas in an Arab country and to analyse the influence of workload and subjective working conditions. The results of this study reveal a worrisome rate of burnout among Lebanese medical residents across both public and private hospitals. It is valuable to know that an important element of the syndrome is a negative impact on job performance with potentially devastating consequences for patients. Besides being implicated for delivery of suboptimal care10 and medical errors,11 12 stress and burnout are related to the desire to give up practice and are, therefore, a human resources issue for the entire healthcare system.13

The prevalence of burnout in the three domains is higher than what is reported in the EGPRN Study revealing 12% for burnout for EE, 35% for DP, and 32% for PA.14 Of significance is the higher proportion of burnout among females, which is contrary to most studies that showed no gender preponderance15 Because women have a higher lifetime risk of developing depression, it is reasonable that this might extend to burnout as well.16 In one study17 lack of workplace control predicted burnout in women but not in men, with increasing odds of burnout for additional hours worked per week. Despite the presence of comforting balancing factors at home, like family support and home helps, Lebanese female residents remain vulnerable. Previous studies demonstrate that female physicians handle more elderly and psychosocial issues and this may be a contributing factor.18 Although the number of married residents was small, burnout was not associated with marital status.

Women in Lebanon also face other social challenges and pressures. There is the social stigma of living alone, away from family, and there is a social urge for females to get married. The medical profession exaggerates both parameters.

Exposure to stressful events is associated with burnout, and the Lebanese are a vulnerable population. Nearly 70% of the Lebanese population is exposed to one or more events associated with war (such as being a civilian in a war zone, or being a refugee), and almost half of the Lebanese population has been exposed to one or two of the 10 war events.19 A serious political crisis occurred in Lebanon 2 weeks before the study and that was a major stressful event which medical residents were exposed to. Despite the notion that burnout is primarily linked to work related stress, personal life events can demonstrate a strong relationship to professional burnout.20

Another important factor associated with burnout in this study is long working hours. Long working hours deprive residents of sleep and this has unfavourable effects on resident training and their patient care.21 In response to the well documented phenomenon of burnout among residents and the established association with medical errors, several countries have introduced limits on the working week and number of consecutive hours worked.22 Studies conducted to evaluate these changes reported a positive impact on the quality of life of residents, their education, and patient care.23 This was emphasised further by a study reporting a reduction in the prevalence of burnout post implementation of a work week hour limit.24 Residency programmes in Beirut have work restriction laws according to the system they follow—American versus Francophone. However, the implementation may not be strictly followed.

One significant finding is the lack of association between burnout and supervision. Although burnout was declared among residents without staff supervision in some reports,15 the lack of this association in our study is of interest. In a country like Lebanon, where health budgets are limited, residents may feel the need for expert opinions to decide on the best cost effective management. On the other hand, medical residents might not feel free to ask for support from their supervisors for fear of seeming incompetent and being put on the spot. As such, it is the balance of these two variables that establish the association between supervision and burnout.

Looking at the demographics, age among residents was not a predictor of burnout, indicating factors other than years of training contribute to burnout. This is not in agreement with other studies showing that older, more experienced doctors report lower psychological distress and burnout than younger doctors, due to lessons learnt over their years of training and practice.25

Worrisome enough is the significance of burnout in the area of DP among residents who cope by either drinking alcohol or smoking a lot. A recent study from the USA revealed that physicians have higher rates of suicide than the general population, and potential contributory factors include drug use and alcohol related disorders together with other factors.26

Despite the fact that physicians working in public hospital based practice reported significantly higher rates of burnout and a trend towards a higher rate of psychiatric morbidity,27 there was no significant difference in the incidence of burnout between private and public hospitals in this study. Distinctive features of public hospitals also applicable to Lebanon include: demoralising situations of dealing with sick patients and lack of hospital beds, necessity of making everyday triage decisions, various conflicts and ethical dilemmas, lack of medical insurance, feeling undervalued and receiving poor support hospital service.

Limitations

The main limitation is that our results were obtained in cross-sectional study, which precludes evaluation of temporality and causality of the observed relationships. Another limitation is the exclusive reliance on self reported scales and rating and professional characteristics, which raises the issue of measurement error, related to systematic positive or negative response tendencies. The study was done in hospitals that follow different benchmarks for training residents (American versus Francophone models), and as such this may have an effect on the results. Also, this study was conducted in tertiary care centres in the capital city, and thus its results do not allow for generalisation and extrapolation to other health centres.

Conclusion

The results of this study indicate that a significant number of Lebanese residents suffer from burnout syndrome. It affects female residents more than males. Addressing workload, work hours and working conditions, and psychological wellbeing may decrease the incidence of burnout in medical residents. The profession's recognition and awareness of the serious issues of stress and burnout might be important in addressing the problem. Peculiar to Lebanon is the need to determine if this high rate of burnout has an impact on the migration of Lebanese medical graduates, and whether it is a contributing factor in addition to other factors. Strategies suggested include screening burnout among residents, improving scheduling, implementing work hour limits in order to fit international standards, and introducing coping mechanisms along with techniques to manage stress.

Main messages

  • A substantial proportion of residents in Lebanon suffer from burnout.

  • Attributed reasons have been linked to three main factors, gender (being a female), working hours (working consecutively for more than 30 h in one shift), and major stress (war is one of the main stressors).

  • Due to the working and sociopolitical conditions, as in many other countries in the region, residents are migrating to developed countries where working conditions are better and there may be less stress.

  • Recognising the problem and addressing working conditions are short term recommendations made in the study.

Current research questions

  • Would changing working conditions in hospitals reduce burnout among residents?

  • Is burnout a contributing factor in the migration of residents?

  • Is burnout a significant problem among attending physicians?

References

View Abstract

Footnotes

  • Competing interests None.

  • Ethics approval This study was conducted with the approval of the American University of Beirut Medical Center and the Hariri Public Hospital.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.