Secondary Logo

Journal Logo

Featured Articles: Original Clinical Research Report

A Cross-Sectional Survey to Determine the Prevalence of Burnout Syndrome Among Anesthesia Providers in Zambian Hospitals

Mumbwe, Mbangu C. MBChB*; McIsaac, Dan MD, MPH, FRCPC†,‡; Jarman, Alison BMBS, BA (Hons)§; Bould, M. Dylan MBChB, Med

Author Information
doi: 10.1213/ANE.0000000000004464

Abstract

See Editorial, p 307

KEY POINTS

  • Question: What is the prevalence of burnout syndrome among anesthetists working in Zambian hospitals?
  • Findings: Burnout syndrome was detected in 51% of the respondents; 86% of respondents were nonphysician anesthetists working in different hospitals in the country.
  • Meaning: The high burnout levels among anesthetists working in Zambian hospitals require attention to minimize risk of negative consequences to the affected anesthetists, the patients, and their respective institutions.

Burnout syndrome can be defined as consisting of emotional exhaustion, depersonalization, and reduced personal accomplishment. It occurs among health care professionals due to a chronic mismatch between job demands and resources to cope with those demands.1 Emotional exhaustion refers to feelings of strain and fatigue. The second dimension, depersonalization, refers to a coping mechanism that involves withdrawal from work and detachment from people entrusted to one's care or feelings of cynicism. Anesthetists with a high level of depersonalization tended to engage in shorter preoperative conversations with patients, provide less information about pain or the procedure, and to have less empathy with them.2 The third dimension of reduced personal accomplishment represents feelings of frustration toward work and lack of successful achievement within one's job and organization.3 Burnout makes the professional more prone to errors, thereby risking patient safety and results in several negative institutional costs such as frequent medicolegal problems, absenteeism, poor staff retention, and increased staff turnover.4

Burnout has been demonstrated to be common in anesthesia providers in middle-income countries. Zambia is defined by the World Bank as a lower-middle income country. In South Africa, Van der Walt et al5 found a burnout prevalence of 21% among 124 physician anesthetists. Of these, more females were affected than males, and less-experienced anesthetists were more affected than their senior colleagues.5

By definition, providing anesthetic care in an under-resourced environment results in a mismatch between demands and resources to meet those demands. Clinical Officers undergo 2 years of training in anesthesia following completion of a 3-year diploma of science in clinical medicine. Physicians take a 4-year Masters in Medicine in anesthesia following completion of internship. The anesthesia workforce in Zambia is overburdened. Anesthesia services outside of the capital city of Lusaka are largely provided by nonphysician anesthetists (clinical officers and nurse anesthetic officers) who work in environments that are severely under-resourced with essential drugs and equipment.6 These nonphysician anesthetists often work without supervision or mentorship as there are not enough physician anesthetists to provide this service. Also, the physicians tend to be concentrated in the capital and the larger cities. Despite significant investment in anesthesia training across the board by the government of Zambia and partners, it is likely that burnout poses a significant risk to the success and sustainability of these efforts.7

With the exception of the South African study cited, there are little data on burnout in the sub-Saharan region where working conditions are very different from high-income countries. Quantifying this issue is the first step toward plans to attenuate the effect of burnout on scaling up human resources for health in anesthesia. The primary objective of the study was to determine the prevalence of burnout syndrome among anesthesia providers working in Zambian hospitals. The secondary objective was to determine which sociodemographic and occupational factors were associated with increased risk of burnout among anesthesia providers in Zambian hospitals.

METHODS

Design

This was a cross-sectional survey evaluating the prevalence of burnout syndrome among all anesthesia care providers in Zambia. We followed best-practice recommendation for survey design and conduct for clinicians. Ethical approval (ref: 004-03-17) for this study was sought and granted by the University of Zambia Biomedical Research Ethics Committee (UNZABREC). Reporting is consistent with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines for the reporting of observational research.8

Population

Our target population was all registered anesthesia providers in Zambia. The Health Profession Council of Zambia (HPCZ, the regulatory body for all anesthesia providers in Zambia) and the Zambian Ministry of Health maintain a registry of regulated anesthesia providers, including contact information. Therefore, our sampling frame consisted of all registered nontrainee anesthesia providers in Zambia. According to the Ministry of Health register, there were 154 nonphysician and 30 physician anesthetists in service at the time of the study.

Recruitment and Data Collection

The study was verbally explained to each participant in person or on the phone and in written communication through an information sheet that was delivered to the study population physically or via e-mail, alongside the consent form and questionnaire. A consent form that was separate to the survey itself had to be completed and sent back to the primary investigator.

Outcome

Our primary outcome was burnout, which we measured using the validated and widely used Maslach Burnout Inventory Human Services Survey (MBI-HSS).9 The English version of the questionnaire was used as it is the official language of Zambia and the language of instruction for all anesthesia training programs in Zambia. The MBI-HSS is a 22-item questionnaire (see Supplemental Digital Content, Appendix, http://links.lww.com/AA/C943) that assesses burnout in 3 dimensions: (1) emotional exhaustion, (2) depersonalization, and (3) personal achievement. Multiple questions within each dimension are scored on a 7-point Likert scale. Following standard scoring practices, burnout was defined as being present if (1) an individual scored >27 on the emotional exhaustion subscale (high emotional exhaustion) and (2) either >10 on the depersonalization subscale (high depersonalization) or <40 on the low personal accomplishment scale.10 A secondary prevalence assessment was made for participants who did not meet the above standard criteria but did score high levels of emotional exhaustion, high depersonalization, or low personal accomplishment on a unique subscale basis. High scores on the emotional exhaustion dimension were defined as 26–54, medium 16–25, and low 0–15. High scores on the depersonalization subscale were defined as 9–30, medium 3–8, and low as 1–2. High scores on the personal accomplishment subscale were defined as 43–48, medium 34–42, and low 0–33, as performed by Natalia and Raluca.11

OTHER VARIABLES COLLECTED

Participants also completed a 2-part questionnaire (see Supplemental Digital Content, Appendix, http://links.lww.com/AA/C943) to allow self-report of sociodemographic data (age, gender, marital status, and number of dependents) and occupational factors (job position, team work perception, and availability of equipment, frequency of negative outcomes, vacation days, and remuneration). These variables were identified as being associated with burnout based on previous studies of burnout in health care professionals, as well as factors that are known locally to risk a mismatch between the demands of anesthesia practice in Zambia and the provider's ability to meet those demands.12 The survey was pretested among the study team and then piloted in senior anesthesia trainees to assess whether it was clear, acceptable, and easily administered. Following minor refinements, the survey (MIB-HSS plus participant questionnaire) was circulated to the target population. Data collection was performed between August 2017 and February 2018. To obtain a maximal response rate, regular follow-up phone calls and e-mails were sent to track the participant's progress and provide an opportunity to clarify any queries by participants.

Statistical Analysis

Data were analyzed using SPSS (version 24; IBM, Armonk, NY) and SAS (version 9.4, SAS Institute, Cary, NC). Descriptive statistics were calculated as percentages for binary and categorical data, or as mean and standard deviation for normally distributed continuous variable or median and interquartile range (IQR) for skewed continuous variables. Distributions of data were determined by visually inspecting histograms of each variable. The proportion of participants who had burnout was calculated as a proportion, and confidence intervals were determined using Wilson's method. Independent variables in anesthesia providers with burnout were compared to those without burnout using χ2 or Fisher exact test for nominal/dichotomous data and the Mann-Whitney U test for continuous data that was not normally distributed. Unadjusted associations between associated variables and burnout were calculated using univariable logistic regression.

To evaluate the adjusted contribution of variables on burnout, we prespecified a set of variables that we postulated would be the most strongly associated with burnout based on literature review and local knowledge: gender (as a binary variable), vacation (dichotomized to a binary variable, either taking vacation or not), having the right equipment to perform work to an appropriate standard (dichotomized to a binary variable), having the right team to carry out work to an appropriate standard (dichotomized to a binary variable), whether physician or nonphysician (binary), workload in hours per week (self-reported as a continuous variable), and years spent in independent practice (reported as a continuous variable). We then ran a multivariable logistic regression model with burnout as the dependent variable and all of the specified independent variables. An α of .05 was used as a significance threshold for all analyses.

Sample Size

We aimed for a census measurement of burnout to obtain as much information as possible about burnout in a relatively small population of anesthesia providers in Zambia (total number of anesthesia providers was 184). A sample size calculation estimated that we would require 125 participants for a margin of error of 5% on a 95% confidence level.13 For the binary logistic regression, we followed a conservative rule of thumb where we allowed 1 degree of freedom for independent variables for every 10 participants who were positive for the dependent variable (ie, were classified with burnout syndrome).14

RESULTS

Table 1. - Respondent Characteristics and Occupational Data
Variable Not Burned Out (n = 78) Burned Out (n = 82) Total P
Age 40 (36–46) 37 (36–42) 38 (36–45) .043a
Gender
 Male 53 (67.9%) 64 (78%) 117 (73.1%) .159b
 Female 25 (32.1%) 18 (22%) 43 (26.9%)
Marital status
 Married 69 (88.5%) 71 (86.6%) 140 (87.5%) .927b
 Single 8 (10.3%) 10 (12.2%) 8 (11.3%)
 Other 1 (1.3%) 1 (1.2%) 2 (1.3%)
Dependents 3 (2–3) 3 (2–4) 3 (2–4) .725a
 Background
  Nonphysician anesthesist 63 (80.8%) 74 (90.2%) 137 (85.6%) .115b
  Physician anesthesiologist 15 (19.2%) 8 (9.8%) 23 (14.4%)
Years spent in training 3 (2–3) 3 (2.5–3) 3 (2–3) .823a
Years in independent practice 3 (1–7) 2 (1–4) 2 (1–6) .059a
Type of employment
 Full time 69 (88.5%) 80 (97.6%) 149 (93.1%) .075b
 Part time 4 (5.1%) 1 (1.2%) 5 (3.1%)
 Contract 5 (6.4%) 1 (1.2%) 6 (3.8%)
Do you receive your pay on time?
 Yes 22 (28.2%) 28 (34.1%) 50 (31.3%) .469b
 No 56 (71.8%) 54 (65.9%) 110 (68.7%)
Pay on par with services rendered?
 Yes 70 (89.7%) 79 (96.3%) 11 (6.9%) .124b
 No 8 (10.3%) 3 (3.7%) 149 (93.1%)
Data refer to the number (percentage or 25th–75th percentile). Statistical tests used were χ
2 or Fisher exact testb for dichotomous/nominal data and the Mann-Whitney U testa for continuous data that was not normally distributed. P values are for comparisons between burned out and not burned out participants for each variable.

Table 2. - Perception of Working Conditions
Never 0% of Time (%) Rarely ≈10% of Time (%) Occasionally ≈30% of Time (%) Sometimes ≈50% of Time (%) Frequently ≈70% of Time (%) Usually ≈90% of Time (%) Every Time 100% of Time (%)
How often do you see patients have negative outcomes at work such as death or permanent disability? 5 (3.1) 130 (81.3) 17 (10.6) 4 (2.5) 4 (2.5)
How often do you feel that you don't have the equipment to carry out your work to an appropriate standard? 45 (28.1) 10 (6.3) 45 (28.1) 64 (40.0) 17 (10.6) 15 (9.4) 8 (5)
How often do you feel you don't have the right team around you to carry out your work to an appropriate standard? 2 (1.3) 19 (11.9) 37 (23.1) 72 (45.0) 20 (12.5) 5 (3.1) 5 (3.1)
How often are you supervised by a senior colleague? 48 (30) 93 (58.1) 8 (5) 5 (3.1) 4 (2.5) 2 (1.3)
How often do you feel like you do cases you are not comfortable with without support? 7 (4.4) 47(29.4) 12 (7.5 18 (11.3) 69 (43.1) 4 (7.3)

One hundred sixty anesthesia providers gave complete responses to the survey, of a total of 184 anesthesia providers in Zambia (154 nonphysician, 30 physician), giving a response rate of 87%. Respondent characteristics and occupational data are detailed in Table 1, and perception of working conditions are detailed in Table 2. The median (IQR [range]) age was 38 (36–45 [31–76]). The median (range) number of years in anesthesia training was 3 (2–3 [1–9]).

Prevalence of Burnout

Burnout criteria were found in 51.3% (95% confidence interval [CI], 43.2–59.2) of respondents. With respect to the subcategories of burnout, high emotional exhaustion was present in 106 (66.3%, 95% CI, 58.7%–73.2%) respondents, 72 (45%, 95% CI, 37.4%–52.7%) had scores indicating high levels of depersonalization, and 38 (23.8%, 95% CI, 17.7%–30.8%) had scores indicating low personal achievement. Figure 1 shows the self-reported weekly working hours of all respondents. The median response indicates around 35 additional hours worked per week in addition to the 40 hours recommended by the Ministry of Health. Most participants took no vacation (Figure 2).

Figure 1.
Figure 1.:
Histogram of hours worked per week by respondents (the vertical black line refers to the recommended working hours as set by the government of the Republic of Zambia; the vertical dashed red line refers to the median).
Figure 2.
Figure 2.:
Histogram of number of days spent on vacation in the previous year by respondents.
Table 3. - Univariate Binary Logistic Regression Output
95% CI for Odds Ratio
Independent Variables P Value Odds Ratio Lower Upper
Vacation .08 0.56 0.30 1.06
Equipment .01 2.36 1.21 4.60
Team .001 3.32 1.68 6.53
Outcomes .07 2.29 0.93 5.66
Gender .15 1.68 0.82 3.40
Background .09 2.20 0.88 5.54
Years in independent practice .21 0.97 0.93 1.01
Workload hours .72 1.00 0.99 1.02
A
ll participants were included in the analysis. The odds ratio refers to the odds of having burnout. Vacation was dichotomized to participants who had or had not taken vacation, with the reference category being those who had not taken vacation; equipment refers to how participants responded to the question, “how often do you feel you don't have the right equipment to carry out work to an appropriate standard?” with the reference condition referring to participants who answered they often had the right equipment; team refers to how participants responded to the question, “do you feel you have the right team around you to do your job?” with the reference category referring to participants who answered they had team support; outcomes refer to how participants responded to the question, “how often do you experience negative outcomes such as death or permanent disability?” with the reference condition referring to participants who answered they experienced negative outcomes infrequently; gender—is the reference category that refers to females; background refers to the position of being either a physician or nonphysician anesthetist with the reference category referring to physician anesthetists; workload hours variable refers to self-reported number of hours worked each week.
A
bbreviation: CI, confidence interval.

Table 4. - Multivariable Binary Logistic Regression Output
95% CI for Odds Ratio
Independent Variables P Value Odds Ratio Lower Upper
Vacation .20 0.63 0.31 1.3
Equipment .38 1.42 0.64 3.2
Team .01 2.91 1.33 6.39
Outcomes .09 2.67 0.86 8.33
Gender .14 1.84 0.82 4.10
Background .02 3.40 1.25 12.36
Years in independent practice .20 0.97 0.92 1.02
Workload hours .93 0.999 0.98 1.02
A
ll participants were included in the analysis. The odds ratio refers to the odds of having burnout. Vacation was dichotomized to participants who had or had not taken vacation, with the reference category being those who had not taken vacation; equipment refers to how participants responded to the question, “how often do you feel you don't have the right equipment to carry out work to an appropriate standard?” with the reference condition referring to participants who answered they often had the right equipment; team refers to how participants responded to the question, “do you feel you have the right team around you to do your job?” with the reference category referring to participants who answered they had team support; outcomes refer to how participants responded to the question, “how often do you experience negative outcomes such as death or permanent disability?” with the reference condition referring to participants who answered they experienced negative outcomes infrequently; gender is the reference category refers to females; background refers to the position of being either a physician or nonphysician anesthetist with the reference category referring to physician anesthetists; workload hours variable refers to self-reported number of hours worked each week.
A
bbreviation: CI, confidence interval.

Table 3 details the output of the unadjusted univariate binary logistic regression analyses. Table 4 details the output of the multivariable binary logistic regression analysis. Following variable selection, perceiving that the anesthesia provider did not “have the right team around [them] to do [their] job” (odds ratio [OR], 2.91, 95% CI, 1.33–6.39) and the position held (a background as a nonphysician) (OR, 3.4, 95% CI, 1.25–12.34) were the only significant variables retained to be associated with burnout.

DISCUSSION

Our study found a high prevalence of burnout among anesthesia providers working in Zambia. Burnout was found to be associated with the perception of infrequently having “the right team around you to carry out your work to an appropriate standard.” Burnout was also associated with being a nonphysician; as physician anesthesiologists tend to be concentrated in the capital or larger cities, and nonphysicians tend to work in more rural areas, it is possible that this is a proxy for isolated and rural practice, rather than a factor specifically relating to training. The “right team” was explicitly defined to the participants as consisting of at least 1 senior physician anesthetic officer to consult regarding difficult cases and colleagues (physician or nonphysician) to share tasks and ideas with. The “right equipment” was defined as the consistent availability of a functional anesthetic machine, oxygen supply, suction machine, airway equipment, and minimum monitoring tools according to Zambian standards, including pulse oximetry, blood pressure, and electrocardiography.

These data suggest that urgent action is required to prevent harm, by all key stakeholders including the Zambian Ministry of Health. The finding that over half of anesthesia providers meet criteria for burnout likely poses a significant ongoing threat to patient safety, anesthetist wellbeing, and quality at institutional level. Early detection and planned interventions are therefore crucial. Our data identify key areas on which interventions could be planned, in particular improving team support structures and training more anesthesia providers, and probably more physician anesthetists. These measures would reduce isolation and provide much needed support to less experienced or junior counterparts. Pay, vacation, gender, and weekly working hours were not found to be statistically significant in this study; however, we note that our sample size is limited by the relatively small number of anesthesia providers in Zambia. This does not exclude these from being factors in burnout and their strong presence in literature from other studies would suggest it is still worth considering them in the planning of interventions to reduce burnout.

In South Africa, a similar study by Van der Walt et al5 among 124 physician anesthetists at the University of Witwatersrand found a burnout prevalence of 21% and scores of the individual dimensions as follows: 45.2% had high scores in emotional exhaustion, 50% had high levels of depersonalization, and 46% had low levels of personal achievement. The relatively lower burnout prevalence could be attributed to better team support that exists in university teaching hospitals compared to the isolated rural practice that majority of the anesthesia providers in Zambia experience. This was consistent with the study by Khetarpal et al15 in India, in which it was shown that burnout was considerably lower among anesthetists working in teaching hospitals with established team structures than community hospitals were anesthetists felt isolated. Team support structures have been shown to provide opportunities for junior colleagues to consult and share ideas and tasks, thereby allowing time for recuperation and minimizing isolation. In another study among 278 intensive care units (ICUs) in France involving 2525 critical care nurses for which 2392 nurses completed the MBI survey (95% response rate), 33% had burnout syndrome. Associated factors were personal characteristics such as age, organizational factors, such as ability to choose days off, participation in an ICU research group, quality of working relations, such as conflicts with patients, and end-of-life–related factors, such as caring for a dying patient.16 In contrast to our study, age was not a significant association and neither was vacation days in the previous year. This might be attributed to the observation by Abdulghafour et al17 that recuperation from work was most likely to achieve a protective effect against burnout if the conditions in the work environment were improved, and unlikely to have a positive effect if the worker returned to stressful work conditions. The quality of working relationships was not assessed in our study but would constitute an important variable to explore in future studies.

Strengths of our study were a high response rate, the use of a validated tool to measure burnout, and piloting the survey to ensure it was well understood. Our study had important limitations. Despite an excellent response rate, the total number of anesthesia providers in Zambia is low, and so we only included a limited number of variables in the regression model to prevent overfitting. In addition, due to the small number of anesthesia providers in Zambia, this study may be underpowered to identify factors which are associated with burnout, but that have relatively small effect size. Due to the limited sample size, we had limited ability to control for potential confounders not included in the model, such as age, and a larger sample size (not possible with the number of anesthesia providers in Zambia) would be required for more robust accounting for potential confounders. Therefore, although it is safe to conclude that unsupported practice and nonphysician status are important factors to improve on, we should be cautious in dismissing the other factors that were not found to be significant but have been found to be relevant in other literature. Our study is not necessarily generalizable to other populations with different contextual factors in other jurisdictions, including gender, age, and working conditions. The study did not specifically assess the nature of locations where service was provided. Noting a rural or urban workplace would have been of benefit in identifying to what extent nonphysician status is associated with burnout when accounting for remote and rural practice. Sensitive information such as frequency of negative outcomes experienced might have been underreported.

The prevalence of burnout syndrome among anesthesia providers in Zambian hospitals is high and of major concern, with potential harms to patients, anesthesia providers, and the health care system including medical migration. Nonphysician status and lack of an effective care team appear to be associated with burnout, and should be the focus of urgent efforts to improve the current situation.

ACKNOWLEDGMENTS

We appreciate the contributions to this work from the following: Nonphysician and physician anesthetists practicing in Zambian hospitals; head of University Teaching Hospital anesthesia department, Dr Christopher Chanda; chief anesthetic officer at Ministry of Health, Mr Wisdom Chelu; University Teaching Hospital anesthesia department for their support; and the various Zambia Anaesthesia Development Program (ZADP) fellows who gave their time and effort to bring this work to completion.

DISCLOSURES

Name: Mbangu C. Mumbwe, MBChB.

Contribution: This author helped with inception and development of research idea, writing the proposal, securing research ethics approval, data collection, data analysis, writing the first draft of the manuscript, and critical reviewing of the final paper.

Name: Dan McIsaac, MD, MPH, FRCPC.

Contribution: This author helped with data analysis, and critical reviewing of the final paper.

Name: Alison Jarman, BMBS, BA (Hons).

Contribution: This author helped with development of research idea, and critical reviewing of the final paper.

Name: M. Dylan Bould, MBChB, Med.

Contribution: This author helped with inception and development of research idea, writing the proposal, data analysis, and critical reviewing of the final paper.

This manuscript was handled by: Angela Enright, MB, FRCPC.

FOOTNOTES

GLOSSARY

CI = = confidence interval

HPCZ = = Health Profession Council of Zambia

ICU = = intensive care units

IQR = = interquartile range

MBI-HSS = = Maslach Burnout Inventory Human Services Survey

STROBE = = Strengthening the Reporting of Observational Studies in Epidemiology

UNZABREC = = University of Zambia Biomedical Research Ethics Committee

REFERENCES

1. Maslach C, Jackson S. The measurement of experienced burnout. J Occup Behav. 1981;2:99–113.
2. Hange L, Mingzhang Z, Adrian W, et al. Chinese anesthesiologists have high burnout and low job satisfaction: a cross-sectional survey. Anesth Analg. 2018;126:1004–1012.
3. Ahola K, Hakanen J. Job strain, burnout, and depressive symptoms: a prospective study among dentists. J Affect Disord. 2007;104:103–110.
4. Hodges SC, Mijumbi C, Okello M, McCormick BA, Walker IA, Wilson IH. Anaesthesia services in developing countries: defining the problems. Anaesthesia. 2007;62:4–11.
5. Van der Walt N, Scribante J, Perrie H. Burnout among anaesthetistsin South Africa. South African. J Anaesth Analg. 2015;21:27–30.
6. Janowicz A, Kasole T, Measures E, et al. Development of education and research in anesthesia and intensive care medicine at the University Teaching Hospital in Lusaka, Zambia: a descriptive observational study. Anesth Analg. 2017;125:287–293.
7. Kudsk-Iversen S, Shamambo N, Bould MD. Strengthening the anesthesia workforce in low- and middle-income countries. Anesth Analg. 2018;126:1291–1297.
8. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol. 2008;61:344–349.
9. Moradi Y, Baradaran HR, Yazdandoost M, Atrak S, Kashanian M. Prevalence of burnout in residents of obstetrics and gynecology: a systematic review and meta-analysis. Med J Islam Repub Iran. 2015;29:235.
10. Kluger MT, Townend K, Laidlaw T. Job satisfaction, stress and burnout in Australian specialist anaesthetists. Anaesthesia. 2003;58:339–345.
11. Natalia H, Raluca P. Prevalence of burnout in Romanian anaesthesia and intensive care physicians and associated factors. Jurnalul Român de Anestezie Terapie intensivă. 2012;19:117–124.
12. Gastão F, Gupta P, Moore R. Occupational well-being in anesthesiologists. Rev Bras Anestesiol. 2014;25:97–119.
13. Lenth RV. Java applets for power and sample size [computer software]. Available at: http://www.stat.uiowa.edu/~rlenth/Power. Accessed October 21, 2019.
14. Vittinghoff E, McCulloch CE. Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol. 2007;165:710–718.
15. Khetarpal R, Chatrath V, Kaur J, Verma A. Occupational stress in anesthesiologists and coping strategies: a review. Int J Sci Stud. 2015;3:188–192.
16. Poncet MC, Toullic P, Papazian L, et al. Burnout syndrome in critical care nursing staff. Am J Respir Crit Care Med. 2007;175:689–704.
17. Abdulghafour YA, Bo-Hamra A, Al-Randi MS, Kamel MI, El-Shazly MK. Burnout syndrome among physicians working in primary health care centers in Kuwait. Alexandria Journal of Medicine. 2011;41:351–357.

Supplemental Digital Content

Copyright © 2019 International Anesthesia Research Society