Associations between emotional intelligence and doctor burnout, job satisfaction and patient satisfaction

Abstract

Medical Education 2011: 45: 835–842

CONTEXT  The occupational health literature has long been dominated by stress-related topics. A more contemporary perspective suggests using a positive approach in the form of a health model focused on what is right with people, such as feelings of well-being and satisfaction.

OBJECTIVES  Using a positive perspective and multi-source data collection, this study investigated the inter-relationships among emotional intelligence (EI), patient satisfaction, doctor burnout and job satisfaction.

METHODS  In this observational study, 110 internists and 2872 out-patients were surveyed in face-to-face interviews.

RESULTS  Higher self-rated EI was significantly associated with less burnout (p < 0.001) and higher job satisfaction (p < 0.001). Higher patient satisfaction was correlated with less burnout (p < 0.01). Less burnout was found to be associated with higher job satisfaction (p < 0.001).

CONCLUSIONS  This study identified EI as a factor in understanding doctors’ work-related issues. Given the multi-dimensional nature of EI, refinement of the definition of EI and the construct validity of EI as rated by others require further examination.

Introduction

Most of the early research on job satisfaction utilised a disease-based model in that it placed a negative focus on what was wrong, including ill health and dissatisfaction in workers. An overview¹ of peer-reviewed journals indicated that occupational health research over the past 15 years has continued to be dominated by stress and stress-related topics, such as burnout and depression. This prevailing negative bias is illustrated by the fact that the total number of publications about negative states exceeds those about positive states by a ratio of 14 : 1.² A more contemporary perspective suggests using a positive approach whereby a health-based model is used to focus on what is right with people, such as present feelings of well-being and satisfaction.³ What are the factors that contribute to the optimal functioning of health care professionals who persist in the face of difficulties? For doctors in daily practice, which factors can enhance the level of patient satisfaction they facilitate, reduce their burnout, and ultimately have an impact on their own job satisfaction? This study refers to one of the important concepts in positive psychology, emotional intelligence (EI), and examines the inter-relationships among EI, patient satisfaction, doctor burnout and doctor job satisfaction.

With regard to the learning and assessment required within the six core competencies defined by the US Accreditation Council for Graduate Medical Education,⁴ EI is proposed to have the potential to deepen and enrich students’ understanding of competency during medical training.⁵ Emotional intelligence is defined as ‘the ability to perceive emotion, integrate emotion to facilitate thought, understand emotions, and regulate emotions to promote personal growth’.⁶ The theory of EI assumes that individuals with high EI have better interpersonal and communication skills. Studies have demonstrated that EI has a positive impact on social relationships⁷ and job satisfaction.⁸ Although it has been suggested that EI should be included among the criteria used to select applicants for medical school,⁹ no association between EI and selection scores¹⁰ or specialty choice¹¹ has been found. One study indicated that the nurse-rated EI of a doctor was positively correlated with patient satisfaction.¹² Another study indicated that EI had a positive impact on reducing occupational stress among dental undergraduates.¹³ However, still another study argued that EI could help moderate lower levels of stress, but that the EI effect might be lessened in the presence of an acute stressor.¹⁴ In addition, the definition and the measurement of multi-dimensionality merit some consideration. There have been serious academic debates as to whether or not EI is an ability that is distinct from personality traits^6,15or whether it is part of a mixed model that combines both.¹⁶ Assessments of EI vary by the model adopted and the purpose of the research. From the perspective of a medical educator, viewing EI as an ability allows for a greater likelihood of change than if it were a personality trait.⁵ Studies^6,15,17 that have adopted an ability-based model for EI suggest that EI is a facet of intelligence which is mildly correlated with general mental ability and that it is developmental in nature. This allows it to increase with age and life experiences.

For individual doctors who are required to spend considerable time in intensive involvement with patients and other health professionals, it is important to understand how EI is associated with burnout, a syndrome of emotional exhaustion, depersonalisation and a reduced sense of personal accomplishment.¹⁸ Some studies have used a multi-dimensional perspective¹⁹ or conservation of resources theory²⁰ to explain the stressors encountered from internal and external sources. Burnout was found to be associated with negative outcomes at both the individual level (e.g. thoughts of dropping out of medical school²¹ and decreased job satisfaction²²) and the organisational level (e.g. diminished productivity, absenteeism, decreased satisfaction with quality of care²³). Doctors who achieve higher satisfaction ratings from patients may experience less burnout and higher levels of job satisfaction. The links between EI, patient satisfaction, doctor burnout and doctor job satisfaction are still unclear in the literature. Therefore, the current study was designed to examine the factors associated with job satisfaction using a health-based model.

Methods

Design, samples and setting

In this observational study, 110 internists and 2872 of their 3682 patients were interviewed between July 2006 and August 2007 (Table 1). Each of the 110 doctors filled out three questionnaires to self-evaluate their own EI, work burnout and level of job satisfaction. Patients were interviewed in face-to-face patient satisfaction interviews conducted by seven nurses in out-patient departments. The protocol for this study was reviewed and approved by the institutional review boards of two hospitals in Taiwan. Informed written consent was obtained from each patient and attending doctor.

Table 1.   Descriptive analysis of variables, reliability and constructs

Variable	Respondent	Cronbach’s α reliability	Mean	SD	Min	Max
SEA = self emotion appraisal; OEA = others’ emotion appraisal; UOE = use of emotion; ROE = regulation of emotion; EE = emotional exhaustion; D = depersonalisation; RPA = reduced personal accomplishment; JS = job satisfaction; PS = patient satisfaction
Emotional intelligence
SEA	Doctors (n = 110)	0.89	5.94	0.81	3	7
OEA		0.91	5.10	0.92	2.25	7
UOE		0.87	5.44	0.80	2.75	7
ROE		0.92	5.22	0.97	1.50	7
Burnout
EE	Doctors (n = 110)	0.79	3.41	1.13	1	6.13
D		0.54	2.59	0.90	1	5
RPA		0.71	2.84	0.68	1	4.75
Job satisfaction
JS1	Doctors (n = 110)	0.92	5.49	0.98	3	7
JS2			5.46	1.17	2	7
JS3			5.21	1.15	2	7
Patient satisfaction
PS1	Patients (n = 2872)	0.88	5.94	0.37	4.61	6.64
PS2			5.81	0.47	3.85	6.64

Instruments

Doctor EI was measured by the Wong and Law Emotional Intelligence Scale (WLEIS),¹⁵ a self-report measure used to determine a doctor’s EI. Adopting Mayer and Salovey’s⁶ definition, this instrument covers the ability to understand one’s own emotions, to understand the emotions of others, to regulate one’s own emotions and to use one’s own emotions. Although this is a newly developed scale of EI, it has been used in several studies.^8,24 Through the use of multitrait-multimethod (MTMM) analyses, the construct validity of self- and others’ ratings has been established. The overall scale score has been shown to be reliable (Cronbach’s α = 0.78) and predictive of job performance and satisfaction.¹⁵ In addition, EI in the WLEIS was found to be conceptually distinct from personality as measured by the five-factor model.¹⁵ The WLEIS contains such items as: ‘I really understand what I feel’ and ‘I have good control of my own emotions.’ These items are scored on a 7-point Likert scale (1 = strongly disagree, 7 = strongly agree). Burnout was assessed using the Maslach Burnout Inventory (MBI).¹⁸ The MBI assesses the three aspects of burnout, namely: emotional exhaustion; depersonalisation, and a reduced sense of personal accomplishment. This questionnaire contains such items as: ‘I feel emotionally drained from my work,’‘I feel I treat some patients as if they were impersonal objects’ and ‘I feel patients blame me for some of their problems.’ Job satisfaction was measured using three self-designed items rated on a 7-point Likert scale ranging from strongly disagree to strongly agree. This questionnaire contains such items as: ‘All in all, I am satisfied with my job,’‘In general, I do not like working here’ and ‘In general, I like working here.’ A meta-analysis of 79 samples reported that this measure yielded acceptable levels of reliability (0.84).²⁵ In the current study, Cronbach’s alpha for doctor variables ranged from 0.54 to 0.92, which was mostly acceptable (Table 1).

Patients provided information on their satisfaction with their doctors. Two purpose-designed questions were used to measure patient satisfaction: ‘I am satisfied with the care provided by my doctor’ and ‘I would recommend this doctor to my friends and family members.’ These two questions were administered in face-to-face interviews conducted by nurses especially trained to carry out these interviews. Both were scored on scales of 1 (totally disagree) to 7 (strongly agree). For children aged < 8 years, the questionnaires were completed by their caregivers. Cronbach’s alpha (0.88) was acceptable for the patient variables. Emotional intelligence, job satisfaction and job satisfaction constructs were reverse-coded prior to analysis (e.g. higher scores represented higher job satisfaction).

Statistical analysis

Descriptive analyses, normality tests and correlation tests were performed using spss Version 12 (SPSS, Inc., Chicago, IL, USA). A p-value of ≤ 0.01 was considered significant. This study focused on the associations among constructs, as well as the relationships between the observed variables and their underlying factors. Confirmatory factor analysis (CFA) and second-order CFA using lisrel Version 8.71 (Scientific Software International, Inc., Lincolnwood, IL, USA) were performed. The unit for CFA analysis was the individual doctor. Emotional intelligence, burnout and job satisfaction for doctors represented ordinal scale measurements. Respondents were presented with statements to be rated on agreement scales of 1–7. For the variable of patient satisfaction, the patient data were averaged for each doctor. The mean number of patients interviewed for each doctor was 26 (range: 10–39). A within-group inter-rater reliability statistic (rwg) > 0.80 was acceptable for aggregation.²⁶ The average rwg for all 110 doctors was 0.85. As the observed variables were a combination of ordinal and interval scales, polyserial correlation using weighted least squares (WLS) was used.²⁷Figure 1 shows the CFA hypothesis that EI can be explained by four factors. Because of the limitations imposed by the sample size and estimated parameters,²⁸ the 16 questionnaire items on EI were grouped into four item parcels and the 22 questionnaire items on burnout were grouped into three item parcels. Patient satisfaction and job satisfaction were covered by two and three questionnaire items, respectively. Data for 10 of the 2872 patients were incomplete. Missing data were replaced by averages for all respondents. There were no missing data for the doctor sample. Correlations among four constructs were obtained from the phi matrix in CFA analysis. The phi coefficient, not conventionally a form of correlation between two binary variables, is a special case of correlation between two constructs in structural equation modelling. By contrast with traditional significance testing, the non-significant chi-squared statistic (p > 0.01) indicates that the predicted model was congruent with the observed data. A root mean square error of approximation (RMSEA) value < 0.05 and a goodness-of-fit index (GFI) value > 0.90 indicated a sound fit in the CFA.²⁵

Figure 1.  Results of confirmatory factor analysis. Ellipses represent latent variables; rectangles represent measured variables. Curved solid lines between ellipses show significant correlations between two latent variables. Curved dashed lines represent non-significant correlations. Solid lines with a single arrowhead represent standardised factor loadings. SEA = self emotion appraisal; OEA = others’ emotion appraisal; UOE = use of emotion; ROE = regulation of emotion; PS = patient satisfaction; JS = job satisfaction; GFI = goodness-of-fit index; RMSEA = root mean square error of approximation. * p < 0.01, ^† p < 0.001

Results

The mean ± standard deviation (SD) age of the doctors was 40.78 ± 6.91 years. Most were male (85.4%) and most held a bachelor’s degree in medicine (86.4%). The most common specialty of the doctors was gastroenterology (n = 21). All doctors from the paediatrics, nephrology and neurology departments in the participating hospitals took part in the study. Only one doctor from a rheumatology department participated. Out of 3682 patients, 2872 agreed to participate in the interview (response rate of 78.0%). Of these, 1686 were men (58.7%) and 1186 were women (41.3%). Their mean ± SD age was 45.51 ± 24.50 years. A total of 32.0% of the patients had completed elementary school or less. The majority of participants were recruited from the gastroenterology clinic (19.3%) and the smallest sample from the rheumatology clinic (0.9%; not shown in Table 1). Descriptive analyses of all observed variables are presented in Table 1. As no variable had a t-value > 3.75,²⁹ measures of skewedness and kutosis are omitted.

As Table 2 shows, older doctors appeared to have higher levels of job satisfaction (p < 0.01–0.001). Doctor EI was not related to patient satisfaction. Emotional intelligence was found to be correlated with almost all sub-dimensions of burnout (p < 0.01–0.001) and job satisfaction (p < 0.01–0.001). Higher EI was shown to be correlated with less burnout and higher job satisfaction. Patient satisfaction was shown to be negatively correlated with one sub-dimension of burnout (depersonalisation, p < 0.0.01). Finally, less doctor burnout was significantly correlated with higher job satisfaction (p < 0.01–0.001). As Fig. 1 shows, evidence of convergent validity was reflected in the high magnitude of the standardised factor loadings (0.66–0.81) at a significant level (p < 0.001) for all four constructs of EI. Findings that the estimated correlations between the factors were not excessively high (e.g. > 0.85)²⁹ indicate discriminate validity. Higher doctor EI was correlated with less burnout (ϕ = − 0.73, p < 0.001) and higher job satisfaction (ϕ = 0.52, p < 0.001). Doctors who achieved higher levels of patient satisfaction perceived less burnout in themselves (ϕ = − 0.31, p < 0.001). In addition, work-related burnout was negatively associated with job satisfaction (ϕ = − 0.50, p < 0.001). The independent model that tested the hypothesis that all latent variables were uncorrelated was easily rejected (χ² = 135.85, d.f. = 49, p < 0.001, GFI = 0.83, RMSEA = 0.13). Support was found for the hypothesised model (χ² = 49.95, d.f. = 43, p = 0.22, GFI = 0.93, RMSEA = 0.04). A chi-squared test (△χ² = 95.9, d.f. = 6, p < 0.001) indicated a significant improvement in fit for the hypothesised model over the independent model.

Table 2.   Correlation matrix of estimated parameters (n = 110)

Variable	Age	Self-ratings					PS1	PS2	Work burnout			JS1	JS2	JS3
		SEA	OEA	UOE	ROE	Total			EE	D	RPA
*p < 0.01; ^†p < 0.001 SEA = self emotion appraisal; OEA = others’ emotion appraisal; UOE = use of emotion; ROE = regulation of emotion; PS = patient satisfaction; EE = emotional exhaustion; D = depersonalisation; RPA = reduced personal accomplishment; JS = job satisfaction
Age	1
Self-ratings
SEA	0.10	1
OEA	0.03	0.55^†	1
UOE	−0.01	0.57^†	0.51^†	1
ROE	0.04	0.56^†	0.52^†	0.61^†	1
Total	0.05	0.81^†	0.80^†	0.81^†	0.84^†	1
Patient satisfaction
PS1	0.02	0.01	0.13	0.09	0.07	0.09	1
PS2	0.14	0.12	0.17	0.16	0.14	0.18	0.87^†	1
Work burnout
EE	− 0.25	− 0.31*	− 0.04	− 0.41^†	− 0.32*	− 0.32*	− 0.13	− 0.20	1
D	− 0.24	− 0.31*	− 0.24	− 0.32*	− 0.31*	− 0.36^†	− 0.26*	− 0.28*	0.47^†	1
RPA	− 0.18	− 0.49^†	− 0.39*	− 0.55^†	− 0.41^†	− 0.56^†	− 0.14	− 0.21	0.54^†	0.52^†	1
Job satisfaction
JS1	0.32*	0.21	0.18	0.44*	0.32*	0.35^†	− 0.16	− 0.06	− 0.43^†	− 0.21	− 0.42^†	1
JS2	0.39^†	0.30*	0.26*	0.47^†	0.47^†	0.46^†	− 0.23	− 0.09	− 0.39^†	− 0.23*	− 0.43^†	0.82^†	1
JS3	0.34^†	0.28*	0.27*	0.47^†	0.45^†	0.45^†	− 0.16	− 0.06	− 0.44^†	− 0.31*	− 0.40^†	0.71^†	0.86^†	1

Discussion

Work–life benefits of EI

What makes the present study unique is its specification of EI as a factor that contributes towards understanding doctors’ work-related issues. Most previous research examined only one subset of these outcomes and few have theoretically integrated both in a single model. Doctor EI is found to be positively correlated with less burnout and higher levels of job satisfaction. Less burnout is found to be significantly correlated with higher patient satisfaction and higher job satisfaction.

The significant link between findings on doctor burnout and patient satisfaction, derived from data drawn from two different sources, reinforces the robustness of the study results. Reliance on doctors’ self-ratings may result in single-rater bias. If patient satisfaction had been rated by the doctors themselves, it would have represented the doctors’ own perceptions of patient satisfaction and would have been much less meaningful. Our findings are consistent with the results of a study that suggested that high levels of burnout in doctors and nurses were associated with poor levels of satisfaction in patients on dialysis units.²³ Our study indicates that lower doctor burnout is not only consistently beneficial in achieving higher levels of patient satisfaction at the organisational level, but it also consistently contributes, with EI, towards increasing doctor job satisfaction at the individual level. Mayer and Salovey⁶ proposed that, in theory, EI consists of the abilities to identify, understand, harness and regulate emotions in oneself and in others. A doctor with higher EI is more likely to have better skills in stress management in the workplace at the individual level. In other words, EI may act as a protecting factor that reduces burnout. Accordingly, significant associations between EI and burnout, as well as job satisfaction, were found in the present study, as hypothesised. The positive contribution of a negative response (i.e. burnout) to EI may be an important issue for future research. Too much burnout may hurt doctors’ well-being, but little or no burnout may have no effect on the development of EI. Future research using a longitudinal study design may help to examine the mechanism by which an individual transforms the negative emotion of burnout into something positive.

No association between EI and patient satisfaction was found in the present study. Previous studies of EI have demonstrated inconsistent findings about the association between EI and patient (or customer) satisfaction.^7,12,30 One study using Bar-On’s measure found that only one subscale of EI (happiness) was related to higher levels of patient satisfaction.³⁰ Another study indicated that the nurse-rated EI of a doctor was positively correlated with patient satisfaction; however, there was no association between self-rated doctor EI and patient satisfaction.¹² The study that found a positive association used the self-perceived quality of social relationships as the measure of satisfaction.⁷ These three studies used different samples (residents versus attending doctors), different measurement instruments of EI (Emotion Quotient Inventory [EQi], MSCEIT, WLEIS), different raters (self versus external) and different sources (patients versus doctors). According to EI theory, a doctor who manifests these abilities is likely to perceive the patient’s needs and immerse him- or herself in the patient’s world. This is supposed to lead to higher patient satisfaction. Given the inconsistent findings of previous studies, it is unclear as to what extent EI is domain- or context-specific. Studies have suggested that patient satisfaction is associated with patient demographics, clinician courtesy, the doctor’s technical skills, hospital service, timing of the measurement and symptoms or functional improvement.^31,32 For example, there may be little possibility that a patient with an adverse outcome will give a high score for satisfaction to a doctor with good manners. Doctor EI may be only one of the factors that affect patient satisfaction. Other factors may be more impactful than doctor EI and this may explain why no correlation between doctor EI and patient satisfaction was found in this study.

Similar reasons may explain the lack of correlation between patient satisfaction and doctor job satisfaction. From the doctor’s perspective, patient satisfaction is only one facet of measuring job performance. Other parameters may include publications, research achievements, salaries, merit raises or bonuses from insurance reimbursement. Rewards such as the opportunity for self-expression and intellectual challenge may differ by doctor specialty or individual career plan. Doctors may have different priorities in pursuing such achievements. Well-being refers not only to psychological well-being, but also to physical well-being. More empirical research utilising a positive approach is needed to provide evidence that EI may benefit job performance, as well as individual well-being.

EI and doctor demographics

This study also found that older doctors had less burnout and higher levels of job satisfaction. Age and gender did not show any correlation with patient satisfaction, but male doctors tended to have higher job satisfaction than female doctors. The small size of the sample of female doctors (n = 16) precluded any multiple-group comparison using CFA to test the moderating effect of gender. Some studies have indicated that EI develops with age and can be increased by life experiences.^6,15 However, a number of questions pertain to such research. Do doctors use different standards to view their own EI? Are some doctors with poor self-ratings seen as having good control of their tempers when others compare them with other doctors? Is it possible that burnout affects EI as age increases? The associations among demographics, EI and burnout need further examination.

Educational implications

The study results have some implications for medical education. Higher EI reduces burnout at the individual level and less burnout is associated with higher levels of patient satisfaction at the organisational level. Using a positive focus on what is right, educational programmes for the promotion of doctor wellness should be developed. In order to motivate the doctor to take the initiative, workshops could be advertised as pertaining to emotional management rather than stress management or burnout prevention. Giving the workshop a positive name that is associated with good health might direct participants to expect positive consequences and enhance the effectiveness of learning. Participants might hesitate about being labelled as ‘burned out’ and not attend. Finally, the trainers’ mindsets should be altered and teaching tools and lesson plans re-designed to incorporate more innovative techniques. For example, educational programmes to teach communication skills may need to be tailored to doctors’ specific interests and needs. These programmes could also incorporate strategies for stress management and the concept of risk management in order to avoid negative and potentially adverse legal actions for malpractice.

Study limitations

This study has some limitations. There may have been some selection bias as doctors who knew they had good relationships with patients may have been more likely to agree to participate. Some selection bias may also have occurred at the patient level. Some patients seek opportunities to lodge complaints, whereas others avoid making criticisms. Furthermore, patients who value doctor–patient relationships may have been more interested in participating than those who do not. Patients who have had negative experiences or have been involved in malpractice actions against certain doctors may have been excluded from the sample. The study may also have been limited by the behaviour of participating doctors, who may have made extra effort to communicate better with their patients and to develop more supportive relationships because they knew they were being evaluated, thus initiating a Hawthorne effect. In addition, our study covered 15 medical specialties. Differences within the field could not be considered because the sample sizes for each specialty were too small to make statistically significant discriminations. Therefore, caution must be applied in any attempt to generalise the findings of this study to other settings.

Conclusions

Using a positive perspective, this study assessed the factors associated with job satisfaction. It found that higher EI was significantly associated with less burnout and higher job satisfaction. In addition, less burnout was not only associated with higher levels of patient satisfaction, but also with higher levels of job satisfaction. Given the multi-dimensional nature of EI, a refinement of the definition of EI and an assessment of the validity of EI as rated by others are needed.

Contributors:  H-CW and Y-JC contributed to the conception and design of the study, the acquisition, analysis and interpretation of data, statistical analysis and the drafting of the manuscript. Y-TL contributed to the conception and design of the study, the acquisition, analysis and interpretation of data and statistical analysis. C-MH, C-YY, C-CC and C-KH contributed to the conception and design of the study, and the acquisition of data. Y-JC and C-CC provided administrative, technical or material support. H-CW obtained funding. All authors contributed to the critical revision of the article and approved the final manuscript for submission.

Acknowledgements:  the authors would like to thank all the doctors who participated in this study. We also thank H-J Chen md, L Kang md, W-H Chen md, S-C Chi rn, W-W Feng md, L-J Lin md, L-C Hsiao ms, D-S Perng md, Y-T Su md, H-B Wu md, L-C Chang rn and Z-H Feng rn, for assistance with data collection, and H-C Chen md and T-J Yu md for their support and advice on this project.

Funding:  this study was funded in part by the Taiwan, National Science Council (NSC97-2745-H-214-004-HPU-1-005-3; NSC 98-2628-S-214 -001-MY3).

Conflicts of interest:  none.

Ethical approval:  this study was approved by the institutional review boards of E-Da Hospital, Yan-chau Shiang, Kaohsiung (E-MRP-095-010) and Kaohsiung Armed Forces General Hospital, Kaohsiung City, Taiwan.