Table 1

Study descriptions and reported patient safety and acceptability related quality of care outcomes

Author(s)Study populationDescription of sampleBurnout measureQuality of care measureQuality of care outcomes
Medical errors (ME)Patient satisfaction (PS)/quality of care (QoC)Communication/attitudes
Anagnostopoulos et al 31
Greece
Physicians working in three large primary healthcare centres.
Patients of participating physicians. Patients selected through systematic random sampling—1:3 consecutive patients.
Physician response rate: 85.8%
Patient response rate: Not reported
n=30 physicians
<10 years practicing: 53%
Specialties:
General practitioners: 63%
Pathologists/internists: 23.3%
Male: n=17
Female: n=13
>50 years: 43%
26–50 years: 40%
n=300 patients
Male: 46%
Female: 54%
Mean age: 54±15 years
Greek translation of the 22-item Maslach Burnout Inventory (MBI)-Human Services SurveyPatient report:
Patient satisfaction assessed using 18-item Consultation Satisfaction Questionnaire (CSQ).52 5-point Likert scale from 1=‘strongly agree’ to 5=‘strongly disagree’.
Satisfaction subscales: (1) general, (2) perceived length of consultation, (3) depth of relationship, and (4) professional care provided
Overall satisfaction: sum of all items (max score=90)
Scale was translated into Greek using back-translation and pilot testing.
English version’s psychometric properties tested.52 63
Correlation between MBI dimensions and PS:
  • EE and PS: r=−0.64, p<0.01

  • DP and PS: r=−0.54, p<0.01

  • PA and PS: r=0.26, p=0.17

Results of mixed effect model with PS as outcome:
  • Low EE associated with highest average PS

Comparison between moderate and high EE: no significant difference in association with PS
Halbesleben and Rathert32
USA
Attending physicians of university students who had been hospitalised in past year.
Student response rate: Not reported.
n=178 physicians
Years practicing: Not reported
Specialties: Not reported
Male: n=84
Female: n=94
Mean age=46±13 years
n=178 patients
Male: n=98
Female: n=80
Mean age: 23±5 years
22-item MBI-Human Services Survey modified to apply to patients rather than general care recipientsPatient report:
Patient satisfaction assessed using 22-item SERVQUAL.50
7-point Likert scale from 1=‘strongly disagree’ to 7=‘strong agree’.
Psychometric properties tested but subsequent study suggested need for further exploration regarding its validity.51
Correlation between MBI dimensions and PS:
  • DP and PS: r=−0.16, p<0.05

Hayashino et al 30
Japan
Members of a panel of 6459 hospital-based physicians recruited through hospital lists and scientific meetings. A randomly selected subsample of 1198 were invited to participate.
Response rate: 70%
n=836 physicians
Years practicing: Not reported
Male: 92%
Female: 8%
28–39 years: 23%
40–49 years: 47%
50–59 years: 26%
60–81 years: 4%
17-item MBI developed for Japanese healthcare professionals based on the MBI-Human Services Survey
Used burnout thresholds:
EE: >21
DP: >18
PA: >16
Physician report:
Perceived medical errors assessed with questions: ‘Are you concerned that you have made any major medical mistakes in the last year?’ If ‘yes’, asked about number of medical errors that concerned respondent.
Psychometric properties not tested.
Association between MBI dimensions and any medical error:
  • Significant differences among tertiles for EE (p=0.026) and DP (p=0.002)

% with ME by burnout dimension tertile:
  • EE 1st tertile: 27.9%

  • EE 2nd tertile: 38.2%

  • EE 3rd tertile: 33.9%

  • DP 1st tertile: 35.0%

  • DP 2nd tertile: 27.8%

  • DP 3rd tertile: 37.2%

  • No significant differences among tertiles for PA (p=0.67)

Klein et al 5
Germany
Physicians in surgery working in >100 beds general hospitals with a general surgical and/or gynaecological ward.
Stratified probability sample based on hospital beds.
Response rates:
Hospital level: 53%
Physician level: 36%
Physicians in participating hospitals: 65%
n=1311 physicians
Mean years practicing: 11 years
Male: 60%
Female: 40%
Mean age=45 ± 8.5 years
Copenhagen Burnout Inventory (CBI). Three scales assessing personal, client and work burnout.
This study focused on personal burnout (ie, degree of physical and psychological fatigue and exhaustion).
Physician report:
Perceived quality of care assessed using short version of Chirurgisches Qualitässiegel. Created three subscales: (1) psychosocial care, (2) diagnosis/therapy and (3) quality assurance. 5-point Likert scale from 1=‘very good’ to 5=‘bad’.
Two questions about frequency of diagnostic and therapeutic errors: ‘I have made mistakes in diagnosis.’ and ‘I have made mistakes in treatment.’ 4-point Likert scale (‘never’ to ‘often’).
Psychometric properties not tested for either set of questions.
Adjusted* ORs (95% CI) for probability of error and high burnout score:
  • Diagnostic error: 1.66 (1.26, 2.20)

  • Therapeutic error: 1.94 (1.39, 2.69)

*Adjusted for gender, occupational position, job experience
Adjusted* ORs (95% CI) for probability of suboptimal care and high burnout score:
  • Psychosocial care=1.58 (1.19, 2.10)

  • Dx/Tx=1.59 (1.17, 2.16)

  • Quality assurance=1.45 (1.10, 1.90)

*Adjusted for gender, occupational position, job experience
Rabatin et al 29
USA
Primary care physicians in New York City, Chicago and rural and urban Wisconsin
Recruited 1–6 patients/physician with diabetes, hypertension or congestive heart failure.
Response rate:
Physicians: 59.6%
Non-participants did not differ from participants in specialty, age or sex.
n=119 practices
n=449 physicians
n=1419 patient charts
Physician characteristics:
Male: n=235
Female: n=187
Mean age: 43±10 years
Specialties:
Family medicine: 47%
General internal medicine: 50%
Patient characteristics: Not reported.
Single-item measure: ‘Using your own definition of burnout… (a) I have no symptoms of burnout; (b) Occasionally I am under stress… but I don’t feel burned out; (c) I am definitely burning out and have one or more symptoms of burnout, such as physical and emotional exhaustion; (d) The symptoms of burnout that I’m experiencing won’t go away…; (e) I feel completely burned out and often wonder if I can go on…’
The question correlates with the EE dimension of MBI.42
Patient chart:
Chart audit using a standardised template to retrospectively assess over 18 months for guideline adherence, responsiveness to ‘recurrent abnormalities’ and missed drug interactions.
Reliability not reported.
Statistics not reported.
No statistically significant differences between physicians with burnout and without.
Ratanawongsa et al 33
USA
Physicians from 15 urban community-based clinics who provided primary care to adult patients enrolled in a randomised controlled trial for hypertensive minority patients.
Response rate: Not reported.
n=40 physicians
Mean years of practice: 11±7.7 years
Male: 47%
Female: 53%
Mean age: 42±8.7 years
Specialties:
Internal medicine: 83%
Family practice: 15%
General practice: 2%
n=235 patients
Male: 34%
Female: 66%
Mean age: 59±13.2 years
A 6-item scale derived from the MBI that captures the domains of EE and PA. 5-point Likert scale from 1=‘strongly agree’ to 3=‘neutral’ to 5=‘strongly agree’.
Based on terciles, burnout scores were categorised as low, average or high.
Physician report:
Physicians completed ‘brief questionnaires indicating the degree to which they knew the patient, their attitudes toward the patient in general, and their attitudes regarding the visit’.
Audiotaped encounters analysed for rapport-building communication behaviours using the Roter Interaction Analysis System. Four types of rapport identified: (1) positive, (2) negative, (3) emotional and (4) social.
Reliability and predictive validity tested.64
Adjusted* ORs (95% CI) for probability of PS with high versus low burnout:
  • PS=0.44 (0.18, 1.08), p=0.07

*Adjusted for patient health insurance, visit length, physician gender, physician international medical school graduate (IMG) status, interaction between IMG status and burnout
ORs (95% CI) for probability of negative rapport building with medium and high versus low burnout:
  • Medium: 1.85 (1.31, 2.61), p=0.001

  • High: 2.06 (1.58, 2.86), p<0.001

*Adjusted for patient health insurance, visit length, physician gender, physician IMG status, interaction between IMG status and burnout
Shanafelt et al 38
USA
American surgeons who were members of the American College of Surgeons who permitted email correspondence.
Response rate: 32%
n=7905 physicians
Specialties:
General: 41%
Cardiothoracic: 6%
Colorectal: 4%
Otolaryngology: 5%
Obstetrics/gynaecology: 1%
Oncological: 5%
Paediatric: 2%
Plastic: 4%
Transplant: 2%
Trauma: 4%
Urological: 4%
Vascular: 6%
Other: 6%
Male: 87%
Female: 13%
Median age (IQR): 51 years (43, 59)
22-item MBI-Human Services SurveyPhysician report:
Response to:
‘Are you concerned you have made any major medical error in the last 3 months?’
Psychometric properties not tested.
ORs (95% CI) for perceived medical error with MBI dimensions:
  • EE=1.048 (1.042, 1.055), p<0.0001

  • DP=1.109 (1.096, 1.122), p<0.0001

  • PA=0.965 (0.955, 0.975), p<0.0001

Shirom et al 39
Israel
Physicians from four health plans specialising in either: ophthalmology, dermatology, otolaryngology, gynaecology (community based), general surgery, cardiology (hospital based). 50% random probability sample drawn from each specialty.
Response rate: 63%
n=890 physicians
Male: 80%
Female: 20%
Median age: 52 years
12-items from the Shirom−Melamed Burnout Measure with three subscales: (1) physical fatigue, (2) cognitive weariness, and (3) emotional exhaustion.Physician report:
Physicians completed a 15-item version of the modified SERVQUAL. 5-point Likert scale from 1=‘to a very small extent’ to 5=‘to a very large extent’.
Psychometric properties of the modified version not tested.
Structural equation model examining relationships of autonomy, burnout and QoC:
  • Relationship between global burnout and QoC not significant (β=−0.12, p>0.05)

  • EE exhaustion negatively related to QoC (β=−40, p<0.05)

Travado et al 34
Italy, Spain, Portugal
Physicians recruited from cancer centres of three hospitals: two general hospitals with a cancer ward and one cancer hospital.
Convenience sample
Response rate: Not reported.
n=125 physicians
Years of practice: 15±9.4 years
Male: 47%
Female: 54%
Mean age: 42±9.7 years
22-item MBI-Human Services Survey
Used Maslach and Jackson6 cut-off scores for no/low burnout, intermediate and high burnout.
Physician report:
Communication skills assessed using two scales: (1) Self−Confidence in Communications Skills (SCSS). 12-item scale rating ability to communicate and manage a series of clinical situations. (2) Expected Outcomes of Communication (EOC). 23-item scale assessing extent to which physician perceives result of communication is positive or negative.
Psychometric testing not completed.56
Correlations between MBI burnout dimensions and communication:
SCSS
  • EE: r=−0.03, not significant

  • DP: r=−0.08, not significant

  • PA: r=0.37, p<0.01

Negative EOC
  • EE: r=−0.21, p<0.05

  • DP: r=−0.25, p<0.01

  • PA: r=0.28, p<0.01

Positive EOC
  • EE: r=0.01, not significant

  • DP: r=0.34, p<0.01

  • PA: r=−0.28, p<0.01

Weigl et al 36
Germany
Physicians working in one academic children’s hospital who were providing patient care.
Response rate: 74%
n=88 physicians
Years of practice: 8±6.7 years
Male: 47%
Female: 53%
Mean age: 37±8.6 years
Two subscales of the German version of the MBI-D: EE and DP.
High burnout defined as Mean EE score >3.5 and Mean DP >2.5
Physician report:
2-item perceived quality of care measure: ‘My workload frequently leads to reduced quality of work’ and ‘Adverse work conditions frequently lead to a loss of quality.’ 5-point Likert scale from 1=‘not at all’ to 5=‘a very great extent’.
Psychometric properties not tested for the two items taken from the German version of the MBI.
Adjusted* ORs (95% CI) for probability of low QoC with MBI dimensions (low vs high):
  • EE=0.75 (0.08, 1.42), p<0.05

  • DP=0.17 (−0.45, 0.80), not significant

*Adjusted for gender, professional tenure, clinical work environment, career stage/position
Wen et al 37
China
Physicians practicing in one of 46 hospitals in 10 provinces
n=12 tertiary hospitals
n=9 secondary hospitals
n=25 primary hospitals
In the secondary and tertiary hospitals, physicians were selected from >10 clinical departments with >10 people in the age groups:<30 years, 30–39 years, 40–49 years and >50 years.
Response rate: 89%
n=1607 total physicians
n=192 physicians from primary hospitals
n=354 physicians from secondary hospitals
n=991 physicians from tertiary hospitals
Primary hospital
Male: 54%
Female: 46%
Mean age: 37±9.9 years
Education:
<high school: 17%
Some college: 47%
Bachelors’ degree: 35%
>Master’s degree: 1.0%
Secondary hospital
Male: 53%
Female: 47%
Mean age: 36±9.4 years
Education:
<high school: 4%
Some college: 17%
Bachelors’ degree: 73%
>Master’s degree: 6%
Tertiary hospital
Male: 61%
Female: 39%
Mean age: 36±8.0 years
Education:
<high school: 1%
Some college: 3%
Bachelors’ degree: 46%
>Master’s degree: 51%
Used 15-item Chinese version of the MBI-General Survey
Respondents grouped into three categories:
(1) no burnout symptoms;
(2) some burnout symptoms;
(3) serious burnout symptoms.
Physician report:
Physicians were asked if they had made any of the following medical errors: (1) patient was harmed, (2) medication error, (3) treatment delayed and (4) incomplete or incorrect item in the patient record.
Psychometric properties not tested.
Adjusted* ORs (95% CI) for probability of any medical error with no burnout symptoms group as reference:
  • Some burnout symptoms: 1.46 (1.13, 1.89)

  • Serious burnout symptoms: 2.28 (1.63, 3.17)

*Adjusted for sex, workload and hospital type
Weng et al 35
Taiwan
Physicians working in two hospitals.
Patients of participating physicians.
Physician response rate:
Not reported.
Patient response rate: 78%
n=110 internists
Male: 85%
Female: 15%
Mean age: 41±6.9 years
n=2872 patients
Male: 59%
Female: 41%
MBI-Human Services SurveyPatient report:
Patient satisfaction assessed with two questions: ‘I am satisfied with the care provided by my doctor’ and ‘I would recommend this doctor to my friends and family members’.
Single item from the CSQ’s52 General Satisfaction subscale not validated for individual use.
Single item about recommendation was correlated with EUROPEP patient satisfaction questionnaire.53
Correlation between MBI burnout dimensions and PS:
  • EE: not significant

  • DP: negative relationship (p<0.01)

  • PA: not significant

  • DP, depersonalisation; EE, emotional exhaustion; PA, personal accomplishment.