Annals of Internal Medicine REVIEW

Evidence Relating Health Care Provider Burnout and Quality of Care

A Systematic Review and Meta-analysis
Daniel S. Tawfik, MD, MS; Annette Scheid, MD; Jochen Profit, MD, MPH; Tait Shanafelt, MD; Mickey Trockel, MD, PhD;
Kathryn C. Adair, PhD; J. Bryan Sexton, PhD; and John P.A. Ioannidis, MD, DSc

Background: Whether health care provider burnout contrib- comes (n = 17), and quality and safety (n = 74). Relations be-
utes to lower quality of patient care is unclear. tween burnout and quality of care were highly heterogeneous
(I2 = 93.4% to 98.8%). Of 114 unique burnout– quality combina-
Purpose: To estimate the overall relationship between burnout tions, 58 indicated burnout related to poor-quality care, 6 indi-
and quality of care and to evaluate whether published studies cated burnout related to high-quality care, and 50 showed no
provide exaggerated estimates of this relationship. significant effect. Excess significance was apparent (73% of stud-
Data Sources: MEDLINE, PsycINFO, Health and Psychosocial ies observed vs. 62% predicted to have statistically significant
Instruments (EBSCO), Mental Measurements Yearbook (EBSCO), results; P = 0.011). This indicator of potential bias was most
EMBASE (Elsevier), and Web of Science (Clarivate Analytics), prominent for the least-rigorous quality measures of best prac-
with no language restrictions, from inception through 28 May tices and quality and safety.
2019. Limitation: Studies were primarily observational; neither causal-
Study Selection: Peer-reviewed publications, in any language, ity nor directionality could be determined.
quantifying health care provider burnout in relation to quality of Conclusion: Burnout in health care professionals frequently is
patient care. associated with poor-quality care in the published literature. The
Data Extraction: 2 reviewers independently selected studies, true effect size may be smaller than reported. Future studies
extracted measures of association of burnout and quality of care, should prespecify outcomes to reduce the risk for exaggerated
and assessed potential bias by using the Ioannidis (excess signif- effect size estimates.
icance) and Egger (small-study effect) tests. Primary Funding Source: Stanford Maternal and Child Health
Data Synthesis: A total of 11 703 citations were identified, from Research Institute.
which 123 publications with 142 study populations encompass-
ing 241 553 health care providers were selected. Quality-of-care Ann Intern Med. 2019;171:555-567. doi:10.7326/M19-1152 Annals.org
outcomes were grouped into 5 categories: best practices (n = For author affiliations, see end of text.
14), communication (n = 5), medical errors (n = 32), patient out- This article was published at Annals.org on 8 October 2019.

H ealth care providers face a rapidly changing land-

scape of technology, care delivery methods, and
regulations that increase the risk for professional burn-
within the field by using a detailed evaluation for re-
porting biases.
Reporting biases take many forms, each contribut-
out. Studies suggest that nearly half of health care pro- ing to overrepresentation of “positive” findings in the
viders may have burnout symptoms at any given time published literature. Publication bias occurs when stud-
(1). Burnout has been linked to adverse effects, includ- ies with negative results are published less frequently
ing suicidality, broken relationships, decreased produc- or less rapidly than those with positive results (14). Se-
tivity, unprofessional behavior, and employee turnover, lective outcome reporting occurs when several out-
at both the provider and organizational levels (2– 6). comes of potential interest are evaluated, but only
Recent attention has been focused on the relation those with positive results are presented or empha-
between health care provider burnout and reduced sized (13). Selective analysis reporting occurs when
quality of care, with a growing body of primary litera- several analytic strategies are used, but those that pro-
ture and systematic reviews reporting associations be- duce the largest effects are presented. Overall, these
tween burnout and adherence to practice guidelines, biases result in an excess of statistically significant re-
communication, medical errors, patient outcomes, and sults in the published literature, threatening reproduc-
safety metrics (7–11). Most studies in this field use ret- ibility of findings, promoting misappropriation of re-
rospective observational designs and apply a wide sources, and skewing the design of studies assessing
range of burnout assessments and analytic tools to interventions to reduce burnout or improve quality (13).
evaluate myriad outcomes among diverse patient pop-
ulations (12). This lack of a standardized approach to
measurement and analysis increases risk of bias, poten-
tially undermining scientific progress in a rapidly ex- See also:
panding field of research by hampering the ability to
decipher which of the apparent clinically significant re- Editorial comment . . . . . . . . . . . . . . . . . . . . . . . . . 589
sults represent true effects (13). The present analysis Web-Only
sought to appraise this body of primary and review lit-
Supplement
erature, developing an understanding of true effects
© 2019 American College of Physicians 555
REVIEW Burnout and Quality of Care

METHODS 24). In line with our aim to look for reporting bias, we
We conducted a systematic literature review and did not expand our search beyond peer-reviewed pub-
meta-analysis to provide summary estimations of the lications and did not contact authors for unpublished
relation between provider burnout and quality of care, data. If an article presented insufficient data to calculate
estimate study heterogeneity, and explore the potential an effect size, we supplemented the information with
of reporting bias in the field. We followed the PRISMA data from subsequent peer-reviewed publications
(Preferred Reporting Items for Systematic reviews and when available; however, we still attributed these effect
Meta-Analyses) and MOOSE (Meta-analysis of Observa- sizes to the initial report. We excluded any studies that
tional Studies in Epidemiology) guidelines for method- were purely qualitative.
ology and reporting (15, 16). All investigators contributed to the development of
study inclusion and exclusion criteria. The literature re-
Data Sources and Searches view and study selection were conducted by 2 inde-
We searched MEDLINE, PsycINFO, Health and Psy- pendent reviewers in parallel (D.S.T. and either A.S. or
chosocial Instruments (EBSCO), Mental Measurements K.C.A.), with ambiguities and discrepancies resolved by
Yearbook (EBSCO), EMBASE (Elsevier), and Web of Sci- consensus.
ence (Clarivate Analytics) from inception through 28
May 2019, with no language restrictions. We used Data Extraction and Quality Assessment
search terms for burnout and its subdomains (emo- We extracted data into a standard template reflect-
tional exhaustion, depersonalization, and reduced per- ing publication characteristics, methods of assessing
sonal accomplishment), health care providers, and burnout and quality metrics, and strength of the re-
quality-of-care markers, as shown in Supplement Ta- ported relationship. Data were extracted by 2 indepen-
bles 1 to 3 (available at Annals.org). dent reviewers (D.S.T. and A.S.), with discrepancies re-
solved by consensus. We estimated effect sizes and
Study Selection precision using the Hedges g and SEs, respectively.
We included all peer-reviewed publications report- The Hedges g estimates effect size similarly to the Co-
ing original investigations of health care provider burn- hen d, but with a bias correction factor for small sam-
out in relation to an assessment of patient care quality. ples. In general, 0.2 indicates small effect; 0.5, medium
Providers included all paid professionals delivering effect; and 0.8, large effect.
outpatient, prehospital, emergency, or inpatient care, We classified each assessment of burnout as over-
including medical, surgical, and psychiatric care, to pa- all burnout, emotional exhaustion, depersonalization,
tients of any age. We chose an inclusive method of or low personal accomplishment. We also identified
identifying burnout studies, considering assessments to burnout assessments as standard if defined as an emo-
be related to burnout if the authors defined them as tional exhaustion score of 27 or greater or a deperson-
such and used any inventory intended to identify burnout, alization score of 10 or greater on the Maslach Burnout
either in part or in full. Likewise, we chose an inclusive Inventory, or as the midpoint and higher on validated
approach to identify quality-of-care metrics, including any single-item scales. We categorized quality metrics within
assessment of processes or outcomes indicative of care 5 groups— best practices, communication, medical errors,
quality. We included objectively measured and subjec- patient outcomes, and quality and safety—and reverse
tively reported quality metrics originating from the pro- coded any “high-quality” metrics such that positive effect
vider, other sources within the health care system, or pa- sizes indicate burnout's relation to poor-quality care.
tients and their surrogates. We considered medical For publications with several distinct (nonover-
malpractice allegations a subjective patient-reported lapping) study populations reported separately, we con-
quality metric. Although patient satisfaction is an impor- sidered each population separately for analytic purposes.
tant outcome, it is not consistently indicative of care qual- For publications with more than 1 outcome for the same
ity or improved medical outcomes, suggesting that it may study population, we decided to perform analyses using
be related to factors outside the provider's immediate only 1 outcome per study, ideally the specified primary
control, such as facility amenities and access to care (17– outcome. If no primary outcome was clear, we chose the
20). Thus, for the purposes of this review, we excluded first-listed outcome, consistent with reporting conventions
metrics solely indicative of patient satisfaction to reduce of presenting the primary outcome first. We considered
bias from these non–provider-related factors that may af- other outcomes secondary, excluding them from the pri-
fect satisfaction. mary analyses to avoid bias from intercorrelation but in-
We included peer-reviewed, indexed abstracts if cluding them in selected descriptive statistics and strati-
they reported a study population not previously or sub- fied analyses when appropriate.
sequently reported in a full-length article. For study
populations described in more than 1 full-length arti- Data Synthesis and Analysis
cle, we included the primary result from the paper with We calculated the Hedges g from odds ratios (di-
the earliest publication date as the primary outcome, chotomized data) by using the transformation
with any unique outcomes from subsequent articles as 冑3
secondary outcomes. We supplemented the database log共OR兲* or from correlation coefficients (unscaled
␲
searches with manual bibliography reviews from in- 2*r
cluded studies and related literature reviews (7–9, 21– continuous data) by using the transformation ,
冑1 ⫺ r2
556 Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 Annals.org
Burnout and Quality of Care REVIEW

both multiplied by a bias correction factor

N
冑
con-
sistent with published norms (25, 26). Further details
N⫺2 ification to estimate the between-study variance ␶2 (32).
We evaluated study heterogeneity using I2. Details re-
garding this meta-analytic approach are presented in
are provided in the Supplement (available at Annals the Supplement.
.org). We performed the Ioannidis test to evaluate for ex-
Most studies reported burnout as a dichotomous cess significance (33) by identifying the study popula-
variable or with unscaled effect size estimates, facilitat- tion with the highest precision (1/SE) among those with
ing the aforementioned transformations. We scaled ef- the lowest risk of bias (studies using a fully validated
fect sizes accordingly for the 6 studies reporting burn- burnout inventory with an objective quality metric). We
out only as a continuous variable in order to maintain then calculated the power of all studies to detect the
comparability, adapting our methods from published effect size of this study and compared the observed
guidelines (27, 28). On the basis of known distributions versus expected number of studies with statistically sig-
of burnout scores among providers (29 –31), we calcu- nificant results by using paired t tests. Next, we strati-
lated the difference between the mean scores of pro- fied excess significance testing by outcome category.
viders with and without burnout to average 47.6% of Because small studies may carry increased risk of
the span of the particular burnout scale used. We thus bias, we performed the Egger test to look for small-
converted effect sizes from continuous scales to the study effects (34). We regressed standard normal devi-
corresponding effect size reflecting a 47.6% change in ate (Hedges g/SE) on precision (1/SE) by using robust
scale score when needed to extrapolate to dichoto- SEs due to clustering of effect sizes at the study popu-
mized burnout. We also performed sensitivity analyses lation level.
excluding these few scaled effect sizes. Details of this
We used Stata 15.0 (StataCorp) for all analyses. All
process are presented in the Supplement.
tests were 2-sided. For summary effects, we considered
Initially, we intended to primarily perform a
2 different thresholds of statistical significance, P <
random-effects meta-analysis including all primary (or
first-listed) effect sizes, with secondary meta-analyses 0.050 and the newly proposed P < 0.005 (35, 36). We
stratified by quality metric category and by each unique made no further corrections for multiple testing.
burnout– quality metric combination. However, because This study was performed in accordance with the
of high heterogeneity in the pooled meta-analyses, we institutional review board requirements of Stanford
report only summary effects from the unique burnout– University and was classified as research not involving
quality metric combinations. We also performed sensi- human subjects.
tivity analyses limited to studies with standard burnout
assessments and those with independently observed or Role of the Funding Source
objectively measured quality-of-care markers. We used The funders had no role in study design, data col-
the empirical Bayes method with Knapp–Hartung mod- lection, analysis, interpretation, or writing of the report.

Figure 1. Evidence search and selection.

Articles identified in MEDLINE Articles identified in

and PsyclNFO (n = 6715) EMBASE (n = 3871)

Articles identified in Web of

Science (n = 3116)
Duplicate publications (n = 1999)

Titles/abstracts screened (n = 11 703)

Not relevant (n = 11 390)

Selected for full-text review (n = 313)

Excluded (n = 193)
No burnout predictor: 123
No quality outcome: 46
Review/repeat population: 16
Bibliographic reviews (n = 3) Not quantitative: 7
Not health care providers: 1

Included in final analysis (n = 123)

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 557
REVIEW Burnout and Quality of Care

Figure 2. Summary of all included burnout– quality metric combinations, showing frequency of effect size reporting (count)
and value of summary effect size (Hedges g).

Burnout Metric

t
en

en
m

m
h

h
is

is
pl

pl
n

n
tio

om

tio

om
n

n
us

us
cc

cc
io

io
ha

ha
la

la
at

at
ex

liz

ex

liz
na

na
na

na
o

o
al

al
rs

rs
n

so

so
ut

ut
pe

pe
io

io
o

er

er
ot

ot
rn

rn
w

w
ep

ep
Em

Em
Bu

Bu
Lo

Lo
D

D
Best practices
Inappropriate laboratory tests
Inappropriate timing of discharge
Suboptimal patient care practices
Inappropriate use of patient restraints
Poor adherence to infection control
Inappropriate antibiotic prescribing
Lack of close monitoring 30
Low best practice score
Neglect of work
Poor adherence to management guidelines
25
Communication

Poor communication
Low patient enablement score 20

Count
Forgetting to convey information
Low attention to patient impact
Low physcian empathy score
Not fully discussing treatment options 15
Poor handoff quality
Short consultation length
Errors 10

Self-reported medical errors 7

Self-reported medication errors 5
Quality Metric

Self-reported treatment/medication errors

Medical error score 3
Observed medical errors 1
Accident propensity
Diagnosis delay 2.0
Diagnostic errors
Observed medication errors
Self-reported impairment
1.5
Outcomes

Adverse events 1.0

Health care–associated infections
Patient falls
Length of stay 0.5
Urinary tract infections

Hedges g
Mortality
Poor pain control
HIV viral load suppression 0
Morbidity
Posthospitalization recovery time
–0.5
Quality and safety
Low quality of care –1.0
Low patient safety score
Low safety climate score
Low quality during most recent shift
Low work unit safety grade –1.5
Poor patient care quality score
Malpractice allegations
Low individual safety grade –2.0
Low safety perceptions
Near-miss reporting
Prolonged emergency department visit

RESULTS study populations included physicians (n = 71 [50%]),

The search identified 11 703 citations. Screening nurses (n = 84 [59%]), and other providers (n = 18
resulted in 313 potentially eligible publications re- [13%]) for a total of 241 553 health care providers eval-
trieved in full text—120 of which were included—plus 3 uated. Quality metrics covered inpatients (n = 122
additional publications identified by bibliography re- [86%]); outpatients (n = 62 [44%]); and adult (n = 134
view (Figure 1). Overall, we included 123 publications [94%]), pediatric (n = 93 [65%]), medical (n = 135
from 1994 through 2019 (37–159), encompassing 142 [95%]), and surgical (n = 89 [63%]) patients. Only 4
distinct study populations, as detailed in Supplement studies explicitly specified a primary outcome. Six stud-
Table 4 (available at Annals.org). The median sample ies did not provide sufficient data to derive an effect
size was 376 (interquartile range, 129 to 1417). The 142 size from the original publication but provided usable
558 Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 Annals.org
Burnout and Quality of Care REVIEW
data published in a subsequent review (39, 66, 69, 107, Results were similar when secondary effect sizes
115, 117). One research group reported results from a were included. Of the 114 distinct burnout– quality met-
single study population in 2 publications; the first pub- ric combinations, 58 (51%) had statistically significant
lished effect was considered primary, with results from summary effects greater than 0, 6 (5%) had statistically
the later publication considered secondary effects significant effects less than 0, and 50 (44%) showed no
(112, 160). difference at the P < 0.050 threshold. When the P <
Overall burnout, emotional exhaustion, and deper- 0.005 threshold was used, the respective numbers
sonalization were the primary predictors for 56, 75, and were 47 (41%), 6 (5%), and 61 (54%). Results from all
11 study populations, respectively, from a variety of sur- burnout– quality metric combinations are shown in Sup-
vey instruments, as outlined in Supplement Table 5 plement Figure 3 (available at Annals.org). Our findings
(available at Annals.org). The 50 distinct quality metrics were similar when limited to studies explicitly using
included 10 best practices, 8 communication, 10 med- standard burnout definitions, but the observed rela-
ical errors, 10 patient outcomes, and 12 quality and tionships were attenuated when limited to indepen-
safety measures (26 measured provider perception of dent or objective quality metrics, as shown in Table 1.
quality, 15 used independent or objective measures of The most precise study with low risk of bias (143)
quality, and 9 included both types of assessments). reported a small effect size (Hedges g = 0.26, analo-
As illustrated in Figure 2, 38 (33%) of the 114 dis- gous to an odds ratio of 1.5 to 1.6). Using this estimate,
tinct burnout– quality combinations were reported 3 or the Ioannidis test found an excess of observed versus
more times. The most frequently reported effect re- predicted statistically significant studies (73% observed
lated emotional exhaustion to low quality of care (n = vs. 62% predicted at the 0.050 significance threshold,
41), with most of the reported effect sizes in the quality P = 0.011) (Table 2). When stratified by quality metric
and safety and medical errors categories. Although all category, an excess of statistically significant studies was
5 categories of outcomes had estimates more fre- seen in the categories of best practices and quality and
quently relating burnout in the direction of poor quality safety. Results were similar for the P < 0.005 threshold.
of care (denoted in red in Figure 2), 7 of the 16 esti- The Egger test did not show small-study effects (inter-
mates pointing in the opposite direction were found in cept, ⫺1.32 [95% CI, ⫺3.48 to 0.85]), indicating that
the communication category. Results were similar when smaller studies did not systematically overestimate effect
limited to primary (or first-listed, when primary was not sizes (Figure 3). A funnel plot relating effect size to SE is
specified) effect sizes only (Supplement Figure 1, avail- shown in Supplement Figure 4 (available at Annals.org).
able at Annals.org).
Meta-analyses combining burnout and quality met-
rics within quality categories revealed I2 values of DISCUSSION
93.4% to 98.8%, indicating extremely high heterogene- This overview extends previous work in the field by
ity; therefore, summary effects are provided only at the including a comprehensive evaluation for reporting bi-
level of the 114 distinct burnout– quality combinations, ases in the health care provider burnout literature, en-
46 of which included primary effect sizes. Meta- compassing 145 published study populations that
analyses of these 46 combinations revealed 24 (52%) quantified the relation between burnout and quality of
with a statistically significant summary effect greater care over 25 years for 241 553 health care profession-
than 0 (burnout related to poor quality of care), 1 (2%) als. Most of the evidence suggests a relationship be-
with statistically significant summary effects less than 0 tween provider burnout and impaired quality of care,
(burnout related to high quality of care), and 21 (46%) consistent with recent reviews of various dimensions (7–
with no difference at the P < 0.050 threshold. When the 10, 22). Although the effect sizes in the published liter-
P < 0.005 threshold was used, the respective numbers ature are modestly strong, our finding of excess signif-
were 18 (39%), 1 (2%), and 27 (59%). Results are sum- icance implies that the true magnitude may be smaller
marized in Table 1, and primary effect sizes from all than reported, and the studies that attempted to lower
included studies are shown in Supplement Figure 2 the risk of bias demonstrate fewer significant associa-
(available at Annals.org). tions than the full evidence base. That only 4 studies

Table 1. Number and Direction of Summary Effect Sizes for Each Combination of Burnout and Quality Metric*

Criteria for Inclusion Burnout–Quality P < 0.050 Threshold, n (%) P < 0.005 Threshold, n (%)
Combinations, n†
Hedges g > 0‡ Hedges g < 0§ No Effect円円 Hedges g > 0‡ Hedges g < 0§ No Effect円円
Primary effects only 46 24 (52) 1 (2) 21 (46) 18 (39) 1 (2) 27 (59)
Primary and secondary effects 114 58 (51) 6 (5) 50 (44) 47 (41) 6 (5) 61 (54)
Standard burnout definitions 24 15 (62) 1 (4) 8 (33) 14 (58) 1 (4) 9 (38)
Independent/objective quality metrics 48 14 (29) 2 (4) 32 (67) 9 (19) 2 (4) 37 (77)
* Summary effect sizes obtained via empirical Bayes meta-analysis.
† Number of distinct burnout– quality combinations represented.
‡ Indicates burnout related to poor-quality care.
§ Indicates burnout related to high-quality care.
兩兩 Not significantly different from 0 at the specified P value threshold.

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 559
REVIEW Burnout and Quality of Care

Table 2. Predicted Versus Observed Significance for Primary* Effect Sizes, Among All Included Studies and Stratified by
Quality Metric Category

Category Studies, n P < 0.050 Threshold P < 0.005 Threshold

Predicted Observed P Value Predicted Observed P Value

Significance, % Significance, n (%) Significance, % Significance, n (%)
Full cohort 142 62 104 (73) 0.011 46 96 (68) <0.001
Best practices 14 12 9 (64) 0.001 2 8 (57) 0.001
Communication 5 43 3 (60) 0.67 40 3 (60) 0.63
Medical errors 32 50 20 (62) 0.169 33 15 (47) 0.182
Patient outcomes 17 64 9 (53) NP 54 9 (53) NP
Quality and safety 74 65 62 (84) <0.001 50 60 (81) <0.001
NP = not pertinent (observed smaller than predicted).
* Or first listed, when the primary effect size was not specified.

specified primary outcomes further supports the possi- care in the published literature is not a result of subop-
bility of reporting bias causing exaggerated effects. timal measures or variability in the definition of burn-
From a 2015 search of MEDLINE, Web of Science, out.
and CINAHL (EBSCO), Salyers and colleagues (9) re- Excess significance in the published literature was
ported effect sizes of r = ⫺0.26 (Hedges g = 0.54) and noted specifically for adherence to best practice guide-
r = ⫺0.23 (Hedges g = 0.47) for the relationship be- lines and for quality and safety metrics. Investigations of
tween burnout and quality and safety outcomes, re- burnout in relation to these outcomes are typically ret-
spectively. These effect sizes are somewhat larger than rospective studies of routinely collected outcome met-
those observed in the present study. However, the pre- rics in existing data sets, without preregistered proto-
vious meta-analysis also included markers of patient cols. The relative ease of defining and evaluating many
satisfaction and included only 82 studies through outcomes in many ways with these data sets increases
March 2015. More recently, a 2017 all-language search the risk for selective outcome and selective analysis re-
of MEDLINE, EMBASE, and CINAHL by Panagioti and porting, which may have contributed to excess signifi-
colleagues (10) identified 47 physician studies and re- cance. We found slightly lower effect sizes, but without
ported a more similar summary odds ratio of 1.96 for excess significance, for the patient outcomes sub-
patient safety incidents (approximate Hedges g = 0.37). group, possibly reflecting the more common use by
However, that review included 42 473 physicians (less these studies of quality metrics with little or no flexibility
than 20% of the number of providers represented here) in their definition and measurement (such as mortality
and did not include diverse health care professionals. or length of stay).
The observed relationships between burnout and In direct assessment, studies using independent or
quality of care are probably multifactorial. Providers objective quality metrics demonstrated less frequent
who have burnout may have less time or commitment significant effects. This finding is not surprising, be-
to optimize the care of their patients, may take more cause previous research suggests that current methods
unnecessary risks, or may be unable to pay attention to of objectively measuring quality of care cannot reliably
necessary details or recognize the consequences of identify certain events, such as errors in judgment,
their actions (71). Conversely, exposure to adverse pa- technical procedural mistakes, or near misses (10, 162).
tient events or recognition of poor-quality care may re- Objective metrics also are costly to measure and diffi-
sult in emotional or other psychological distress among cult to connect to an individual provider because of the
providers. This phenomenon often is referred to as sec- team-based nature of most clinical care, limiting appli-
ondary trauma, particularly in relation to sentinel events cation to smaller studies and those in which a quality
or important safety incidents, but it might also arise metric can be connected reliably to a provider. On the
from repeated minor incidents (161). The true effect other hand, subjective quality metrics may be more
sizes relating burnout and quality of care in both direc- sensitive and comprehensive but more prone to bias
tions are important to understand in order to make (for example, having burnout may create recall bias).
sound decisions regarding resource allocation and Further research is needed to determine the appropri-
study design of interventions, both to improve quality ate balance between insensitivity of objective quality
of care and to diminish burnout. metrics and potential for recall bias with subjective
Recent concerns have arisen regarding variability quality metrics.
in burnout assessment methods, and this inconsistency Our analysis found no evidence specifically for
was evident in the body of literature compiled here small-study effects, that is, small (more imprecise) stud-
(12). In this regard, the subset of studies in our analysis ies reporting larger effects than large studies. These
that used the most widely accepted “standard” burnout findings are consistent with those of previous meta-
assessment methods demonstrated a similar to slightly analyses, which traditionally evaluated for small-study
increased frequency of significant associations com- effects as a surrogate for all forms of reporting bias (9,
pared with the full evidence base. This finding suggests 10). The discrepancy between our findings of overall
that the relationship between burnout and quality of excess significance without evidence of small-study ef-
560 Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 Annals.org
Burnout and Quality of Care REVIEW
fects may highlight the insensitivity of the latter test as a widely accepted tools exist. Salyers and colleagues (9)
marker of all forms of bias. Moreover, smaller studies in created a 10-item tool to assess quality aspects in 82
this field are more likely to have objective measure- burnout and quality-of-care studies and did not identify
ments, whereas larger studies are more likely to have any relationship between study quality score and effect
subjective measurements. This would dilute the ability size.
of the small-study effect test to show a typical bias Our findings carry several important implications
pattern. for future intervention trials and observational studies.
Our study should be viewed in light of its design. For intervention trials, the potential for exaggerated
Although most included studies were cross-sectional, published effects should be considered in power calcu-
observational, and unable to determine the directional- lations to lower the risk for false-negative results (type II
ity of a causal relationship, longitudinal studies suggest error). In addition, future studies should attempt to re-
bidirectional causality (62, 149, 151, 152). Although 2 duce the risk of reporting biases. Standardization and
independent reviewers conducted extensive searches, consensus on core outcomes may be useful for future
they may have missed some relevant studies. Burnout studies if appropriate targets can be identified (164).
has several important outcomes beyond its effects on Such standardization may improve comparability
quality of care that were not the focus of our analysis among studies, facilitating traditional meta-analysis es-
(2– 6). Finally, excess significance may be a result of timates of the relevant effect sizes. Some outcomes,
genuine heterogeneity of effects across studies rather such as self-reported medical errors, low quality of
than reporting bias (33). The effects reported here rep- care, and low patient safety score, are particularly prev-
resent the results of heterogeneous studies; therefore, alent in the literature, suggesting that researchers al-
we do not report a single summary effect size. Rather, ready consider these outcomes either important or fea-
we report frequencies of significant summary effect sible to measure. However, if core outcomes are to be
sizes within burnout– quality metric combinations to widely accepted, they must be both important and fea-
provide a quantitative framework for interpretation sible to measure. Thus, in addition to this “popular
while acknowledging that a distribution of true effect vote” approach, expert consensus is needed to curate
sizes is expected in this field-wide assessment, in con- an appropriate list of core outcomes for this field. Other
trast to a traditional meta-analysis (163). outcome evaluations might then be discouraged unless
We avoided scoring quality assessments of the in- a unique justification is present.
cluded studies, choosing instead to analyze key aspects Study registration may further reduce the risk of
of study quality, as suggested by the proposed report- study publication bias and increase transparency of un-
ing guidelines for meta-analyses of observational stud- published studies. By registering a study publicly at its
ies (16). Judging the quality of mostly cross-sectional outset, researchers can reduce the likelihood that a
observational studies is notoriously difficult, and no study was conceived and conducted but remains un-

Figure 3. Standard normal deviate (Hedges g/SE) in relation to precision (1/SE).

80
95% CI
Fitted values
60
Standard Normal Deviate

40

20

–20

0 20 40 60 80
Precision
Parameter Robust
Estimate SE 95% CI P Value
Intercept –1.32 1.10 –3.48 to 0.85 0.23
Slope 0.54 0.10 0.33 to 0.75 <0.001

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 561
REVIEW Burnout and Quality of Care

published because of undesirable or lackluster results Corresponding Author: Daniel S. Tawfik, MD, MS, 770 Welch
(165). In a similar manner, protocol prespecification Road, Suite 435, Palo Alto, CA 94304; e-mail, dtawfik
may reduce the risk for selective outcome and selective @stanford.edu.
analysis reporting within published studies, allowing
easier identification of any post hoc analyses. Published Current author addresses and author contributions are avail-
analyses that deviate from the prespecified protocol able at Annals.org.
would require justification from the authors, and this
approach would alert the readers that those results
may be more susceptible to bias. Currently, these
mechanisms are used rarely in any field of medicine References
1. Shanafelt TD, West CP, Sinsky C, et al. Changes in burnout and
outside clinical trials, but they could become widely ad-
satisfaction with work-life integration in physicians and the general
opted with sufficient advocacy by researchers, publish- US working population between 2011 and 2017. Mayo Clin Proc.
ers, funders, and other stakeholders. 2019. [PMID: 30803733] doi:10.1016/j.mayocp.2018.10.023
In conclusion, burnout among health care provid- 2. Shanafelt TD, Boone SL, Dyrbye LN, et al. The medical marriage: a
ers is frequently associated with reduced quality of care national survey of the spouses/partners of US physicians. Mayo Clin
in the published literature. However, few rigorous stud- Proc. 2013;88:216-25. [PMID: 23489448] doi:10.1016/j.mayocp.2012
ies exist, and the effect size may be smaller than report- .11.021
3. Shanafelt TD, Mungo M, Schmitgen J, et al. Longitudinal study
ed—and may be particularly smaller for objective quality
evaluating the association between physician burnout and changes
measures. Whether curtailing burnout improves quality in professional work effort. Mayo Clin Proc. 2016;91:422-31. [PMID:
of care, or whether improving quality of care reduces 27046522] doi:10.1016/j.mayocp.2016.02.001
burnout, is not yet known, and adequately powered 4. Windover AK, Martinez K, Mercer MB, et al. Correlates and out-
and designed randomized trials (91, 166, 167) will be comes of physician burnout within a large academic medical center.
indispensable in answering these questions. JAMA Intern Med. 2018;178:856-858. [PMID: 29459945] doi:10
.1001/jamainternmed.2018.0019
From Stanford University School of Medicine, Stanford, Cali- 5. Hamidi MS, Bohman B, Sandborg C, et al. Estimating institutional
physician turnover attributable to self-reported burnout and associ-
fornia (D.S.T., T.S., M.T.); Brigham and Women's Hospital and
ated financial burden: a case study. BMC Health Serv Res. 2018;18:
Harvard Medical School, Boston, Massachusetts (A.S.); Stan-
851. [PMID: 30477483] doi:10.1186/s12913-018-3663-z
ford University School of Medicine, Stanford, California, and 6. van der Heijden F, Dillingh G, Bakker A, et al. Suicidal thoughts
California Perinatal Quality Care Collaborative, Palo Alto, Cal- among medical residents with burnout. Arch Suicide Res. 2008;12:
ifornia (J.P.); Duke University School of Medicine, Duke Uni- 344-6. [PMID: 18828037] doi:10.1080/13811110802325349
versity Health System, and Duke Patient Safety Center, Dur- 7. Dewa CS, Loong D, Bonato S, et al. The relationship between
ham, North Carolina (K.C.A., J.B.S.); and Stanford University physician burnout and quality of healthcare in terms of safety and
School of Medicine, Stanford University School of Humanities acceptability: a systematic review. BMJ Open. 2017;7:e015141.
and Sciences, and Meta-Research Innovation Center at Stan- [PMID: 28637730] doi:10.1136/bmjopen-2016-015141
ford (METRICS), Stanford, California (J.P.I.). 8. Hall LH, Johnson J, Watt I, et al. Healthcare staff wellbeing, burn-
out, and patient safety: a systematic review. PLoS One. 2016;11:
e0159015. [PMID: 27391946] doi:10.1371/journal.pone.0159015
Note: The lead author had full access to all data in the study 9. Salyers MP, Bonfils KA, Luther L, et al. The relationship between
and affirms that the manuscript is an honest, accurate, and professional burnout and quality and safety in healthcare: a meta-
transparent account of the study; that no important aspects of analysis. J Gen Intern Med. 2017;32:475-482. [PMID: 27785668] doi:
the study have been omitted; and that any discrepancies from 10.1007/s11606-016-3886-9
the study as originally planned have been explained. 10. Panagioti M, Geraghty K, Johnson J, et al. Association
between physician burnout and patient safety, professionalism, and
patient satisfaction: a systematic review and meta-analysis. JAMA In-
Financial Support: By the Stanford Maternal and Child Health tern Med. 2018;178:1317-1330. [PMID: 30193239] doi:10.1001
Research Institute. /jamainternmed.2018.3713
11. Rathert C, Williams ES, Linhart H. Evidence for the quadruple
aim: a systematic review of the literature on physician burnout and
Disclosures: Dr. Tawfik reports grants from Stanford Maternal patient outcomes. Med Care. 2018;56:976-984. [PMID: 30339573]
and Child Health Research Institute during the conduct of the doi:10.1097/MLR.0000000000000999
study. Dr. Profit reports grants from the Eunice Kennedy 12. Rotenstein LS, Torre M, Ramos MA, et al. Prevalence of burnout
Shriver National Institute of Child Health and Human Develop- among physicians: a systematic review. JAMA. 2018;320:1131-1150.
[PMID: 30326495] doi:10.1001/jama.2018.12777
ment during the conduct of the study and has received hon-
13. Ioannidis JP, Munafò MR, Fusar-Poli P, et al. Publication and
oraria for speaking at scientific meetings on the topic of burn-
other reporting biases in cognitive sciences: detection, prevalence,
out. Dr. Sexton reports grants from the National Institutes of and prevention. Trends Cogn Sci. 2014;18:235-41. [PMID:
Health during the conduct of the study. Authors not named 24656991] doi:10.1016/j.tics.2014.02.010
here have disclosed no conflicts of interest. Disclosures can 14. Dwan K, Gamble C, Williamson PR, et al; Reporting Bias Group.
also be viewed at www.acponline.org/authors/icmje/Conflict Systematic review of the empirical evidence of study publication bias
OfInterestForms.do?msNum=M19-1152. and outcome reporting bias - an updated review. PLoS One. 2013;
8:e66844. [PMID: 23861749] doi:10.1371/journal.pone.0066844
15. Moher D, Liberati A, Tetzlaff J, et al; PRISMA Group. Preferred
Reproducible Research Statement: Study protocol, statistical reporting items for systematic reviews and meta-analyses: the
code, and data set: Available from Dr. Tawfik (e-mail, dtawfik PRISMA statement. Ann Intern Med. 2009;151:264-9, W64. [PMID:
@stanford.edu). 19622511]

562 Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 Annals.org
Burnout and Quality of Care REVIEW
16. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observa- 36. Ioannidis JPA. The proposal to lower P value thresholds to .005.
tional studies in epidemiology: a proposal for reporting. Meta- JAMA. 2018;319:1429-1430. [PMID: 29566133] doi:10.1001/jama
analysis Of Observational Studies in Epidemiology (MOOSE) group. .2018.1536
JAMA. 2000;283:2008-12. [PMID: 10789670] 37. Abe K, Ohashi A. Development and testing of a staff question-
17. Chang JT, Hays RD, Shekelle PG, et al. Patients' global ratings of naire for evaluating the quality of services at nursing homes in Japan.
their health care are not associated with the technical quality of their JAmMedDirAssoc.2009;10:189-95.[PMID:19233059]doi:10.1016/j
care. Ann Intern Med. 2006;144:665-72. [PMID: 16670136] .jamda.2008.10.004
18. Kennedy GD, Tevis SE, Kent KC. Is there a relationship between 38. Ožvačić Adžić Z, Katić M, Kern J, et al. Is burnout in family phy-
patient satisfaction and favorable outcomes? Ann Surg. 2014;260: sicians in Croatia related to interpersonal quality of care? Arh Hig
592-8; discussion 598-600. [PMID: 25203875] doi:10.1097/SLA Rada Toksikol. 2013;64:255-64. [PMID: 23819934] doi:10.2478
.0000000000000932 /10004-1254-64-2013-2307
19. Rao M, Clarke A, Sanderson C, et al. Patients' own assessments 39. Aiken LH, Sermeus W, Van den Heede K, et al. Patient safety,
of quality of primary care compared with objective records based satisfaction, and quality of hospital care: cross sectional surveys of
measures of technical quality of care: cross sectional study. BMJ. nurses and patients in 12 countries in Europe and the United States.
2006;333:19. [PMID: 16793783] BMJ. 2012;344:e1717. [PMID: 22434089] doi:10.1136/bmj.e1717
20. Schmocker RK, Cherney Stafford LM, Winslow ER. Satisfaction 40. Angermeier I, Dunford BB, Boss AD, et al. The impact of partic-
with surgeon care as measured by the Surgery-CAHPS survey is not ipative management perceptions on customer service, medical er-
related to NSQIP outcomes. Surgery. 2019;165:510-515. [PMID: rors, burnout, and turnover intentions. J Healthc Manag. 2009;54:
30322662] doi:10.1016/j.surg.2018.08.028 127-40; discussion 141. [PMID: 19413167]
21. Chuang CH, Tseng PC, Lin CY, et al. Burnout in the intensive care 41. Baer TE, Feraco AM, Tuysuzoglu Sagalowsky S, et al. Pediatric
unit professionals: a systematic review. Medicine (Baltimore). 2016; resident burnout and attitudes toward patients. Pediatrics. 2017;139.
95:e5629. [PMID: 27977605] [PMID: 28232639] doi:10.1542/peds.2016-2163
22. Dewa CS, Loong D, Bonato S, et al. The relationship between 42. Baier N, Roth K, Felgner S, et al. Burnout and safety outcomes - a
resident burnout and safety-related and acceptability-related quality cross-sectional nationwide survey of EMS-workers in Germany. BMC
of healthcare: a systematic literature review. BMC Med Educ. 2017; Emerg Med. 2018;18:24. [PMID: 30126358] doi:10.1186/s12873
17:195. [PMID: 29121895] doi:10.1186/s12909-017-1040-y -018-0177-2
23. Scheepers RA, Boerebach BC, Arah OA, et al. A systematic re- 43. Balch CM, Oreskovich MR, Dyrbye LN, et al. Personal conse-
view of the impact of physicians' occupational well-being on the quences of malpractice lawsuits on American surgeons. J Am Coll
Surg. 2011;213:657-67. [PMID: 21890381] doi:10.1016/j.jamcollsurg
quality of patient care. Int J Behav Med. 2015;22:683-98. [PMID:
.2011.08.005
25733349] doi:10.1007/s12529-015-9473-3
44. Bao Y, Vedina R, Moodie S, et al. The relationship between value
24. West CP, Dyrbye LN, Erwin PJ, et al. Interventions to prevent and
incongruence and individual and organizational well-being out-
reduce physician burnout: a systematic review and meta-analysis.
comes: an exploratory study among Catalan nurses. J Adv Nurs.
Lancet. 2016;388:2272-2281. [PMID: 27692469] doi:10.1016/S0140
2013;69:631-41. [PMID: 22632178] doi:10.1111/j.1365-2648.2012
-6736(16)31279-X
.06045.x
25. Chinn S. A simple method for converting an odds ratio to effect
45. Basar U, Basim N. A cross-sectional survey on consequences of
size for use in meta-analysis. Stat Med. 2000;19:3127-31. [PMID:
nurses' burnout: moderating role of organizational politics. J Adv
11113947]
Nurs. 2016;72:1838-50. [PMID: 26988276] doi:10.1111/jan.12958
26. Lajeunesse M. Recovering Missing or Partial Data from Studies: A
46. Beckman TJ, Reed DA, Shanafelt TD, et al. Resident physician
Survey of Conversions and Imputations for Meta-analysis. Princeton:
well-being and assessments of their knowledge and clinical perfor-
Princeton Univ Pr; 2013.
mance. J Gen Intern Med. 2012;27:325-30. [PMID: 21948207] doi:
27. Guyatt GH, Thorlund K, Oxman AD, et al. GRADE guidelines: 13. 10.1007/s11606-011-1891-6
Preparing summary of findings tables and evidence profiles- 47. Block L, Wu AW, Feldman L, et al. Residency schedule, burnout
continuous outcomes. J Clin Epidemiol. 2013;66:173-83. [PMID: and patient care among first-year residents. Postgrad Med J. 2013;
23116689] doi:10.1016/j.jclinepi.2012.08.001 89:495-500. [PMID: 23852828] doi:10.1136/postgradmedj-2012
28. Hasselblad V, Hedges LV. Meta-analysis of screening and diag- -131743
nostic tests. Psychol Bull. 1995;117:167-78. [PMID: 7870860] 48. Boamah SA, Read EA, Spence Laschinger HK. Factors influenc-
29. Shanafelt TD, Boone S, Tan L, et al. Burnout and satisfaction with ing new graduate nurse burnout development, job satisfaction and
work-life balance among US physicians relative to the general US patient care quality: a time-lagged study. J Adv Nurs. 2017;73:1182-
population. Arch Intern Med. 2012;172:1377-85. [PMID: 22911330] 1195. [PMID: 27878844] doi:10.1111/jan.13215
30. Tawfik DS, Phibbs CS, Sexton JB, et al. Factors associated with 49. Bronkhorst B, Vermeeren B. Safety climate, worker health and
provider burnout in the NICU. Pediatrics. 2017;139. [PMID: organizational health performance: testing a physical, psychosocial
28557756] doi:10.1542/peds.2016-4134 and combined pathway. International Journal of Workplace Health
31. West CP, Shanafelt TD, Kolars JC. Quality of life, burnout, edu- Management. 2016;9:270-89.
cational debt, and medical knowledge among internal medicine res- 50. Brunsberg KA, Landrigan CP, Garcia BM, et al. Association of
idents. JAMA. 2011;306:952-60. [PMID: 21900135] doi:10.1001/jama pediatric resident physician depression and burnout with harmful
.2011.1247 medical errors on inpatient services. Acad Med. 2019;94:1150-1156.
32. Knapp G, Hartung J. Improved tests for a random effects meta- [PMID: 31045601] doi:10.1097/ACM.0000000000002778
regression with a single covariate. Stat Med. 2003;22:2693-710. 51. Chao M, Shih CT, Hsu SF. Nurse occupational burnout and
[PMID: 12939780] patient-rated quality of care: the boundary conditions of emotional
33. Ioannidis JP, Trikalinos TA. An exploratory test for an excess of intelligence and demographic profiles. Jpn J Nurs Sci. 2016;13:156-
significant findings. Clin Trials. 2007;4:245-53. [PMID: 17715249] 65. [PMID: 26542752] doi:10.1111/jjns.12100
34. Sterne JA, Sutton AJ, Ioannidis JP, et al. Recommendations for 52. Chen KY, Yang CM, Lien CH, et al. Burnout, job satisfaction, and
examining and interpreting funnel plot asymmetry in meta-analyses medical malpractice among physicians. Int J Med Sci. 2013;10:
of randomised controlled trials. BMJ. 2011;343:d4002. [PMID: 1471-8. [PMID: 24046520] doi:10.7150/ijms.6743
21784880] doi:10.1136/bmj.d4002 53. Cheng C, Bartram T, Karimi L, et al. The role of team climate in
35. Benjamin DJ, Berger JO, Johannesson M, et al. Redefine statis- the management of emotional labour: implications for nurse reten-
tical significance. Nat Hum Behav. 2018;2:6-10. [PMID: 30980045] tion. J Adv Nurs. 2013;69:2812-25. [PMID: 23834619] doi:10.1111
doi:10.1038/s41562-017-0189-z /jan.12202

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 563
REVIEW Burnout and Quality of Care

54. Cimiotti JP, Aiken LH, Sloane DM, et al. Nurse staffing, burnout, 72. Halbesleben JR, Wakefield BJ, Wakefield DS, et al. Nurse
and health care-associated infection. Am J Infect Control. 2012;40: burnout and patient safety outcomes: nurse safety perception versus
486-90. [PMID: 22854376] doi:10.1016/j.ajic.2012.02.029 reporting behavior. West J Nurs Res. 2008;30:560-77. [PMID:
55. Colindres CV, Bryce E, Coral-Rosero P, et al. Effect of effort- 18187408] doi:10.1177/0193945907311322
reward imbalance and burnout on infection control among Ecuador- 73. Hansen RP, Vedsted P, Sokolowski I, et al. General practitioner
ian nurses. Int Nurs Rev. 2018;65:190-199. [PMID: 29114886] doi:10 characteristics and delay in cancer diagnosis. a population-based co-
.1111/inr.12409 hort study. BMC Fam Pract. 2011;12:100. [PMID: 21943310] doi:10
56. Cummings GG, Estabrooks CA, Midodzi WK, et al. Influence of .1186/1471-2296-12-100
organizational characteristics and context on research utilization. 74. Hayashino Y, Utsugi-Ozaki M, Feldman MD, et al. Hope modified
Nurs Res. 2007;56:S24-39. [PMID: 17625471] the association between distress and incidence of self-perceived
57. Davenport DL, Henderson WG, Mosca CL, et al. Risk-adjusted medical errors among practicing physicians: prospective cohort
morbidity in teaching hospitals correlates with reported levels of study. PLoS One. 2012;7:e35585. [PMID: 22530055] doi:10.1371
communication and collaboration on surgical teams but not with /journal.pone.0035585
scale measures of teamwork climate, safety climate, or working con- 75. Holden RJ, Patel NR, Scanlon MC, et al. Effects of mental de-
ditions. J Am Coll Surg. 2007;205:778-84. [PMID: 18035261] mands during dispensing on perceived medication safety and em-
58. de Oliveira GS Jr, Chang R, Fitzgerald PC, et al. The prevalence
ployee well-being: a study of workload in pediatric hospital pharma-
of burnout and depression and their association with adherence to
cies. Res Social Adm Pharm. 2010;6:293-306. [PMID: 21111387] doi:
safety and practice standards: a survey of United States anesthesiol-
10.1016/j.sapharm.2009.10.001
ogy trainees. Anesth Analg. 2013;117:182-93. [PMID: 23687232]
76. Holden RJ, Scanlon MC, Patel NR, et al. A human factors frame-
doi:10.1213/ANE.0b013e3182917da9
work and study of the effect of nursing workload on patient safety
59. De Stefano C, Philippon AL, Krastinova E, et al. Effect of emer-
gency physician burnout on patient waiting times. Intern Emerg and employee quality of working life. BMJ Qual Saf. 2011;20:15-24.
Med. 2018;13:421-428. [PMID: 28677043] doi:10.1007/s11739-017 [PMID: 21228071] doi:10.1136/bmjqs.2008.028381
-1706-9 77. Huang CH, Wu HH, Chou CY, et al. The perceptions of physicians
60. Deckard G, Meterko M, Field D. Physician burnout: an examina- and nurses regarding the establishment of patient safety in a re-
tion of personal, professional, and organizational relationships. Med gional teaching hospital in Taiwan. Iran J Public Health. 2018;47:852-
Care. 1994;32:745-54. [PMID: 8028408] 860. [PMID: 30087871]
61. Dorigan GH, Guirardello EB. Effect of the practice environment 78. Huang CH, Wu HH, Lee YC. The perceptions of patient safety
of nurses on job outcomes and safety climate. Rev Lat Am Enferma- culture: a difference between physicians and nurses in Taiwan. Appl
gem. 2018;26:e3056. [PMID: 30379243] doi:10.1590/1518-8345 Nurs Res. 2018;40:39-44. [PMID: 29579497] doi:10.1016/j.apnr
.2633.3056 .2017.12.010
62. Fahrenkopf AM, Sectish TC, Barger LK, et al. Rates of medication 79. Huang EC, Pu C, Huang N, et al. Resident burnout in Taiwan
errors among depressed and burnt out residents: prospective co- Hospitals-and its relation to physician felt trust from patients. J For-
hort study. BMJ. 2008;336:488-91. [PMID: 18258931] doi:10.1136 mos Med Assoc. 2019. [PMID: 30626545] doi:10.1016/j.jfma.2018.12
/bmj.39469.763218.BE .015
63. Faivre G, Kielwasser H, Bourgeois M, et al. Burnout syndrome in 80. Johnson J, Louch G, Dunning A, et al. Burnout mediates the
orthopaedic and trauma surgery residents in France: a nationwide association between depression and patient safety perceptions: a
survey. Orthop Traumatol Surg Res. 2018;104:1291-1295. [PMID: cross-sectional study in hospital nurses. J Adv Nurs. 2017;73:1667-
30341030] doi:10.1016/j.otsr.2018.08.016 1680. [PMID: 28072469] doi:10.1111/jan.13251
64. Galletta M, Portoghese I, D’Aloja E, et al. Relationship between 81. Kang EK, Lihm HS, Kong EH. Association of intern and resident
job burnout, psychosocial factors and health care-associated infec- burnout with self-reported medical errors. Korean J Fam Med. 2013;
tions in critical care units. Intensive Crit Care Nurs. 2016;34:51-8. 34:36-42. [PMID: 23372904] doi:10.4082/kjfm.2013.34.1.36
[PMID: 26961918] doi:10.1016/j.iccn.2015.11.004 82. Kim MH, Mazenga AC, Simon K, et al. Burnout and self-reported
65. Garrouste-Orgeas M, Perrin M, Soufir L, et al. The Iatroref study: suboptimal patient care amongst health care workers providing HIV
medical errors are associated with symptoms of depression in ICU care in Malawi. PLoS One. 2018;13:e0192983. [PMID: 29466443] doi:
staff but not burnout or safety culture. Intensive Care Med. 2015;41: 10.1371/journal.pone.0192983
273-84. [PMID: 25576157] doi:10.1007/s00134-014-3601-4 83. Kirwan M, Matthews A, Scott PA. The impact of the work envi-
66. Gasparino RC, Guirardello Ede B, Aiken LH. Validation of the ronment of nurses on patient safety outcomes: a multi-level model-
brazilian version of the nursing work index-revised (B-NWI-r). J Clin ling approach. Int J Nurs Stud. 2013;50:253-63. [PMID: 23116681]
Nurs. 2011;20:3494-501. [PMID: 21749511] doi:10.1111/j.1365-2702
doi:10.1016/j.ijnurstu.2012.08.020
.2011.03776.x
84. Klein J, Grosse Frie K, Blum K, et al. Burnout and perceived
67. Gopal R, Glasheen JJ, Miyoshi TJ, et al. Burnout and internal
quality of care among German clinicians in surgery. Int J Qual Health
medicine resident work-hour restrictions. Arch Intern Med. 2005;
Care. 2010;22:525-30. [PMID: 20935011] doi:10.1093/intqhc
165:2595-600. [PMID: 16344416]
/mzq056
68. Guirardello EB. Impact of critical care environment on burnout,
perceived quality of care and safety attitude of the nursing team. Rev 85. Kwah J, Fallar R, Weintraub JP, Ripp J. The impact of job burnout
Lat Am Enfermagem. 2017;25:e2884. [PMID: 28591294] doi:10 on measures of professionalism in first-year internal medicine resi-
.1590/1518-8345.1472.2884 dents at a large urban academic medical center. J Gen Intern Med.
69. Gunnarsdóttir S, Clarke SP, Rafferty AM, et al. Front-line manage- 2014;29:S228.
ment, staffing and nurse-doctor relationships as predictors of nurse 86. Lafreniere JP, Rios R, Packer H, et al. Burned out at the bedside:
and patient outcomes. a survey of Icelandic hospital nurses. Int J patient perceptions of physician burnout in an internal medicine res-
Nurs Stud. 2009;46:920-7. [PMID: 17229425] ident continuity clinic. J Gen Intern Med. 2016;31:203-208. [PMID:
70. Gupta K, Lisker S, Rivadeneira NA, et al. Decisions and repercus- 26340808] doi:10.1007/s11606-015-3503-3
sions of second victim experiences for mothers in medicine (SAVE 87. Spence Laschinger HK, Leiter MP. The impact of nursing work
DR MoM). BMJ Qual Saf. 2019;28:564-573. [PMID: 30718333] doi: environments on patient safety outcomes: the mediating role of
10.1136/bmjqs-2018-008372 burnout/engagement. J Nurs Adm. 2006;36:259-67. [PMID:
71. Halbesleben JR, Rathert C. Linking physician burnout and patient 16705307]
outcomes: exploring the dyadic relationship between physicians 88. Laschinger H, Shamian J, Thomson D. Impact of magnet hospital
and patients. Health Care Manage Rev. 2008;33:29-39. [PMID: characteristics on nurses' perceptions of trust, burnout, quality of
18091442] care, and work satisfaction. Nursing Economic$. 2001;19(5):209-19.

564 Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 Annals.org
Burnout and Quality of Care REVIEW
89. Lewis EJ, Baernholdt MB, Yan G, et al. Relationship of adverse during the long-call shift. Health Commun. 2012;27:449-56. [PMID:
events and support to RN burnout. J Nurs Care Qual. 2015;30:144- 21970629] doi:10.1080/10410236.2011.606527
52. [PMID: 25148522] doi:10.1097/NCQ.0000000000000084 107. Patrician PA, Shang J, Lake ET. Organizational determinants of
90. Linzer M, Manwell LB, Williams ES, et al; MEMO (Minimizing work outcomes and quality care ratings among Army Medical De-
Error, Maximizing Outcome) Investigators. Working conditions in partment registered nurses. Res Nurs Health. 2010;33:99-110.
primary care: physician reactions and care quality. Ann Intern Med. [PMID: 20151409] doi:10.1002/nur.20370
2009;151:28-36, W6-9. [PMID: 19581644] 108. Pedersen AF, Carlsen AH, Vedsted P. Association of GPs' risk
91. Linzer M, Poplau S, Brown R, et al. Do work condition interven- attitudes, level of empathy, and burnout status with PSA testing in
tions affect quality and errors in primary care? results from the primary care. Br J Gen Pract. 2015;65:e845-51. [PMID: 26541183]
healthy work place study. J Gen Intern Med. 2017;32:56-61. [PMID: doi:10.3399/bjgp15X687649
27612486] doi:10.1007/s11606-016-3856-2 109. Poghosyan L, Clarke SP, Finlayson M, et al. Nurse burnout and
92. Liu X, Zheng J, Liu K, et al. Hospital nursing organizational fac- quality of care: cross-national investigation in six countries. Res Nurs
tors, nursing care left undone, and nurse burnout as predictors of Health. 2010;33:288-98. [PMID: 20645421] doi:10.1002/nur.20383
patient safety: a structural equation modeling analysis. Int J Nurs 110. Pratt M, Kerr M, Wong C. The impact of ERI, burnout, and car-
Stud. 2018;86:82-89. [PMID: 29966828] doi:10.1016/j.ijnurstu.2018 ing for SARS patients on hospital nurses' self-reported compliance
.05.005 with infection control. Can J Infect Control. 2009;24:167-72, 174.
93. Liu Y, Aungsuroch Y. Factors influencing nurse-assessed quality [PMID: 19891170]
nursing care: a cross-sectional study in hospitals. J Adv Nurs. 2018; 111. Prins JT, van der Heijden FM, Hoekstra-Weebers JE, et al. Burn-
74:935-945. [PMID: 29148146] doi:10.1111/jan.13507 out, engagement and resident physicians' self-reported errors. Psy-
94. Loerbroks A, Glaser J, Vu-Eickmann P, et al. Physician burnout, chol Health Med. 2009;14:654-66. [PMID: 20183538] doi:10.1080/
work engagement and the quality of patient care. Occup Med 13548500903311554
(Lond). 2017;67:356-362. [PMID: 28510762] doi:10.1093/occmed 112. Profit J, Sharek PJ, Amspoker AB, et al. Burnout in the NICU
/kqx051 setting and its relation to safety culture. BMJ Qual Saf. 2014;23:806-
95. Lorenz VR, Sabino MO, Corrêa Filho HR. Professional exhaustion, 13. [PMID: 24742780] doi:10.1136/bmjqs-2014-002831
quality and intentions among family health nurses. Rev Bras Enferm. 113. Qureshi HA, Rawlani R, Mioton LM, et al. Burnout phenomenon
2018;71:2295-2301. [PMID: 30365797] doi:10.1590/0034-7167 in U.S. plastic surgeons: risk factors and impact on quality of life. Plast
-2016-0510 Reconstr Surg. 2015;135:619-26. [PMID: 25357156] doi:10.1097/PRS
96. Lu DW, Dresden S, McCloskey C, et al. Impact of burnout on .0000000000000855
self-reported patient care among emergency physicians. West J 114. Rafferty AM, Ball J, Aiken LH. Are teamwork and professional
autonomy compatible, and do they result in improved hospital care?
Emerg Med. 2015;16:996-1001. [PMID: 26759643] doi:10.5811
Qual Health Care. 2001;10 Suppl 2:ii32-7. [PMID: 11700377]
/westjem.2015.9.27945
115. Ridley J, Wilson B, Harwood L, et al. Work environment, health
97. MacPhee M, Dahinten VS, Havaei F. The impact of heavy per-
outcomes and magnet hospital traits in the Canadian nephrology
ceived nurse workloads on patient and nurse outcomes. Administra-
nursing scene. CANNT J. 2009;19:28-35. [PMID: 19354155]
tive Sciences. 2017;7:7.
116. Riquelme I, Chacón JI, Gándara AV, et al; PAINBO Study
98. Martinussen M, Kaiser S, Adolfsen F, et al. Reorganisation of
Group. Prevalence of burnout among pain medicine physicians and
healthcare services for children and families: improving collabora-
its potential effect upon clinical outcomes in patients with oncologic
tion, service quality, and worker well-being. J Interprof Care. 2017;
pain or chronic pain of nononcologic origin. Pain Med. 2018;19:
31:487-496. [PMID: 28481168] doi:10.1080/13561820.2017.1316249
2398-2407. [PMID: 29361180] doi:10.1093/pm/pnx335
99. Mazurkiewicz RA, Smith KL, Korenstein D, Ripp J. The impact of
117. Rochefort CM, Clarke SP. Nurses' work environments, care ra-
resident physician burnout on the quality of care of hospitalized pa-
tioning, job outcomes, and quality of care on neonatal units. J Adv
tients. J Gen Intern Med. 2012;27:S323-S4. Nurs. 2010;66:2213-24. [PMID: 20626479] doi:10.1111/j.1365-2648
100. Mion G, Libert N, Journois D. [Burnout-associated factors in .2010.05376.x
anesthesia and intensive care medicine. 2009 survey of the French 118. Salyers MP, Fukui S, Rollins AL, et al. Burnout and self-reported
Society of Anesthesiology and Intensive Care]. Ann Fr Anesth Re- quality of care in community mental health. Adm Policy Ment Health.
anim. 2013;32:175-88. [PMID: 23395149] doi:10.1016/j.annfar.2012 2015;42:61-9. [PMID: 24659446]
.12.004 119. Schmidt SG, Dichter MN, Bartholomeyczik S, et al. The satisfac-
101. Mohr DC, Eaton JL, Meterko M, et al. Factors associated with tion with the quality of dementia care and the health, burnout and
internal medicine physician job attitudes in the Veterans Health Ad- work ability of nurses: a longitudinal analysis of 50 German nursing
ministration. BMC Health Serv Res. 2018;18:244. [PMID: 29622008] homes. Geriatr Nurs. 2014;35:42-6. [PMID: 24131899] doi:10.1016
doi:10.1186/s12913-018-3015-z /j.gerinurse.2013.09.006
102. Molina Siguero A, Garcı́a Pérez MA, Alonso González M, et al. 120. Schwartz SP, Adair KC, Bae J, et al. Work-life balance behav-
[Prevalence of worker burnout and psychiatric illness in primary care iours cluster in work settings and relate to burnout and safety culture:
physicians in a health care area in Madrid]. Aten Primaria. 2003;31: a cross-sectional survey analysis. BMJ Qual Saf. 2019;28:142-150.
564-71. [PMID: 12783745] [PMID: 30309912] doi:10.1136/bmjqs-2018-007933
103. Nantsupawat A, Nantsupawat R, Kunaviktikul W, et al. Nurse 121. Shanafelt TD, Balch CM, Bechamps G, et al. Burnout and med-
burnout, nurse-reported quality of care, and patient outcomes in thai ical errors among American surgeons. Ann Surg. 2010;251:995-
hospitals. J Nurs Scholarsh. 2016;48:83-90. [PMID: 26650339] doi: 1000. [PMID: 19934755] doi:10.1097/SLA.0b013e3181bfdab3
10.1111/jnu.12187 122. Shanafelt TD, Bradley KA, Wipf JE, et al. Burnout and self-
104. O’Connor P, Lydon S, O’Dea A, et al. A longitudinal and reported patient care in an internal medicine residency program.
multicentre study of burnout and error in Irish junior doctors. Ann Intern Med. 2002;136:358-67. [PMID: 11874308]
Postgrad Med J. 2017;93:660-664. [PMID: 28600343] doi:10 123. Shields CG, Fuzzell LN, Christ SL, et al. Patient and provider
.1136 characteristics associated with communication about opioids: an ob-
/postgradmedj-2016-134626 servational study. Patient Educ Couns. 2019;102:888-894. [PMID:
105. Panunto MR, Guirardello Ede B. Professional nursing practice: 30552013] doi:10.1016/j.pec.2018.12.005
environment and emotional exhaustion among intensive care 124. Shirom A, Nirel N, Vinokur AD. Overload, autonomy, and burn-
nurses. Rev Lat Am Enfermagem. 2013;21:765-72. [PMID: out as predictors of physicians' quality of care. J Occup Health Psy-
23918023] doi:10.1590/S0104-11692013000300016 chol. 2006;11:328-42. [PMID: 17059297]
106. Passalacqua SA, Segrin C. The effect of resident physician 125. Sillero-Sillero A, Zabalegui A. Safety and satisfaction of patients
stress, burnout, and empathy on patient-centered communication with nurse's care in the perioperative. Rev Lat Am Enfermagem.

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 565
REVIEW Burnout and Quality of Care

2019;27:e3142. [PMID: 31038636] doi:10.1590/1518-8345.2646 vey.IntJNursStud.2014;51:1123-34.[PMID:24444772]doi:10.1016/j

.3142 .ijnurstu.2013.12.009
126. Sokolova O, Pogosova N, Isakova S, et al. Impact of primary 142. Bogaert PV, Heusden DV, Slootmans S, et al. Staff empower-
care physicians burnout on their adherence to national guidelines for ment and engagement in a magnet® recognized and joint commis-
common CVD. Global Heart. 2018;13:486. sion international accredited academic centre in Belgium: a cross-
127. Squires M, Tourangeau A, Spence Laschinger HK, et al. The link sectional survey. BMC Health Serv Res. 2018;18:756. [PMID:
between leadership and safety outcomes in hospitals. J Nurs Manag. 30285735] doi:10.1186/s12913-018-3562-3
2010;18:914-25. [PMID: 21073565] doi:10.1111/j.1365-2834.2010 143. Van Gerven E, Vander Elst T, Vandenbroeck S, et al. Increased
.01181.x risk of burnout for physicians and nurses involved in a patient safety
128. Sturm H, Rieger MA, Martus P, et al; WorkSafeMed Consortium. incident. Med Care. 2016;54:937-43. [PMID: 27213542] doi:10
Do perceived working conditions and patient safety culture correlate .1097/MLR.0000000000000582
with objective workload and patient outcomes: a cross-sectional ex- 144. Vifladt A, Simonsen BO, Lydersen S, et al. The association be-
plorative study from a German university hospital. PLoS One. 2019; tween patient safety culture and burnout and sense of coherence: a
14:e0209487. [PMID: 30608945] doi:10.1371/journal.pone.0209487 cross-sectional study in restructured and not restructured intensive
129. Sulaiman CFC, Henn P, Smith S, et al. Burnout syndrome care units. Intensive Crit Care Nurs. 2016;36:26-34. [PMID:
among non-consultant hospital doctors in Ireland: relationship with 27212614] doi:10.1016/j.iccn.2016.03.004
self-reported patient care. Int J Qual Health Care. 2017;29:679-684. 145. Vogus TJ, Cooil B, Sitterding M, et al. Safety organizing,
[PMID: 28992145] doi:10.1093/intqhc/mzx087 emotional exhaustion, and turnover in hospital nursing units.
130. Sun BZ, Chaitoff A, Hu B, et al. Empathy, burnout, and antibiotic Med Care. 2014;52:870-6. [PMID: 25222533] doi:10.1097/MLR
prescribing for acute respiratory infections: a cross-sectional primary .0000000000000169
care study in the US. Br J Gen Pract. 2017;67:e565-e571. [PMID: 146. Wawrzyniak AJ, Rodriguez AE. The association between physi-
28717000] doi:10.3399/bjgp17X691901 cian burnout and satisfaction on health outcomes in HIV-infected
131. Tawfik DS, Profit J, Morgenthaler TI, et al. Physician burnout, outpatients. Psychosomatic Medicine. 2017;79:A102.
well-being, and work unit safety grades in relationship to reported 147. Weigl M, Schneider A, Hoffmann F, et al. Work stress, burnout,
medical errors. Mayo Clin Proc. 2018;93:1571-1580. [PMID: and perceived quality of care: a cross-sectional study among hospi-
30001832] doi:10.1016/j.mayocp.2018.05.014 tal pediatricians. Eur J Pediatr. 2015;174:1237-46. [PMID: 25846697]
132. Teng CI, Shyu YI, Chiou WK, et al. Interactive effects of nurse- doi:10.1007/s00431-015-2529-1
experienced time pressure and burnout on patient safety: a cross- 148. Welp A, Meier LL, Manser T. Emotional exhaustion and work-
sectional survey. Int J Nurs Stud. 2010;47:1442-50. [PMID: load predict clinician-rated and objective patient safety. Front Psy-
20472237] doi:10.1016/j.ijnurstu.2010.04.005 chol. 2014;5:1573. [PMID: 25657627] doi:10.3389/fpsyg.2014
133. Toral-Villanueva R, Aguilar-Madrid G, Juárez-Pérez CA. Burnout .01573
and patient care in junior doctors in Mexico City. Occup Med (Lond). 149. Welp A, Meier LL, Manser T. The interplay between teamwork,
2009;59:8-13. [PMID: 18796698] doi:10.1093/occmed/kqn122 clinicians' emotional exhaustion, and clinician-rated patient safety: a
134. Trockel M, Bohman B, Lesure E, et al. A brief instrument to longitudinal study. Crit Care. 2016;20:110. [PMID: 27095501] doi:10
assess both burnout and professional fulfillment in physicians: reli- .1186/s13054-016-1282-9
ability and validity, including correlation with self-reported medical 150. Wen J, Cheng Y, Hu X, et al. Workload, burnout, and medical
errors, in a sample of resident and practicing physicians. Acad Psy- mistakes among physicians in China: a cross-sectional study. Biosci
chiatry. 2018;42:11-24. [PMID: 29196982] doi:10.1007/s40596-017 Trends. 2016;10:27-33. [PMID: 26961213] doi:10.5582/bst.2015
-0849-3 .01175
135. Tsiga E, Panagopoulou E, Montgomery A. Examining the link 151. West CP, Huschka MM, Novotny PJ, et al. Association of per-
between burnout and medical error: a checklist approach. Burnout ceived medical errors with resident distress and empathy: a prospec-
Research. 2017;6:1-8. tive longitudinal study. JAMA. 2006;296:1071-8. [PMID: 16954486]
136. Van Bogaert P, Clarke S, Roelant E, et al. Impacts of unit-level 152. West CP, Tan AD, Habermann TM, et al. Association of resident
nurse practice environment and burnout on nurse-reported out- fatigue and distress with perceived medical errors. JAMA. 2009;302:
comes: a multilevel modelling approach. J Clin Nurs. 2010;19:1664- 1294-300. [PMID: 19773564] doi:10.1001/jama.2009.1389
74. [PMID: 20579204] doi:10.1111/j.1365-2702.2009.03128.x 153. Williams ES, Manwell LB, Konrad TR, et al. The relationship of
137. Van Bogaert P, Clarke S, Wouters K, et al. Impacts of unit-level organizational culture, stress, satisfaction, and burnout with
nurse practice environment, workload and burnout on nurse- physician-reported error and suboptimal patient care: results from
reported outcomes in psychiatric hospitals: a multilevel modelling the MEMO study. Health Care Manage Rev. 2007;32:203-12. [PMID:
approach. Int J Nurs Stud. 2013;50:357-65. [PMID: 22695484] doi: 17666991]
10.1016/j.ijnurstu.2012.05.006 154. Winning AM, Merandi JM, Lewe D, et al. The emotional impact
138. Van Bogaert P, Dilles T, Wouters K, et al. Practice environment, of errors or adverse events on healthcare providers in the NICU: the
work characteristics and levels of burnout as predictors of nurse re- protective role of coworker support. J Adv Nurs. 2018;74:172-180.
ported job outcomes, quality of care and patient adverse events: a [PMID: 28746750] doi:10.1111/jan.13403
study across residential aged care services. Open Journal of Nurs- 155. Yanos PT, Vayshenker B, DeLuca JS, et al. Development and
ing. 2014;4:343-55. validation of a scale assessing mental health clinicians' experiences
139. Van Bogaert P, Kowalski C, Weeks SM, et al. The relationship of associative stigma. Psychiatr Serv. 2017;68:1053-1060. [PMID:
between nurse practice environment, nurse work characteristics, 28617207] doi:10.1176/appi.ps.201600553
burnout and job outcome and quality of nursing care: a cross- 156. Yassi A, Cohen M, Cvitkovich Y, et al. Factors associated with
sectional survey. Int J Nurs Stud. 2013;50:1667-77. [PMID: staff injuries in intermediate care facilities in British Columbia, Can-
23777786] doi:10.1016/j.ijnurstu.2013.05.010 ada. Nurs Res. 2004;53:87-98. [PMID: 15084993]
140. Van Bogaert P, Meulemans H, Clarke S, et al. Hospital nurse 157. You LM, Aiken LH, Sloane DM, et al. Hospital nursing, care
practice environment, burnout, job outcomes and quality of care: quality, and patient satisfaction: cross-sectional surveys of nurses and
test of a structural equation model. J Adv Nurs. 2009;65:2175-85. patients in hospitals in China and Europe. Int J Nurs Stud. 2013;50:
[PMID: 20568322] 154-61. [PMID: 22658468] doi:10.1016/j.ijnurstu.2012.05.003
141. Van Bogaert P, Timmermans O, Weeks SM, et al. Nursing unit 158. Yuguero O, Marsal JR, Buti M, et al. Descriptive study of asso-
teams matter: impact of unit-level nurse practice environment, nurse ciation between quality of care and empathy and burnout in primary
work characteristics, and burnout on nurse reported job outcomes, care. BMC Med Ethics. 2017;18:54. [PMID: 28950853] doi:10.1186
and quality of care, and patient adverse events—a cross-sectional sur- /s12910-017-0214-9

566 Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 Annals.org
Burnout and Quality of Care REVIEW
159. Zarei E, Khakzad N, Reniers G, et al. On the relationship be- 164. Dodd S, Clarke M, Becker L, et al. A taxonomy has been devel-
tween safety climate and occupational burnout in healthcare organi- oped for outcomes in medical research to help improve knowledge
zations. Safety Science. 2016;89:1-10. discovery. J Clin Epidemiol. 2018;96:84-92. [PMID: 29288712] doi:
160. Tawfik DS, Sexton JB, Kan P, et al. Burnout in the neonatal 10.1016/j.jclinepi.2017.12.020
intensive care unit and its relation to healthcare-associated infec- 165. Dal-Ré R, Ioannidis JP, Bracken MB, et al. Making prospective
tions. J Perinatol. 2017;37:315-320. [PMID: 27853320] doi:10.1038 registration of observational research a reality. Sci Transl Med. 2014;
/jp.2016.211 6:224cm1. [PMID: 24553383] doi:10.1126/scitranslmed.3007513
161. Van Gerven E, Vander Elst T, Vandenbroeck S, et al. Increased 166. Parshuram CS, Amaral AC, Ferguson ND, et al; Canadian Crit-
risk of burnout for physicians and nurses involved in a patient safety
ical Care Trials Group. Patient safety, resident well-being and conti-
incident. Med Care. 2016;54:937-43. [PMID: 27213542] doi:10
nuity of care with different resident duty schedules in the intensive
.1097/MLR.0000000000000582
care unit: a randomized trial. CMAJ. 2015;187:321-9. [PMID:
162. Sari AB, Sheldon TA, Cracknell A, et al. Sensitivity of routine
system for reporting patient safety incidents in an NHS hospital: ret- 25667258] doi:10.1503/cmaj.140752
rospective patient case note review. BMJ. 2007;334:79. [PMID: 167. West CP, Dyrbye L, Satele D, et al. A randomized controlled trial
17175566] evaluating the effect of Compass (Colleagues Meeting to Promote
163. Ioannidis JP, Patsopoulos NA, Rothstein HR. Reasons or ex- and Sustain Satisfaction) small group sessions on physician well-
cuses for avoiding meta-analysis in forest plots. BMJ. 2008;336: being, meaning, and job satisfaction. J Gen Intern Med. 2015;30:
1413-5. [PMID: 18566080] doi:10.1136/bmj.a117 S89.

INFORMATION FOR AUTHORS

The Annals Information for Authors section is available at www.annals.org

/aim/pages/authors. All manuscripts must be submitted electronically us-
ing the manuscript submission option at Annals.org.

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019 567
Current Author Addresses: Dr. Tawfik: 770 Welch Road, Suite Author Contributions: Conception and design: D.S. Tawfik,
435, Palo Alto, CA 94304. J.P.A. Ioannidis.
Dr. Scheid: Office BL341G, 221 Longwood Avenue, Boston, Analysis and interpretation of the data: D.S. Tawfik, J. Profit, T.
MA 02115. Shanafelt.
Dr. Profit: 1265 Welch Road, MSOB x1C07, Stanford, CA Drafting of the article: D.S. Tawfik, T. Shanafelt, J.P.A. Ioannidis.
94305. Critical revision for important intellectual content: D.S. Tawfik,
Dr. Shanafelt: 300 Pasteur Drive, Room H3215, Stanford, CA A. Scheid, T. Shanafelt, M. Trockel, J.B. Sexton, J.P.A. Ioannidis.
94305. Final approval of the article: D.S. Tawfik, A. Scheid, J. Profit, T.
Dr. Trockel: 401 Quarry Road, Room 2303, Stanford, CA Shanafelt, M. Trockel, K.C. Adair, J.B. Sexton, J.P.A. Ioannidis.
Provision of study materials or patients: D.S. Tawfik.
94305.
Statistical expertise: D.S. Tawfik.
Drs. Adair and Sexton: 3100 Tower Boulevard, Suite 300, Dur-
Obtaining of funding: D.S. Tawfik.
ham, NC 27707.
Administrative, technical, or logistic support: D.S. Tawfik, A.
Dr. Ioannidis: 1265 Welch Road, MSOB x306, Stanford, CA
Scheid, J.B. Sexton.
94305. Collection and assembly of data: D.S. Tawfik, A. Scheid, K.C.
Adair.

Annals.org Annals of Internal Medicine • Vol. 171 No. 8 • 15 October 2019