RESEARCH IN PEDIATRIC EDUCATION

Burnout in Pediatric Residents: Comparing Brief

Screening Questions to the Maslach Burnout
Inventory
Kathi J. Kemper, MD, MPH; Paria M. Wilson, MD, MEd; Alan Schwartz, PhD;
John D. Mahan, MD; Maneesh Batra, MD, MPH; Betty B. Staples, MD;
Hilary McClafferty, MD; Charles J. Schubert, MD; Janet R. Serwint, MD;
on behalf of the Pediatric Resident Burnout-Resilience Study Consortium
From the College of Medicine, The Ohio State University (KJ Kemper); Nationwide Children’s Hospital and the Department of Pediatrics,
College of Medicine, The Ohio State University College of Medicine (JD Mahan), Columbus, Ohio; Department of Pediatrics, University of
Cincinnati and Cincinnati Children’s Hospital Medical Center (CJ Schubert), Cincinnati, Ohio; Department of Pediatrics, Division of Emergency
Medicine (P Wilson), UPMC Children’s Hospital of Pittsburgh, Pittsburgh, Pa; Department of Medical Education and Pediatrics (A Schwartz),
University of Illinois College of Medicine, Chicago, Ill; Department of Pediatrics (M Batra), Seattle Children’s Hospital, Seattle, Wash;
Department of Pediatrics (B Staples), Duke University Medical Center, Durham, NC; Department of Medicine (H McClafferty), University of
Arizona College of Medicine, Tucson, Ariz; and Department of Pediatrics (JR Serwint), The Johns Hopkins University, Baltimore, Md
The authors have no conflicts of interest to disclose.
Address correspondence to Kathi J. Kemper, MD, MPH, College of Medicine,The Ohio State University, 370 W 9th Ave, Columbus, OH
43210 (e-mail: [email protected]).
Received for publication June 29, 2018; accepted November 2, 2018.

TAGEDPABSTRACT
BACKGROUND: Measuring burnout symptoms is important, item and a depersonalization item generated burnout esti-
but the Maslach Burnout Inventory (MBI) has 22 items. This mates of 53% in both years and, compared with the full MBI,
project compared 3 single-item measures with the MBI and had sensitivities of 85% to 87%, specificities of 84% to 85%,
other factors related to burnout. positive likelihood ratios of 5.7 to 6.4, and negative likeli-
METHODS: Data were analyzed from the 2016 and 2017 Pedi- hood ratios of 0.18 for both years. Both items were signifi-
atric Resident Burnout-Resilience Study Consortium surveys, cantly correlated with their parent subscales. The single items
which included standard measures of perceived stress, mind- were significantly correlated with stress, mindfulness, resil-
fulness, resilience, and self-compassion; the MBI; and the ience, and self-compassion.
1- and 2-item screening questions. CONCLUSIONS: The 1- and 2-item screens generated preva-
RESULTS: In 2016 and 2017, data were collected from 1785/ lence estimates similar to the MBI and were correlated with
2723 (65%) and 2148/3273 (66%) eligible pediatric residents, variables associated with burnout. The 1- and 2-item screens
respectively. Burnout rates on the MBI were 56% in 2016 may be useful for pediatric residency training programs track-
and 54% in 2017. The Physician Work Life Study item gen- ing burnout symptoms and response to interventions.
erated estimates of burnout prevalence of 43% to 49% and,
compared with the MBI for 2016 and 2017, had sensitivities TAGEDPKEYWORDS: burnout; education; pediatric; resident; resil-
of 69% to 72%, specificities of 79% to 82%, positive likeli- ience; screening
hood ratios of 3.4 to 3.8, and negative likelihood ratios of
0.35 to 0.38. The combination of an emotional exhaustion ACADEMIC PEDIATRICS 2019;19:251−255

TAGEDPWHAT’S NEW TAGEDPBURNOUT IS AN increasing focus of pediatric physician

training, highlighting the importance of accurately mea-
Little is known about the performance of brief screening
suring burnout.1 The most widely used instrument for
tools for burnout in pediatric residents. We found that 1-
assessing burnout is the 22-item Maslach Burnout Inven-
and 2-item screening questions have reasonable sensi-
tory (MBI), which has 3 subscales: emotional exhaustion
tivities, specificities, likelihood ratios, and construct
(EE), depersonalization (DP), and personal accomplish-
validity, and they may be useful for tracking burnout
ment (PA).2 Because burnout is often measured in lengthy
symptoms in pediatric residency training programs.
surveys that assess risk and protective factors associated

ACADEMIC PEDIATRICS Volume 19, Number 3

Copyright © 2018 by Academic Pediatric Association 251 April 2019
TAGEDEN252 KEMPER ET AL ACADEMIC PEDIATRICS

with it, a screening instrument with fewer items and lower they answered all 22 items on the MBI. For the MBI, burn-
participant burden is desirable. out as a dichotomous variable was defined as having high
In 2009, West and colleagues3 reported that 2 items from subscale scores for personal emotional exhaustion (≥27)
the MBI—“I feel burned out from my work” (an item from and/or depersonalization (≥10).9,10 Burnout as a continu-
the EE subscale) and “I have become more callous to other ous variable was defined as the total of the EE score plus
people since I took this job” (an item from the DP subscale) the DP score. The continuous variable was used not to
—were consistent with the full MBI in predicting major identify burnout but to assess correlations with other varia-
medical errors and suicidality in medical students and inter- bles. For this study, consistent with the PWLS study, we
nal medicine residents; however, they did not include a considered the single screening question positive if the
large sample of pediatric residents. In 2015, Dolan and col- score was ≥3 on the 5-point scale (1 or more symptoms of
leagues4 reported that a single non-proprietary item from burnout).4 Consistent with the West et al3 study, the 2
the Physician Work Life Study (PWLS)—“Overall, based screening items were considered positive for burnout if the
on your definition of burnout, how would you rate your responses on either the EE or DP item were endorsed at
level of burnout?” with a 5-point scale ranging from 1 (“I least weekly (>3 on the 7-point Likert scale from never to
enjoy my work. I have no symptoms of burnout.”) to 5 (“I daily). We determined the sensitivity, specificity, positive
feel completely burned out and often wonder if I can go and negative predictive values, and likelihood ratios of the
on.”), where scores of 3 or more are considered positive for PWLS screen compared with the full MBI in 2016 and
burnout—had a sensitivity of 83% and specificity of 87% 2017, as well as the EE and DP items alone and together
compared with the full MBI in a sample of practicing pri- compared with the EE and DP subscales and the full MBI
mary care physicians and nurses, which did not include for respondents in 2016 and 2017.
pediatric residents.4 Waddimba and colleagues5 tested all 3 The survey also included widely used measures of
single-item measures against the full MBI in a sample of stress (Perceived Stress Scale),11 mindfulness (Cognitive
308 physicians and advanced practice clinicians in rural and Affective Mindfulness Scale-Revised),12 resilience
upstate New York and reported that the 3 single-item meas- (Brief Resilience Scale),13 and self-compassion (Neff’s
ures varied in their effectiveness as screeners for EE or DP Self-Compassion Scale, short form).14 To answer the sec-
dimensions of burnout. Thus, although using just one or ond study question, consistent with the study by West et
two items is appealing, there is a gap in understanding the al,3 we multiplied the numeric response to the first screen-
test parameters of these 2 questions in pediatric residents ing question by 9 (because the emotional exhaustion sub-
who may differ from other groups of practicing physicians; scale of the MBI has 9 items), we multiplied the numeric
for example, pediatric residents typically include a higher response to the second screening question by 5 (because
percentage of women and are likely to be younger than the depersonalization subscale of the MBI has 5 items),
practicing clinicians. Being able to use single-item screens and then we added those results together.3 We then com-
may be useful at the program level and in larger epidemio- pared correlations of the full MBI versus the 2-item screen
logic studies looking at trends and responses to interven- to standard measures of stress, mindfulness, resilience,
tions. Furthermore, there are limited data on the correlation and self-compassion. Because there was substantial over-
between these screening questions and commonly measured lap in respondents from 2016 and 2017, we analyzed data
risk and protective factors for burnout (ie, stress, mindful- from each year separately.
ness, resilience, and self-compassion), and it would be use- Survey data were exported from LimeSurvey to R 3.3.1
ful to understand these relationships before relying on for analysis. Descriptive statistics were used to character-
single-item measures alone to track burnout symptoms. ize participants. Spearman’s rank correlations were used
The Pediatric Resident Burnout-Resilience Study Consor- to calculate correlations between variables. The Nation-
tium (PRB-RSC) has been conducting annual national sur- wide Children’s Hospital Institutional Review Board and
veys of burnout in pediatric residents since 2016, and we other participating programs approved this study.
were considering using 1 or 2 screening questions rather
than the full 22-item MBI in future studies. Before doing so,
however, we aimed to answer 2 research questions: 1) What TAGEDH1RESULTSTAGEDEN
are the sensitivity, specificity, positive and negative predic- Among the 1758/2723 (65%) eligible pediatric resident
tive values, and likelihood ratios of these screening items respondents in 2016 and 2148/3273 (66%) eligible
compared to the MBI in pediatric residents? 2) How do the respondents in 2017, 72% were women, 60% were mar-
correlations compare between the screening questions and ried, and 62% reported being more than $100,000 in debt.
full MBI to the variables of interest (stress, mindfulness, In 2016, 56% met the criteria for burnout on the full MBI;
resilience, and self-compassion) in pediatric residents? in 2017, 54% met the criteria. Table 1 shows the value of
1- and 2-item screens for predicting subscales and full-
scale scores of the MBI. Results from the 2016 survey are
TAGEDH1METHODSTAGEDEN presented on top, and a replication based on the 2017 sur-
The samples and survey instrument have been described vey is presented on the bottom. The PWLS single-item
fully elsewhere.6−8 Briefly, pediatric residents were sur- screen generated an estimated prevalence of burnout of
veyed in the spring of 2016 and 2017 from 34 and 46 pro- 46% in 2016 and 49% in 2017; it had a sensitivity of 69%
grams, respectively. Resident responses were eligible if and specificity of 82% (positive likelihood ratio [LR
 TAGEDENACADEMIC PEDIATRICS
Table 1. Sensitivity, Specificity, Likelihood Ratios, and Positive and Negative Predictive Values of Screening Questions in 2016 and 2017

Burnout Rate (95% CI) Sensitivity (95% CI) Specificity (95% CI) LR+ (95% CI) LR− (95% CI) PPV NPV
2016 56% on MBI
PWLS screening question compared with MBI 46% (44−49) 69% (66−72) 82% (79−84) 3.8 (3.2−4.4) 0.38 (0.35−0.42) 83% 67%
EE question compared with EE subscale of MBI 44% (43−46) 72% (69−74) 82% (79−84) 4.0 (3.5−4.5) 0.34 (0.31−0.38) 79% 75%
DP question compared with DP subscale of MBI 38% (36−40) 69% (66−72) 82% (80−84) 6.4 (5.3−7.8) 0.18 (0.16−0.21) 89% 71%
Combined EE and DP questions compared with full MBI 53% (50−55) 84% (82−86) 87% (84−89) 6.4 (5.3−7.8) 0.18 (0.16−0.21) 89% 81%
2017 54% on MBI
PWLS screening question compared with MBI 49% (47−51) 72% (69−75) 79% (76−81) 3.4 (3.0−3.9) 0.35 (0.32−0.39) 80% 71%
EE question compared with EE subscale of MBI 41% (39−43%) 78% (75−80) 90% (88−92) 7.7 (6.5−9.2) 0.25 (0.22−0.28) 87% 82%
DP question compared with DP subscale of MBI 35% (33−37) 77% (74−80) 88% (86−90) 6.6 (5.6−7.6) 0.26 (0.23−0.30) 78% 87%
Combined EE and DP questions compared with full MBI 53% (50−55) 85% (83−87) 85% (83−87) 5.7 (4.9−6.6) 0.18 (0.16−0.21) 87% 82%
CI indicates confidence interval; LR+, positive likelihood ratio; LR−, negative likelihood ratio; PPV, positive predictive value; NPV, negative predictive value; PWLS, Physician Work Life Study; MBI, Mas-
lach Burnout Inventory; EE, emotional exhaustion subscale of the MBI; DP, depersonalization subscale of the MBI.

Table 2. Spearman’s Rank Correlations of 2 Screening Items and Full-Scale MBI with Stress, Mindfulness, Resilience, and Self-Compassion among Pediatric Residents in 2016 and 2017

Full MBI Screen MBI (2 Questions) MBI EE Subscale MBI DP Subscale Stress Mindfulness Resilience Self-Compassion
2016 Residents
Full MBI 1 0.90 0.95 0.85 0.61 −0.44 −0.37 −0.46
PWLS item 0.64 0.68 0.68 0.43 0.59 −0.39 −0.36 −0.42

BURNOUT IN PEDIATRIC RESIDENTS

EE single item 0.81 0.92 0.86 0.52 0.61 −0.42 −0.37 −0.42
DP single item 0.73 0.79 0.84 0.84 0.36 −0.24 −0.19 −0.29
Combined EE+DP items 0.90 1 0.86 0.75 0.59 −0.40 −0.34 −0.42
2017 Residents
Full MBI 1 0.90 0.95 0.84 0.61 −0.46 −0.38 −0.47
PWLS item 0.65 0.68 0.69 0.44 0.61 −0.43 −0.40 −0.44
EE single item 0.82 0.92 0.87 0.54 0.61 −0.42 −0.38 −0.42
DP single item 0.73 0.79 0.85 0.85 0.36 −0.31 −0.24 −0.35
Combined EE+DP items 0.90 1 0.86 0.76 0.59 −0.43 −0.36 −0.43
MBI indicates Maslach Burnout Inventory; EE, emotional exhaustion subscale of the MBI; DP, depersonalization subscale of the MBI; PWLS, Physician Work Life Study.
All correlations are statistically significant at P < .01.

253
TAGEDEN254 KEMPER ET AL ACADEMIC PEDIATRICS

+] = 3.8, 95% confidence interval [CI] 3.2−4.4; LR students, residents, and practicing surgeons. Our data are
− = 0.38, 95% CI 0.32−0.42) compared with the full MBI also consistent with Waddimba et al,5 who reported that
in 2016 and similar values in 2017. The combined EE the 3 single items varied in effectiveness as screeners for
+DP screening questions generated a burnout prevalence the EE and DP dimensions of burnout; for example, the
of 53% in both 2016 and 2017. In 2016, for the combined DP item alone generated an estimated prevalence of burn-
EE+DP screen, the specificity was 87% and sensitivity out of 34% in 2016, substantially below the prevalence of
was 84% (LR+ = 6.4, 95% CI 5.3−7.8; LR− = 0.18, 95% 56% generated by the full MBI, and it showed lower cor-
CI 0.16−0.21), with similar values in 2017. Predictive relations with other variables, such as stress, mindfulness,
values in both years ranged between 81% and 89%. resilience, and self-compassion, than the PWLS item, the
Table 2 presents correlations among different measures of EE item alone, or the combination of EE and DP. In our
burnout, the MBI, and measures of related constructs. The study, the combined EE+DP item performed better than
top row presents results for the full MBI, which as the indus- either single item alone compared with the full MBI.
try standard presents a basis of comparison for the 1- and 2- The use of a single- or 2-item screen is appealing as a
item screens presented in subsequent rows. The single substitute for the full MBI, and, based on our findings rep-
PWLS item correlations with widely used measures of stress, licated across 2 years, these measures can be considered
mindfulness, resilience, and self-compassion were similar to to measure epidemiologic trends over time in pediatric
the full MBI, statistically significant, and in expected direc- residency programs. Pediatric program directors may opt
tions. Correlations between the combined EE and DP items to use the PWLS item or the 2-item (EE+DP) screening
and other variables were also similar to the full MBI. questions to gain insight into the magnitude of the prob-
lem in their programs and response to programmatic inter-
ventions. However, no screening test or even the industry
TAGEDH1DISCUSSIONTAGEDEN standard, MBI, can perfectly measure risk in individual
This study of large national samples of pediatric resi- residents, and if program directors are concerned about
dents over 2 years showed that 3 single-item screens, par- the potential for depression, substance abuse, or suicidal-
ticularly the combined EE+DP screening questions, ity, then additional assessments should be considered.18
generated similar prevalence estimates for burnout as the This study is limited by having been conducted among
full MBI. The single EE and DP items showed moderate pediatric residents who were part of training programs in
to strong correlations with the EE subscale and the DP the Pediatric Resident Burnout-Resilience Study Consor-
subscale, respectively. Furthermore, the single items also tium, and it is unclear whether there was a non-response
had similar, significant correlations with other variables bias. This study assessed burnout, not consequences of burn-
associated with burnout: stress, mindfulness, resilience, out such as suicidality, substance abuse, or medical errors.
and self-compassion. These findings extend previous The study included only pediatric residents, and it is not
research by focusing specifically on pediatric residents clear that results can be generalized to students, other spe-
and testing both PWLS items and the single items for EE cialties, or practicing pediatricians. It is also not clear that
and DP against the MBI and its subscales, as well as by these results can be generalized to other groups of health
assessing their correlations with other factors thought to professionals with markedly different rates of burnout.
be associated with burnout. Finally, these questions were included as part of a larger sur-
Although the MBI has been widely used among physi- vey on burnout in pediatric residents, and the items may
cians,15 its length limits its use for repeated assessment of function differently when used alone or in other contexts.
burnout and other risk and protective factors in large
groups of individuals. West et al3 demonstrated that the 2-
TAGEDH1CONCLUSIONSTAGEDEN
item measure of burnout could identify the likelihood of
high burnout with a Spearman’s rank correlation coeffi- The single PWLS item and the combined EE+DP
cient of 0.76 to 0.83 across the groups of individuals screening items have good correlations with variables
tested, which is similar to the correlations we found in thought to be associated with burnout, and they generate
both years between single items for each domain (EE and prevalence estimates similar to those of the full MBI.
DP). The single item from the PWLS has been reportedly They may be considered for tracking epidemiologic
correlated to emotional exhaustion in a large sample of trends in burnout and the response to programmatic inter-
residents of different specialties with r = 0.64,16 which ventions in pediatric residencies.
echoes our findings in pediatric residents.
Dolan et al4 reported on the test parameters of the TAGEDH1ACKNOWLEDGMENTSTAGEDEN
PWLS screening item relative to the full MBI, but the The authors thank all of the residents who completed the surveys. The
reported prevalence of burnout in their study was almost authors also thank the staff at the Association of Pediatric Program
half of what was found in our study, which can alter the Directors Longitudinal Educational Assessment Research Network office
negative and positive predictive values of this screen. and, in particular, Ms. Beth King, whose diligence and tireless efforts
Similar to our findings of high correlations with measures were instrumental in establishment of the consortium.
The 46 participating residency programs listed below are all members
of stress, mindfulness, resilience, and self-compassion, of the Pediatric Resident Burnout-Resilience Study Consortium and partic-
West et al17 found that the EE+DP items exhibited a ipated in one or both years of the study. Every institution had site principal
strong association with outcomes of suicidality in medical investigators who served as collaborators for this study and manuscript.
TAGEDENACADEMIC PEDIATRICS BURNOUT IN PEDIATRIC RESIDENTS 255

Pediatric Resident Burnout-Resilience Study Consortium in 2018: 2. Maslach C, Jackson S, Leiter M. Maslach Burnout Inventory Man-
Megan McCabe, Albert Einstein College of Medicine (Montefiore); ual. 3rd ed. Palo Alto, Calif: Consulting Psychologists Press; 1996.
Molly Shane; Sharon Unti, Ann & Robert H. Lurie Children’s Hosp of 3. West CP, Dyrbye LN, Sloan JA, et al. Single item measures of emo-
Chicago; Sanghamitra Misra, Baylor College of Medicine (Houston); tional exhaustion and depersonalization are useful for assessing
Colin Sox, Boston Children’s Hospital/Boston Medical Center; Violet burnout in medical professionals. J Gen Intern Med. 2009;24:1318–
Borowitcz, Carillion Clinic - Virginia Tech Carillion School of Medi- 1321.
cine; Sydney Primis, Carolinas Medical Center; Ross Myers, Kathy 4. Dolan ED, Mohr D, Lempa M, et al. Using a single item to measure
Mason, Case Western Reserve University/University Hospital Case burnout in primary care staff: a psychometric evaluation. J Gen
Medical Center/Rainbow Babies and Children’s Hospital; Miriam Stew- Intern Med. 2015;30:582–587.
art, Lisa Zaoutis, Oana Tomescu, Children’s Hospital of Philadelphia; 5. Waddimba AC, Scribani M, Nieves MA, et al. Validation of single-
Chuck Schubert, Cincinnati Children’s Hospital Medical Center/Univer- item screening measures for provider burnout in a rural health care
sity of Cincinnati College of Medicine; Frances Zappalla, Crozer-Ches- network. Eval Health Prof. 2016;39:215–225.
ter Medical Center; Kim Gifford, Dartmouth Hitchcock Medical Center; 6. Batra M, Kemper KJ, Serwint JR, et al. Burnout in pediatric resi-
Betty Staples, Duke University Hospital Medical Center; Jerry Rushton, dents: a national survey to inform future interventions (platform pre-
Adam Hill, Kim Schneider, Indiana Univeristy School of Medicine; sentation). Acad Pediatr. 2017;17:e39–e40.
Kathleen Donnelly, Inova Fairfax Medical Campus/Inova Children’s 7. Batra M, Kemper KJ, Mahan JD, et al. Promoting resilience among
Hospital; Jordan Watson, Jefferson Medical College/duPont Hospital for pediatric trainees and practitioners Paper presented at: In: Pediatric
Children; Janet R. Serwint, Johns Hopkins University School of Medi- Academic Societies Meeting; May 6, 2018; Toronto.
cine; Jerussa Levy, Louisiana State University (New Orleans); Pam 8. Kemper KJ, McClafferty H, Wilson PM, et al. Do mindfulness and
Dietz, Maine Medical Center; Jay Homme, Mayo Clinic College of Med- self-compassion predict burnout in pediatric residents? Acad Med.
icine (Rochester); Leah Phillipi, Medical College of Wisconsin; John 2018; in press.
Mahan, Nationwide Children’s Hospital/Ohio State University; Jennifer 9. Dyrbye LN, Massie Jr FS, Eacker A, et al. Relationship between
DiPace, New York Presbyterian Hospital (Cornell Campus); Bridget burnout and professional conduct and attitudes among US medical
VoigtRush University Medical Center; Nicole Paradise Black, Denise students. JAMA. 2010;304:1173–1180.
McIntyre, Shands Hospital at the University of Florida; Katherine Baker, 10. Shanafelt TD, Boone S, Tan L, et al. Burnout and satisfaction with
St. Christopher’s Hospital for Children; Priya Garg, Tufts Medical Cen- work-life balance among US physicians relative to the general US
ter; Lauren Nassetta, University of Alabama Medical Center; Hilary population. Arch Intern Med. 2012;172:1377–1385.
McClafferty, University of Arizona Health Sciences Center; Albina 11. Cohen S. Perceived stress in a probability sample of the United
Gogo, University of California (Davis) Health System; Kate Perkins, Jes- States. In: Spacapan S, Oskamp S, eds. The Claremont Symposium
sica Lloyd, Alan Chin, University of California (Los Angeles) Medical on Applied Social Psychology: The Social Psychology of Health,
Center; Deborah Rana, University of California (San Diego); Joseph Thousand Oaks, Calif: Sage Publications; 1988:31–67.
Hageman, University of Chicago Corner Children’s Hospital; Jennifer 12. Feldman G, Hayes A, Kumar S, et al. Mindfulness and emotion reg-
Reese, University of Colorado Denver Children’s Hospital; Amanda ulation: the development and initial validation of the Cognitive and
Osta, University of Illinois College of Medicine at Chicago; Lisa Gilmer, Affective Mindfulness Scale-Revised (CAMS-R). J Psychopathol
University of Kansas School of Medicine; Tara McKinley, Michelle Ste- Behav Assess. 2007;29:177–190.
venson, University of Louisville School of Medicine; Hilary Haftel, 13. Smith BW, Dalen J, Wiggins K, et al. The brief resilience scale: assess-
Thomas Saba, University of Michigan Medical Center; Maren Olson, ing the ability to bounce back. Int J Behav Med. 2008;15:194–200.
Michael Pitt, University of Minnesota Medical School; Carmen Herrera; 14. Raes F, Pommier E, Neff KD, et al. Construction and factorial vali-
Lanier Lopez, University of New Mexico Children’s Hospital; Keith dation of a short form of the self-compassion scale. Clin Psychol
Mather, University of Oklahoma College of Medicine (Tulsa); Casey Psychother. 2011;18:250–255.
Hester, University of Oklahoma Health Sciences Center (Oklahoma 15. Rafferty JP, Lemkau JP, Purdy RR, et al. Validity of the Maslach
City); Stephanie Dewar, University of Pittsburgh Children’s Hospital; Burnout Inventory for family practice physicians. J Clin Psychol.
Sophia Goslings, University of South Alabama Children’s and Women’s 1986;42:488–492.
Hospital; Maneesh Batra, University of Washington-Seattle Children’s 16. Rohland BM, Kruse GR, Rohrer JE. Validation of a single-item
Hospital; Grant Syverson, University of Wisconsin Hospitals and Clin- measure of burnout against the Maslach Burnout Inventory among
ics; Ann Burke, Wright State University. physicians. Stress Health. 2004;20:75–79.
17. West CP, Dyrbye LN, Satele DV, et al. Concurrent validity of
TAGEDH1REFERENCESTAGEDEN single-item measures of emotional exhaustion and depersonaliza-
tion in burnout assessment. J Gen Intern Med. 2012;27:1445–
1. Shanafelt T, Hasan O, Dyrbye L, et al. Changes in burnout and satis- 1452.
faction with work-life balance in physicians and the general US 18. Omurtag A, Fenton AA. Assessing diagnostic tests: how to correct
working population between 2011 and 2014. Mayo Clin Proc. for the combined effects of interpretation and reference standard.
2015;90:1600–1613. PLoS ONE. 2012;7:e52221.