Journal of Rehabilitation Medicine 51-11 | Page 57

Systematic review of burnout in PM&R
871
Table I. Summary of included studies
Author, year,
country
Shanafelt et al. (2),
2012, USA
Study type Recruitment method Outcome measures Study population
Survey MBI PM&R specialists and other 97
physicians
Modified stratified random
sample
2-item PCEMD
PM&R physicians
Prevalence of
burnout (%)
48
1-item suicidal ideation
question a
Shanafelt et al. (1),
2015, USA
Modified stratified random
sample
62
1-item suicidal ideation
question a
Kao et al. (15),
2016, USA
Survey
Convenience sample
Likert work-life satisfaction
question b
MBI
PM&R residents
53 (distribution not
reported)
2014–2015
83.3 (PGY2)
37.5 (PGY3)
37.5 (PGY4)
2015–2016
58.3 (PGY2)
46.2 (PGY3)
Dyrbye et al. (11),
2018, USA
Survey
Modified stratified random
sample
MBI
PM&R residents and other
2-item career choice regret residents (PGY2)
questions c
30 (PGY2) 22.2 (PGY4)
50
1,536 50.7
PROMIS anxiety short form
8-item JSPE d
Sliwa et al. (12),
2019, USA
Survey
Likert work-life satisfaction
question b
MBI
PM&R specialists and other 170
physicians
2-item PCEMD
Survey
Purposive sample
TSES subscale of MOSSM
Mini-Z Burnout Survey
MBI single item d
American board-certified
PM&R specialists
Author-developed questions
about potential drivers of
burnout
a
”During the past 12 months, have you had thoughts of taking your own life?”. b My work schedule leaves me enough time for my personal/family life. c ”If you
could revisit your career choice, would you choose to become a physician again?” and “If you could revisit your specialty choice, would you choose the same
specialty again?”. d Maslach Burnout Inventory single item “I have become more callous towards people since I took this job.”
MBI: Maslach Burnout Inventory (22-item); PCEMD: Primary Care Evaluation of Mental Disorders; JSPE: Jefferson Scale of Physician Empathy; PROMIS: Patient-
Reported Outcomes Measurement Information System; TSES: Tangible Support and Emotional Support subscale of the Medical Outcomes Study Social Support
Measure; MOSSM: Medical Outcomes Study Social Support Measure.
summarized in Table I. Convenience sampling was used
in one study (15) of residents and fellows; purposive
sampling was used in one study (12) of PM&R specia-
lists; and modified stratified random sampling was used
in 2 studies (1, 2) of PM&R and other physician specia-
lists and one study (11) of residents in PM&R and other
specialties. The year of publication ranged from 2012
to 2019. All 5 studies were based in the USA. A total
of 1,886 PM&R trainees and specialists were surveyed
(83 trainees, 1,803 specialists). One study surveyed
American board-certified PM&R specialists, 2 studies
surveyed PM&R and other specialists, one study sur-
veyed PM&R residents stratified by postgraduate year,
and one surveyed PM&R and other residents. All studies
used the MBI or its components to measure burnout,
except one study (12) which used one item from the
Mini-Z Burnout Survey. In addition, 2 studies measured
depression and suicidal ideation using the 2-item Pri-
mary Care Evaluation of Mental Disorders, and a single
item for suicidal ideation and self-reported satisfaction
(1, 2). One study evaluated career choice regret using 2
items for which PM&R-specific data were available; this
study also evaluated anxiety using the Patient-Reported
Outcome Measurement Information System (PROMIS)
anxiety short form, empathy using 8 items from the Jef-
ferson Scale of Physician Empathy, and availability of
social support using the Tangible Support and Emotional
Support subscale of the Medical Outcomes Study Social
Support Measure, but PM&R-specific results were not
published (11). One study identified risk or protective
factors for burnout in PM&R specialists. No studies
reported on wellness or resilience.
Risk of bias
All of the studies meeting inclusion criteria were
surveys (level of evidence III); none of the studies
meeting inclusion criteria were controlled trials. None
of the included studies reported a priori clinical trial
registration. None of the included studies described
investigator blinding for MBI or other outcome mea-
sure assessment. Three studies (1, 2, 11) did not advise
participants that they were measuring burnout as the
purpose of assessment, consistent with best practices
for administering the MBI; one study (15) did not
disclose this practice; one study (12) asked partici-
J Rehabil Med 51, 2019