Journal of Rehabilitation Medicine 51-11 | Page 56

870
E. A. Bateman and R. Viana
with the general population, other factors that increase
a physician’s risk of developing burnout vary by study,
with few studies yielding reproducible results (1, 2,
9–12). Among the heterogeneous findings, individual,
system, institutional, and patient factors have been
shown to play a role (4, 6, 10, 12).
Medical specialty has been shown to play a role in
predisposing specialists and trainees to burnout, but the
specialties found to be at highest risk varies by study
(1–3, 10–11). Little is known about burnout in PM&R.
In their role caring for persons with disabilities and
after catastrophic accidents, physiatrists are adept at
fostering resilience amongst their patients, which may
confer resilience as physicians, although the high de-
gree of empathic demand may, conversely, increase the
risk of emotional exhaustion. This systematic review
aims to consolidate available literature to enhance our
understanding of this critical problem in a potentially
vulnerable physician population.
Data extracted included number of participants, recruitment met-
hod, participant characteristics (career stage, sex, age), country
of study, year of publication, study methodology, tool or criteria
used to assess burnout, incidence and severity of burnout and, if
available, prevalence of depression, suicide, measures of well-
ness including resiliency and satisfaction, and risk or protective
factors. Data were compiled using a standardized collection form.
Data analysis
Data were analysed using descriptive statistics. Content analysis
was performed for qualitative data. Risk of bias was assessed
using the Cochrane Risk of Bias Tool (14).
RESULTS
METHODS
To evaluate the prevalence of burnout in PM&R specialists
and trainees. This systematic review was registered with the
International Prospective Register of Systematic Reviews
(PROSPERO, registration #CRD42018104133).
Inclusion criteria
www.medicaljournals.se/jrm
All studies meeting the inclusion criteria were cross-
sectional surveys. Their characteristics and findings are
Articles Identified
MEDLINE (n=284)
CINAHL (n=22)
EMBASE (n=43)
Hand Search (n=10)
Total # Articles Identified
(n=359)
Duplicates
(n=128)
Systematic literature searches were conducted in
MEDLINE, CINAHL and EMBASE for peer-
reviewed full text articles and published abstracts in
English published before 1 July 2018, using a com-
bination of keywords and Medical Subject Headings
(MeSH). The search strategy is shown in Appendix I.
Hand searches of studies for which the full text was re-
viewed were used to identify other relevant citations.
During manuscript preparation, the authors became
aware of additional potentially relevant publications
from September 2018 and January 2019, after the
initial systematic review search strategy was com-
pleted. Updated systematic searches were performed
in February 2019 using the same search strategy to
ensure the inclusion of relevant studies. The authors
became aware of a retraction and replacement (13)
applicable to one of the included studies during the
course of manuscript preparation; the reasons for
retraction were reviewed and data extracted was
revised to reflect the replacement. Citations obtained
from the searches were screened for eligibility by 2
independent reviewers (EAB and RV). Disagreements
Study features
# Articles after Duplicates
Removed
(n=231)
Potentially Relevant Studies
(n=33)
Search strategy
Studies reporting the prevalence of burnout in specialists and/
or residents in PM&R were included. Studies that grouped
PM&R specialists or residents with healthcare providers from
other medical disciplines were excluded, unless PM&R-specific
outcomes were reported separately.
The search strategy yielded 359 articles, of which 128
were duplicates. A total of 33 articles underwent full text
review, and 5 met the inclusion criteria (Fig. 1). Four
articles (1, 2, 11, 12) were peer-reviewed publications,
and one (15) was a peer-reviewed abstract. Cohen’s
kappa was 0.89 between the 2 reviewers, indicating
excellent agreement in identifying the relevant studies.
# Articles Undergoing Full
Text Review
(n=33)
Data extraction
Study selection
Objective
between the reviewers were resolved by discussion to reach
consensus. Agreement between reviewers was determined
using Cohen’s kappa.
Total # Studies Included
(n=3)
Removed by Screening Abstracts
Non-Physician Health Professional (n=69)
Patient or Non-Physician Caregiver (n=49)
Non-PM&R Physician (n=36)
Non-Bumout or Wellness Outcome (n=24)
Comment or Opinion (n=8)
Not English (n=12)
Removed by Full Text Review
Non-Physician Health Professional (n=12)
Patient or Non-Physician Caregiver (n=3)
Non-PM&R Physician (n=5)
Non-Bumout or Wellness Outcome (n=3)
Comment or Opinion (n=85)
Fig. 1. PRISMA study selection flow chart. Flow chart of strategy used for study
selection.