Original Manuscript

Nursing Ethics
1–27
Compassion fatigue in healthcare ª The Author(s) 2019
Article reuse guidelines:
sagepub.com/journals-permissions
providers: A systematic review 10.1177/0969733019889400
journals.sagepub.com/home/nej
and meta-analysis

Nicola Cavanagh
University of Calgary, Canada
Grayson Cockett and Christina Heinrich
Alberta Health Services, Canada

Lauren Doig
Carleton University, Canada
Kirsten Fiest, Juliet R Guichon, Stacey Page, Ian Mitchell and
Christopher James Doig
University of Calgary, Canada

Abstract
Background: Compassion fatigue is recognized as impacting the health and effectiveness of healthcare
providers, and consequently, patient care. Compassion fatigue is distinct from “burnout.” Reliable
measurement tools, such as the Professional Quality of Life scale, have been developed to measure the
prevalence, and predict risk of compassion fatigue. This study reviews the prevalence of compassion fatigue
among healthcare practitioners, and relationships to demographic variables.
Methods: A systematic review was conducted using key words in MEDLINE, PubMed, and Ovid databases.
Data were extracted from a total of 71 articles meeting inclusion criteria, from studies measuring
compassion fatigue in healthcare providers using a validated instrument. Quantitative and qualitative data
were extracted and compiled by three independent reviewers into an evidence table that included basic
study characteristics, study strength and quality determination, measurements of compassion fatigue, and
general findings. Meta-analysis, where data allowed, was stratified by Professional Quality of Life version,
heterogeneity was quantified, and pooled means were reported with 95% confidence interval. A table of
major study characteristics and results was created.
Ethical consideration: This paper contains no primary data obtained directly from research participants.
Data obtained from previously published resources have been acknowledged within references.
Psychological distress, particularly compassion fatigue, can be insidious, no health profession is immune,
and may significantly impact the ability to provide care.
Results: A total of 71 studies were included. Compassion fatigue was reported across all practitioner
groups studied. Relationships to most demographic variables such as years of experience and specialty were
either not statistically significant or unclear. Variability in reporting of Professional Quality of Life results was
found.

Corresponding author: Christopher J Doig, Department of Critical Care Medicine, Cumming School of Medicine, University of
Calgary, 3134 Hospital Drive NW, Calgary, AB T2N 5A1, Canada.
Email: [email protected]
2 Nursing Ethics XX(X)

Interpretation: Compassion fatigue exists across diverse practitioner groups. Prevalence is highly
variable, and its relationship with demographic, personal, and/or professional variables is inconsistent.
Questions are raised about how to mitigate compassion fatigue.

Keywords
burnout, compassion fatigue, healthcare practitioner, nursing, Professional Quality of Life, secondary
trauma

Introduction
Compassion fatigue is described as a healthcare practitioner’s diminished capacity to care as a consequence
of repeated exposure to the suffering of patients, and from the knowledge of their patient’s traumatic
experiences.1 Compassion fatigue is a result of providing patient care, and is more often considered as the
result of many events, though it could arise from the experience of caring for an individual patient or event.
Compassion fatigue is closely related to the concepts of “vicarious trauma” and “secondary traumatic stress
(STS),” both of which also result from exposure to the trauma experienced by patients, rather than to the
trauma itself.2 However, other factors contribute to the development of compassion fatigue, for example,
burnout (BO) may develop through factors such as work hours (i.e. shift work) and work environment,
impacting an individual’s capacity to care.3
Instruments have been developed to measure the prevalence of, and examine risk factors for, compas-
sion fatigue in care providers. One such instrument is the Professional Quality of Life (ProQOL) scale
developed by Stamm. The ProQOL began as the “Compassion Satisfaction and Fatigue Test (1993),” and
was developed to measure both the positive and negative elements experienced by persons who act as
professional helpers.4 The instrument uses three subscales: Compassion Satisfaction (CS), BO, and STS/
compassion fatigue (depending on the version). The subscales are scored separately, and scores are not
considered cumulatively; there is no accepted method of combining subscales to report an overall score.4
Instead, combinations of high and low scores in the subscales indicate an overall level of compassion
fatigue. Within each of the 3 domains, a score greater than 42 is considered as high, 23–41 as average, and
22 or less as low. CS indicates the positive feelings associated with doing a type of work, and a higher
score is suggestive of more contentment. Higher scores in STS and BO would indicate more negative
feelings and distress. The ProQOL does not have a scale specific control value for use as a diagnostic
instrument. The scale has evolved as more is learned about the factors contributing to compassion fatigue;
the current instrument represents the understood relationships between the positive and negative effects
of work on care providers.

Aim and objectives

This systematic review sought to describe and summarize the prevalence of compassion fatigue in healthcare
practitioners using narrative synthesis, and meta-analytic methods where data presented permitted. We asked
the following:

1. What is the frequency of reported compassion fatigue and in which care providers?
2. What variables impact compassion fatigue?
3. What are the common values reported in the subscales of ProQOL?
Cavanagh et al. 3

Methods
Search strategy
A broad electronic search was conducted using the MEDLINE (Ovid interface) and PubMed databases, using
MeSH term “compassion fatigue,” with additional keyword searches for “secondary traumatic stress,”
“secondary traumatization,” and “vicarious traumatization.” Search terms were combined using Boolean
operator “OR.” Articles were included from earliest available content up to 31 December 2017. The bib-
liographies of articles meeting inclusion criteria were also searched by reviewers for any additional articles.

Inclusion criteria
To be included in this review, studies were required to meet the following inclusion criteria: English language,
containing data that included a measurement of “compassion fatigue,” using a validated measurement tool and
reporting mean or median scores, using a participant sample consisting of practitioners from a professional
healthcare discipline providing frontline patient care, and originating from a peer reviewed scientific journal.
Studies that were specific to an examination of interventions or mitigation strategies were excluded.

Study selection and reporting

This systematic review and meta-analysis was conducted consistent with recommendations from “The Meta-
analysis of Observational Studies in Epidemiology” guidelines, and reported in accordance with the PRISMA
statement.5,6 No protocol was registered. A total of 1302 records were returned from the electronic search
strategy. An additional 46 records were found through reference searches. Two reviewers (N.L.C. and C.H.)
reviewed the abstracts of those articles and included them for full review if the abstract suggested they met
inclusion criteria. Of the 113 abstracts identified, the same reviewers independently and in parallel stringently
applied the inclusion and exclusion criteria to the full text of the articles to identify the 71 articles that remained
for this study. See the PRISMA diagram (Figure 1). Fifteen studies were excluded due to an issue with data
reporting (including incomplete data), 17 did not specifically report prevalence, 4 studies did not use a validated
instrument, 3 studies were not presented as journal articles, and 2 did not study healthcare practitioners.

Data extraction
After the final list of articles for inclusion was identified, data were extracted by two authors (N.L.C. and
G.C., or N.L.C. and C.D.) with all data independently verified by each author, and discrepancies resolved by
consensus. Quantitative data used in the meta-analysis (e.g. mean score, standard deviation, sample size)
were abstracted by two authors (C.D. and L.D.) and any discrepancies were resolved through discussion.
Full text review of included articles resulted in a list of study characteristics for extraction. These
characteristics were used to develop an evidence table that included the following: the study’s demographic
details (source, year, country, number of participants) and information pertaining to the overall strength and
design (population, total sample size, sampling method, research design, level of statistical analysis,
response rate, method of assessment, and the measurement tools utilized). Study result data were recorded
as quantitative results (mean scores, risk/incidence) and narrative summary.

Assessment of quality
Compassion fatigue is most often investigated using descriptive cross-sectional survey methodology, which
is appropriate for this subject matter. During full-text review, studies were assessed under guidelines from
4 Nursing Ethics XX(X)

Records idenfied through Addional records idenfied

Idenficaon

database searching through other sources

(n=1302) (n=46)

Records aer duplicates removed

(n=1224)
Screening

Records screened Records excluded

(n=1224) (From tle n=984
from abstract n=127)

Full-text arcles assessed Full-text arcles excluded

for eligibility (n=42)
(n=113)
Eligibility

Exclusions:
Studies included in -17 no reported prevalence
quantave synthesis -15 incomplete data reporng
(n=71) -4 no validated instrument
-3 not journal arcles
-2 did not study health care
providers
Included

Figure 1. PRISMA diagram of study selection.

the National Health Institute Quality Assessment Tool for Observational Cohort and Cross-Sectional
Studies.

Data synthesis and analysis

As described, the ProQOL is measured on three sub scales: CS, BO, and compassion fatigue/STS. The third
scale is sometimes measured as “compassion fatigue” and sometimes as “STS,” and each is scored differ-
ently. The versions of the ProQOL used in the reviewed studies varied (when reported) among versions III,
IV, and V, and both “compassion fatigue” and “STS” were reported as the third subscale.
Data were reported in individual studies one of three ways: (1) a detailed measure of risk based on
participant’s mean score on each of three subscales compared with established cut off scores, (2) a limited
measure of risk based on participant’s mean scores in two subscales (usually the CS and BO scales), (3) a
Cavanagh et al. 5

proportional analysis where the proportion of participants reporting values above (or below) established cut-
off scores are reported, or (4) a combination of the above. For all studies, we applied standard content analysis
of study results and discussion to create a narrative summary. In studies with detailed measure of risk, mean
subscale scores (and accompanying standard deviations) were pooled using random effects meta-analysis in
STATA 14. Meta-analyses were stratified by the version of the ProQOL to ensure appropriate pooling.
Heterogeneity was quantified using the I2 statistic and accompanying Q p-value. Pooled means are reported
with 95% confidence intervals (CIs). In cases where studies reported estimates from multiple groups (e.g.
nurses, physicians, social workers), results were pooled as long as the groupings were mutually exclusive. A
summary evidence table was created presenting the study results (title, author, year, country, sample pop-
ulation, measurement tools used, overall findings, quantitative prevalence, or risk measured by ProQOL).

Results
Study characteristics
Table 1 presents a summary of the study characteristics and major quantitative and overall findings,
including narrative summary. Of the 71 articles, 50 (70.4%) were exclusive to one professional care
provider group, whereas the remaining 21 (29.6%) included 2 or more care providers. Of the 50 studies
restricted to one profession, nursing professionals were most commonly represented (n ¼ 30), followed by
physicians or medical trainees (n ¼ 7), social workers (n ¼ 3), genetic counselors (n ¼ 3), and a mix of other
care providers (n ¼ 7). Nursing professionals included were as follows: (1) broadly representative of care
environments such as acute care hospitals—including emergency departments, critical care units, and acute
care wards—and varied settings outside of hospital including clinics and outpatient treatment settings; and
(2) broadly representative of clinic conditions and patient groups cared including examples such as pallia-
tive care, pediatrics, geriatrics, oncology and as transplant coordinators, and oncology. Of the 21 studies that
included multiple providers, these studies were not as expansive in care environments or patients served but
did include mental health providers/therapists, hospice care providers, and care in hospitals or acute care.
The studies had a significant focus on economically developed western healthcare systems with most
(41/71, 57.8%) being from North America and 37 of these exclusively the United States, Europe being the
next most frequent (10/71, 14.1%)—with 1 study comparing Brazil and Portugal, and 6 (8.5%) from
Australia–New Zealand. Only 4 of 71 studies (5.6%) were from Asian rim countries, and 2 of 71 (2.8%)
from the Indian sub-continent despite the large populations served by healthcare providers in these regions.
There was only one study from the African continent, and no study exclusively from South America (apart
from the aforementioned one study examining Portugal and Brazil).
Of the 71 articles, 67 (94.4%) used the ProQOL scale: 6 (9.0%) used version 3, 3 (4.5%) used version 4,
and 28 (41.8%) used version 5. Thirty of the 67 (44.8%) studies did not explicitly identify which ProQOL
version was used. All studies included some form of demographic questionnaire, and 56% (40/71) of studies
used one or more additional tools to further assess domains of psychological distress. Fifty six (78.9%) of
studies were published in the last 5 years, with 30 (42.3%) published since 2016, suggesting a recent
increased interest in this subject. The reporting of the results from the 71 studies varied. Of the 67 studies
that reported results from ProQOL, 41 studies reporting detailed measures of risk only from the 3 subscales
(61.2%), 12 (17.9%) studies reported only a proportional or percentage risk, 2 (3.0%) reported limited
results on only two of the three subscales, and the remaining 10 (14.9%) reported a combination of mean
results from the 3 subscales and proportional values above cutoffs.
The correlations between demographic and work variables and reported values on the ProQOL were
inconsistent between study results. A number of studies demonstrated that psychological distress and
underlying mental health conditions were associated with worse scores, but as cross-sectional surveys, it
6
Table 1. Summary of studies.
Author Sample Measurement
(reference) Year Country n instruments Overall findings Quantitative findings

Abendroth and 2006 United States Hospice care RNs ProQOL-III Overall sample at risk for compassion 26.4% high risk, 52.3% moderate
Flannery7 n¼216 fatigue; 91% scoring moderate high risk, and 21.3% low risk
risk for BO also scored moderate high Mean STS: 13.6
risk for compassion fatigue. No
correlation with personal
demographics, some correlation with
work-related variables.
35% of those diagnosed with depression
or PTSD scored high risk.
Adams et al.8 2006 United States Social workers Compassion Fatigue Scale Supported notion that job burnout and No scores provided
n¼274 General Health secondary trauma were separate
Questionnaire contributors to psychological distress.
Benoit et al.9 2007 United States Genetic Focus Group questions All participants described experience None reported
counselors indicating compassion fatigue. Themes
n¼12 included being overwhelmed due to
caring, feeling responsible, and not
being able to control patient suffering.
Difficulty differentiating experience of
compassion fatigue from BO.
Delivering bad news prevalent trigger
of compassion fatigue. Traumatic
memories also a trigger.
Emotional responses from work
interfered with ability to cope in
personal life.
Alkema et al.10 2008 United States Hospice ProQOL Self Care Burnout and compassion fatigue Mean CS: 40.5 (high), BO: 23.8
healthcare Assessment Worksheet negatively correlated to all aspects of (average), and compassion
practitioners self-care except physical care. CS fatigue: 17.5 (high)
n¼37 significantly positively correlated to SCAW showed higher scores for
emotional care, spiritual care, and self-care activities.
work life balance.
As compassion fatigue increased,
reported self-care decreased.
Yoder11 2008 United States RNs from special ProQOL-IV BO and compassion fatigue scales Mean overall scores: CS ¼ 40.3,
care units Narrative response strongly correlated. 15.8% of the BO ¼ 19.2, and compassion
n¼106 sample fell into area of risk for fatigue ¼ 12.3
compassion fatigue. CS found to be
higher in those working longer shifts in
ICU and with less experience.

(continued)
 Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings

Craig and Sprang12 2009 United States Trauma treatment ProQOL-III Sample reported significantly lower 32% above cut off for CS
therapists Trauma Practices compassion fatigue than in other 6% above cut off for compassion
n¼532 Questionnaire studies of mental health professionals. fatigue
Younger, less experienced reported 12% above cut off for BO
higher burnout; more experience had Mean CS: 43.8
higher CS. Mean BO: 7.4
Meadors et al.13 2009 United States Pediatric ProQOL (version not No statistically significant relationships 1.2% of the participants showed
healthcare specified) between professional groups and high BO scores, 76% scored
providers Secondary Traumatic Stress subscales. Higher compassion fatigue low. 7.3% scored high on
n¼167 Scale scores for those who had experienced compassion fatigue, 43%
Impact of Events Scale- a loss within the last 30 days. scored low.
Revised Mean CS: 38.2
Mean BO: 14.8
Mean STS: 9.1
Hooper et al.14 2010 United States Emergency ProQOL-IV ED nurses found to have lower levels of Risk percentages: 20.2% low risk
department CS compared to other services. 82% for CS, 26.6% high risk for BO,
RNs had moderate–high levels of BO, 86% 28.4% high risk for compassion
n¼114 had moderate–high levels of fatigue.
compassion fatigue.
Potter et al.15 2010 United States Oncology RNs ProQOL-IV Risk of compassion fatigue relatively Mean overall scores: CS ¼ 38.3,
n¼153 equal between work settings BO ¼ 21.5, and compassion
(inpatient/outpatient units). Higher fatigue ¼ 15.2
level of BO for those on outpatient
units. Staff with 11–20 years
experience showed highest risk scores
on all sub scales.
Circenis and 2011 Latvia Acute care ProQOL-V 53% of the sample had higher than the Mean scores: CS ¼ 37.42,
Millere16 hospital RNs Workplace questionnaire mean score for CS, 54% had higher BO ¼ 23.5, and STS ¼ 19.59
n¼129 Maslach Burnout Inventory than the mean score for BO, and 50%
had higher than the mean score for
STS.
Slocum-Gori 2011 Canada Palliative care ProQOL (version not Practitioners providing care for Mean scores: CS ¼ 43.9,
et al.17 providers specified) psychological, emotional, and physical compassion fatigue ¼ 18.6, and
n¼630 distress had significantly higher levels BO ¼ 20.8
of compassion fatigue and BO. Nurses
reported highest level of compassion
fatigue. Those who worked part time
had higher CS.

(continued)

7
8
Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings

Young et al.18 2011 United States Cardiovascular ProQOL-V Statistically significant differences Mean scores
RNs between nursing units for CS and BO HVICU: CS ¼ 36.6, BO ¼ 24.82,
n¼70 subscales. and STS ¼ 21.88
HVIMC: CS ¼ 41.84, BO ¼ 19.48,
and STS ¼ 19.44
Bhutani et al.19 2012 India Non-physician ProQOL-V CS highest for those with most years of Mean scores: CS ¼ 40.63,
clinicians Questionnaire including practice or in private practice. Those BO ¼ 22.8, and STS ¼ 23.52
n¼60 personal, professional, reported “poor working conditions”
anthropometric, and had higher BO. No significant
metabolic profiles correlation between sub scales and
demographic variables.
Rossi et al.3 2012 Italy Mental health ProQOL-III Those with reported psychological Mean scores: CS ¼ 32,
providers General Health distress had significantly lower CS. BO ¼ 21.15, and compassion
n¼260 Questionnaire (GHQ- Distress and prior trauma showed fatigue ¼ 10.2
12) higher BO and compassion fatigue.
Severn et al.20 2012 New Zealand Audiologists ProQOL-III Highest level of CS in private practice. 25% high level CS, 22% low
n¼82 Audiology Occupational Lower CS and higher BO with 20% high level risk of BO, 26% low
Stress Questionnaire increased age. Stress associated with 22% high risk of compassion
(AOSQ) patient contact strongest predictor of fatigue, 29% low
compassion fatigue.
El-Bar et al.21 2013 Israel Physicians CFST Compassion fatigue most prevalent. 35.2% at extremely high risk for
n¼128 Sharp divide in risk for compassion compassion fatigue, 9.4% high
fatigue, either extremely high or risk for BO, and 21.1% at risk
extremely low. for low CS.
Immigration can involve a great deal of
personal trauma, which increases risk
for compassion fatigue. Levels found
here significantly higher than other
studies.
Being born abroad/having no academic
affiliation increased compassion
fatigue risk.
No impact with other variables.
Kim22 2013 United States Liver and kidney ProQOL-V No significant correlation between Mean scores: CS ¼ 40, BO ¼ 24,
nurse demographic variables and subscale and STS ¼ 23
transplant scores.
coordinators Majority of sample had average levels of
n¼14 CS, BO, and STS.

(continued)
 Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings

Michalec et al.23 2013 United States Nursing students ProQOL-V Significantly lower levels of BO in first Overall sample categorization: CS
n¼436 Maslach Burnout Inventory year students than second year, but no levels ¼ high, BO levels ¼
(MBI) significant increase in third and fourth average, and STS levels ¼ low
Semi-structured interview year.
Mizuno et al.24 2013 Japan Nurses ProQOL ProQOL scores significantly associated Mean CS: 33.5
Midwives Frankfurt Emotional Work with stress factors and emotion work. Mean BO: 26.9
n¼255 Scale Particularly making value judgments, and Mean compassion fatigue: 21.3
having to control emotions. No
relationship seen between scores and
work experience.
Negative feelings about accepting certain
aspects of the work significantly
associated with compassion fatigue.
No significant differences in ProQOL
scores between nurses/midwives. No
high risk for compassion fatigue.
Sodeke-Gregson 2013 United Kingdom Trauma therapists ProQOL-V Majority in average range for CS and BO, CS: 8% low, 53.2% average
et al.25 n¼253 Coping Strategies Inventory and high range for STS. CS negatively BO: 64.2% average, 25.8% high
correlated with BO, BO positively STS: 70% high, 30% average
correlated with STS. Age, managerial
support positive predictors of CS.
Perceived management support and
older age significant negative
predictors of BO. More self-care, past
trauma for self-positive predictors for
STS.
Whitebird et al.26 2013 United States Hospice care ProQOL-III Higher levels of BO than compassion Mean scores: Compassion fatigue
providers of Short Form-12 fatigue, both less than average for ¼ 9.9,
various Health Survey Version 2 norms associated with ProQOL. Both BO ¼ 13.9.
designations (SF-12) were moderately correlated with
n¼547 Generalized Anxiety anxiety and depression.
Disorder Scale (GAD-7)
Patient Health
Questionnaire (PHQ-8)
Medical Outcomes Social
Support Survey (MOS6)
Job Satisfaction question
Coping Strategies

(continued)

9
10
Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings
27
Zeidner et al. 2013 United States Health ProQOL-III Women reported higher levels of No overall scores provided.
professionals of Schutte Self Report compassion fatigue. No significant Scores by practice/gender.
various Inventory difference between professional
designations Emotion-management groups.
and specialties subscale of the Mayer-
n¼182 Salovey-Caruso
emotional intelligence
test
Coping inventory for
stressful situations-
situation specific coping
Mood subscales of the
Dundee stress state
questionnaire
Ariapooran28 2014 Iran RNs ProQOL Inverse correlation between social 45.7% at risk for compassion
n¼164 Multidimensional Scale of support and BO; no meaningful fatigue
Perceived Social Support correlation between social support 54.3% of ER and 35.4% on non-ER
and compassion fatigue. nurses suffered from
May indicate that nurses working with compassion fatigue;
victims or survivors could be at risk 15.03% at risk for BO
for developing compassion fatigue or 19.2% of ER and 11.4% of non-ER
BO. Social support, a key variable in at risk for BO
determining risk for compassion
fatigue and BO.
Bellolio et al.29 2014 United States Resident ProQOL-V No statistically significant impact of any CS: 59% average level, 41% high
physicians demographic variable in any of the level
n¼188 subscales. BO scores higher when BO: 57% low, 43% average
more hours and night shifts worked. STS: 77% low, 23% average
Found average levels of CS, low levels of Mean CS: 40.4
BO and STS in sample; no difference in Mean BO: 21.8
comparison to other specialties. Mean STS: 19.8
Higher risk of developing compassion
fatigue with night shifts.

(continued)
 Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings
30
Hegney et al. 2014 Australia RNs ProQOL-V Approximately 20% in sample showed Mean CS: 35.66
n¼132 DASS potential risk and higher compassion Mean BO: 23.66
fatigue. 12% high risk, 7.6% very Mean STS: 18.6
distressed. Positively correlated with age and
STS and BO related and associated with years in nursing.
negative mood.
No significant correlation with variables.
Lower levels of CS in nurses with less
experience.
Hinderer et al.31 2014 United States RNs ProQOL-V 35.9% nurses scored for high BO, 27.3% Mean BO: 20.56 (suggested
n¼128 Penn Inventory for compassion fatigue, and more than higher risk of BO)
75% high for CS. Compassion fatigue Mean compassion fatigue: 13.94
less prevalent than BO. BO related to (reported by 27.3% of the
longer shifts, work relationships. sample)
Compassion fatigue triggered by Mean CS: 37.96 (majority above
challenging patients, futile care, average CS)
environment, and personal
experience. High prevalence of CS,
particularly in older nurses.
Martin et al.32 2014 United States Abortion ProQOL Significantly higher CS and lower BO than CS: 49% high, 30% average, and
providers Ways of Coping other healthcare providers. 21% low
(range of Questionnaire Compassion fatigue consistent with BO: 4% high, 21% average, and
clinicians) Workgroup other studies. 75% low
n¼79 Characteristics Measure compassion fatigue: 46% average,
People and Organizational 43% low
Culture Profile
APSS
Mason et al.33 2014 United States RNs ProQOL-V As work engagement increases, so does CS: 73.1% average, 26.9% high
n¼26 Utrecht Work Engagement CS. No other relationships between BO: 42.3% low, 57.7% average
Scale scores and variables. STS: 61.5% low, 38.5 average
Moral Distress Scale
Smart et al.34 2014 United States RNs ProQOL-V Mean scores did not differ among Mean CS: 39.1
Physicians departments. Mean BO: 23.4
Nursing assistants General medical workers showed Mean STS: 19.0
n¼139 evidence of higher burnout.
Significant increase in level of BO for
caregivers of non-critical patients
(compared to those in ED or ICU).
Findings contrary to other studies.
Night shifts associated with higher
BO, less CS.

11
(continued)
 Table 1. (continued)

12
Author Sample Measurement
(reference) Year Country n instruments Overall findings Quantitative findings

Amin et al.35 2015 India NICU RNs ProQOL-V Perceiving high stress experienced Perceived stress: most nurses
n¼129 Perceived Stress Scale 14 prolonged distress, might be higher perceived moderate stress
risk of BO, STS, or low CS. Weak to (47.3%), mild stress in 29.5%,
moderate positive correlation and high stress in 23.2%.
between perceived stress and BO/STS. High CS found in 19.4%, high BO
Almost 25% positive for BO and STS. in 23.3%, and high STS in
Similar to reports in western 23.3%.
literature.
Berger et al.36 2015 United States Pediatric RNs ProQOL-V STS due to frequent traumatic patient 71.5% moderate to high CS. Over
n¼239 Two open ended questions situations. End of life situations a quarter had low CS, high BO,
particularly difficult. Continually caring and high STS.
for critical patients’ frequent trigger of Mean CS: 49.1
compassion fatigue and BO. Mean BO: 50.2
Respondents reported decrease in Mean STS: 50.1
care and issues in personal life.
Younger nurses had lower CS, higher BO
and STS. Lower CS with 6–10 years
experience, higher after 20 years
experience. Lower CS, higher BO on
medical/surgical units.
Branch and 2015 United States Pediatric RNs, ProQOL-V Staff on units with long periods of stability 25% high risk for CS, 30.9% high
Klinkenberg37 social workers, and support showed higher CS, lower risk for BO, and 26.9% high risk
RTs, PTs, OTs, BO. Oncology units avoid compassion for STS.
psychologists, fatigue through meaningful Mean CS: 50.3
child life interactions with patients. Mean BO: 49.7
therapists, Significant score differences by clinical Mean STS: 50
PCAs unit.
n¼296
Dasan et al.38 2015 United Emergency ProQOL Levels of compassion fatigue found to be 98% of the consultants were
Kingdom medicine Potential impacts of low. Those with compassion fatigue average or high for CS, only
consultants in compassion fatigue more likely to report irritability and 2.3% low. Of the 2.3%, only
United questionnaire reduced standards of care, as well as two consultants had high BO,
Kingdom ED Qualitative interview intent to retire earlier. and one of those also high STS.
n¼681 Mean CS reduced marginally over first 10 Mean CS: 50.9
years, increased after 20 years. Job Mean BO: 50.8
demand, control, and support Mean STS: 50.5
impacted levels of satisfaction or
fatigue in interviews.

(continued)
 Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings
39
Hunsaker et al. 2015 United States Nurses ProQOL-V Overall low to average level of Mean CS: 39.77
n¼284 compassion fatigue and BO among ED compassion fatigue: 21.57
nurses. Compassion fatigue less BO: 23.66
prevalent with increasing age/ 56.8% average CS; 65.9% low
experience. compassion fatigue; 54.1%
Less compassion fatigue and BO with average BO.
increased managerial support. Older
nurses had higher levels of CS,
younger nurses had higher levels of
BO. More years nurse had practiced,
higher the level of CS, lower the level
of BO. Lower BO with shorter shifts.
Kelly et al.40 2015 United States RNs ProQOL-V Nurses would be in normal range for BO Overall mean ProQOL scores:
n¼491 and CS, low range for STS. BO ¼ 25.63, STS ¼ 20.86,
Some findings related to demographics: CS ¼ 40.51
“millennial” generation more likely to
be experiencing higher BO/STS and
lower CS than “Baby boomer” nurses.
As nurses gained experience,
increased compassion fatigue and
lowered CS.
Kim et al.41 2015 Korea RNs ProQOL-V Significant differences in CS and BO by Mean CS: 32.8
n¼488 Ethical Dilemmas age, marital status, religion, Mean BO: 29.0
Questionnaire educational status, position, and Mean STS: 27.3
Professional Nursing Values number of years in practice. Low CS
Questionnaire and high BO associated with nurses
with less than 3 years experience.
Level of CS higher than levels of BO
and STS. CS increases as age increases.
Lee et al.42 2015 United States Genetic ProQOL-V High CS found in counselors at high risk Mean CS: 41.15 (4% low range)
counselors STAI for compassion fatigue; both can exist Mean BO: 21.11 (19% high, 68%
n¼467 together. average)
Burnout a significant predictor of Mean STS: 19.37 (61% high, 39%
compassion fatigue. average)
Mangoulia et al.43 2015 Greece RNs ProQOL-V Overall showed low CS, high risk for BO Mean CS: 28.36 (low)
n¼174 and compassion fatigue. No relationship Mean BO: 25.17 (average)
shown between previous trauma and Mean compassion fatigue: 16.45
STS. Knowledge powerful protective (borderline high)
factor.

13
(continued)
14
Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings
44
Meyer et al. 2015 United States RNs starting ProQOL Compassion fatigue: 23.56 +
pediatric Mueller McCloskey 13.29
residency Satisfaction Scale CS: 93.56 + 16.77
program Life Events Checklist BO: 24.01 + 11.67
n¼231
Osland45 2015 Australia Dieticians ProQOL-V Significantly higher STS in those working CS: 78% average, 20% high
n¼87 with pediatrics. BO: 37% low, 63% average
Overall showed average CS, average BO, STS: 66% low, 34% average
and low STS; overall positive
professional quality of life. Higher STS
in high risk areas (ICU, mental health).
More years of practice showed higher
STS/BO.
Sacco et al.46 2015 United States RNs ProQOL-V Overall, nurses scored in average range CS ¼ 51.7
n¼221 for all three sub scales. STS ¼ 48.7
Suggested that age has impact on BO ¼ 48.2
ProQOL, nurses over 50 had higher
CS, lower BO and STS.
Younger less experienced nurses, higher
risk for compassion fatigue.
Significant difference in compassion
fatigue by age, unit acuity, and
management/practice change.
Harker et al.47 2016 Australia Psychology, social General Well Being Scale Higher level resilience significant Mean scores: BO ¼ 49.4,
work, (GWBS) predictor of lower distress (BO and STS ¼ 49.3
counseling ProQOL-V STS); age predictor of resilience and
n¼133 Resilience Factor Inventory mindfulness, not BO or STS. No other
(RFI) variable found to be predictors.
Freiburg Mindfulness
Inventory (FMI)
Kolthoff and 2016 United States Geriatric RN ProQOL-V Average levels of compassion fatigue and Mean scores: BO ¼ 26.7,
Hickman48 n¼42 BO. Higher BO and compassion STS ¼ 23, CS ¼ 37.4
fatigue in newer nurses.
Weintraub et al.49 2016 United States Neonatology Compassion Fatigue and The prevalence of compassion fatigue, Mean scores: BO ¼ 19.4,
physicians Satisfaction Self-Test BO, and CS 15.7%, 20.8%, and 21.9%, STS ¼ 17, CS ¼ 37.5
n¼443 (modified) respectively.

(continued)
 Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings

Wu et al.50 2016 Canada/United Oncology nurses ProQOL-V No statistical differences in CS, BO, and Mean scores Canada: BO ¼ 22.5,
States n¼549 Abendroth Demographic STS between United States and STS ¼ 22.4, CS ¼ 42.6
Questionnaire Canada. Younger nurses experienced Mean scores United States:
more STS. Cohesive team BO ¼ 22.7, STS ¼ 22.7,
environment buffered negative CS ¼ 42.4
feelings.
Beaumont et al.51 2016 United Midwife students ProQOL Above average BO in just over 50% of the Mean scores: BO ¼ 22.9,
Kingdom n¼103 Self-Compassion Scale sample. Participants with less self- STS ¼ 21.8, CS ¼ 41.6
Short Warwick and compassion reported higher
Edinburgh Mental Well- compassion fatigue and BO.
being Scale
Compassion for Others
Scale
Duarte and Pinto- 2016 Portugal Oncology nurses ProQOL Decrease in compassion fatigue and BO Mean scores, intervention:
Gouveia52 n¼48 with mindfulness intervention. BO ¼ 26.6, STS ¼ 25.7,
CS ¼ 40
Montross-Thomas 2016 United States Hospice staff and ProQOL Personal rituals, personal and No overall mean scores available.
et al.53 volunteers Questionnaire re: personal professional support, and age
n¼390 rituals and practices correlated with higher CS and lower
BO.
Yu et al.54 2016 China Oncology nurses ProQOL More years experience, passive coping, Mean scores: BO ¼ 21.1,
n¼650 Jefferson Scales of Empathy and working in secondary hospitals STS ¼ 21.4, CS ¼ 31.8
Simplified Coping Styles associated with higher compassion
Questionnaire fatigue and BO. “Perspective taking”
Perceived Social Support strong predictor of CS.
Scale
Chinese Big Five Personality
Inventory
Muliira and 2016 Uganda Midwives ProQOL Overall average levels of CS, STS, and Mean scores: BO ¼ 36.9,
Ssendikadiwa55 n¼224 BO. CS related to psychological well- STS ¼ 22.9, CS ¼ 19
being and job satisfaction.
Pruginin et al.56 2016 Israel Social workers ProQOL Average levels of CS, STS, and BO. No Mean scores, urban: BO ¼ 22.8,
n¼125 difference between study groups. STS ¼ 22.4, CS ¼ 38.6
Mean scores, regional: BO ¼ 23.1,
STS ¼ 22.8, CS ¼ 38.7
Cragun et al.57 2016 United States Military ProQOL-V Comparison between those who had Mean scores, deployed:
emergency been deployed and those who had not. BO ¼ 22.3, STS ¼ 20.7,
department No statistically significant differences CS ¼ 37.2
personnel between groups in any of the sub Mean scores, never deployed: BO
n¼105 scales. ¼ 21.8, STS ¼ 18.9, CS ¼ 39

(continued)

15
 Table 1. (continued)

16
Author Sample Measurement
(reference) Year Country n instruments Overall findings Quantitative findings

Craigie et al.58 2016 Australia Hospital nursing ProQOL Trait-negative affect important factor in Mean scores: BO ¼ 23.5,
staff Depression Anxiety Stress compassion fatigue and STS. CS STS ¼ 19.6, CS ¼ 37.6
n¼273 Scale protective factor to BO, not to STS.
State-Trait Anxiety
Inventory
Allsbrook et al.59 2016 United States Genetic ProQOL Overall, supervision not associated with Overall mean scores: BO ¼ 53,
counselors State-Trait Anxiety increase compassion fatigue. Less STS ¼ 59.6, CS ¼ 54.1
n¼320 Inventory experienced supervisors reported
Supervision questionnaire more STS, more likely to be at risk.
Tucker et al.60 2017 Canada Medical students ProQOL-V Decrease in compassion fatigue and Mean scores: BO ¼ 22.1,
n¼59 Self-Reflection exercises increase in burnout during medical STS ¼ 20.4, CS ¼ 40
student’s training; workshop on
compassion fatigue at start of program
had limited impact.
Wu et al.61 2017 United States Surgeons ProQOL-V CS significantly lower in trauma surgeon Overall mean scores: BO ¼ 25.3,
n¼178 sub-group; female surgeons across STS ¼ 22.6, CS ¼ 39.2
specialties reported higher
compassion fatigue and BO.
Mooney et al.62 2017 United States Nurses ProQOL-V CS lower and BO higher in ICU nurses Overall mean scores: BO ¼ 25,
n¼86 Demographic survey compared to oncology. CS higher and STS ¼ 20.5, CS ¼ 38.7
BO lower in male nurses.
Davidson et al.63 2017 United States Multidisciplinary ProQOL 48% of the participants reported a Overall mean scores: BO ¼ 21.7,
hospital staff stressful event at work. STS ¼ 21.8, CS ¼ 40.8
n¼164
Galiana et al.64 2017 Spain/Brazil Palliative care ProQOL Study to report scores as well as validate Mean scores Spain: BO ¼ 15.6,
professional s the instrument in Spanish. Study found STS ¼ 12.4, CS ¼ 41.1
n¼546 ProQOL to be valid. Mean scores Brazil: BO ¼ 15.1,
STS ¼ 14.2, CS ¼ 41.6
Zajac et al.65 2017 United States Oncology care ProQOL Average levels of CS and STS, low BO in Mean scores: BO ¼ 22.5,
providers Press Ganey survey sample. STS ¼ 24.2, CS ¼ 40.8
n¼107
Kelly and Lefton66 2017 United States ICU nurses ProQOL Sample reported average CS and STS, low Mean scores: BO ¼ 23.5,
n¼1136 BO. Significantly higher CS, lower BO STS ¼ 22, CS ¼ 38.8
in nurses who had received meaningful
recognition.
Cohen et al.67 2017 Israel Midwives ProQOL Relatively high levels CS reported, Mean scores: BO ¼ 13.2,
n¼93 PSS-SR seniority positively correlated with STS ¼ 12.6, CS ¼ 43.1
BO.
Cetrano et al.68 2017 Italy Mental health ProQOL Ergonomic problems and work impacting Mean scores: BO ¼ 22.9,
professionals Quality of Working Life personal life associated with higher STS ¼ 14.4, CS ¼ 32.7
n¼416 Questionnaire compassion fatigue and BO.

(continued)
 Table 1. (continued)

Author Sample Measurement

(reference) Year Country n instruments Overall findings Quantitative findings
69
Kim et al. 2017 South Korea Nurses ProQOL Type D personality significantly Mean scores: BO ¼ 28.7,
n¼875 Type D Personality Scale-14 associated with all three sub scales. STS ¼ 25.6, CS ¼ 32.1
Allen et al.70 2017 Australia Obstetricians/ ProQOL Intervention-based study. Scores in BO No mean scores available
Gynecologists and STS decreased with intervention,
n¼23 CS increased.
Austin et al.71 2017 United States Physicians/nurses Moral Distress Scale- Participants scoring as most at risk for Mean scores physicians:
n¼330 revised STS and BO also scored highest for BO ¼ 19.4, STS ¼ 20,
ProQOL moral distress. CS ¼ 40.7
Mean scores nurses: BO ¼ 20.6,
STS ¼ 20.2, CS ¼ 39.7
Trowbridge 2017 United States Pediatric hospital ProQOL Intervention-based study. Mindfulness Mean scores: BO ¼ 24.2,
et al.72 social workers Mindful Attention and intervention related to decrease in STS ¼ 21.5, CS ¼ 38.7
n¼33 Awareness Scale STS scores.
Caring Efficacy Scale
Klein et al.73 2017 United States Palliative care ProQOL Intervention-based study. Education- Mean scores: BO ¼ 27.3,
professionals based program resulted in increase in STS ¼ 26.1, CS ¼ 35.2
n¼12 CS, slight decrease in BO, and no
change in STS.
Pfaff et al.74 2017 Canada Palliative care ProQOL Evaluation of pilot program intervention Mean scores: BO ¼ 22.1,
professionals for compassion fatigue. Reduced STS ¼ 24.8, CS ¼ 42.2
n¼12 clinical stress seen at program
completion.
Hamid and Musa75 2017 UAE Psychologists and ProQOL Coping seen as factor in mediating No overall mean scores available
social workers General Health relationship between BO and STS. BO
n¼502 Questionnaire positively related to STS. CS negatively
Maslach Burnout Inventory associated with STS.
(HSS)
Endler and Parker’s Coping
Inventory
Duarte and Pinto- 2017 Portugal Oncology nurses ProQOL Psychological factors accounted for 26% Overall mean scores: BO ¼ 25.3,
Gouveia76 n¼221 of CS, 29% of BO, and 18% of STS ¼ 25.8, CS ¼ 38
compassion fatigue.

CS: Compassion Satisfaction; BO: Burnout; STS: Secondary Traumatic Stress; RN: registered nurse; PTSD: posttraumatic stress disorder; SCAW: self care assessment
worksheet; ICU: intensive care unit; ED: emergency department; ER: emergency room; NICU: neonatal intensive care unit; CFST: Compassion Satisfaction and Fatigue Test;
HVIMC: Heart and Vascular Intermediate Care Unit; HVICU: Heart and Vascular Intensive Care Unit; APSS: Accumulated Pain/Stressor Scale; DASS: Depression Anxiety Stress
Scale; PCAs: Patient Care Associates; STAI: State-Trait Anxiety Inventory; PSS-SR: PTSD Symptom Scale-Self Report; MBI-HSS: Maslach Burnout Inventory: Human Services
Survey.

17
18 Nursing Ethics XX(X)

was not possible to determine if one or the other was causal, or simply part of the same experience. CS was
variably reported with some studies suggesting it increased with experience15,20,39,41 or age,46,47 and other
studies suggesting it was inversely associated.12,19,30,38 BO and compassion fatigue were correlated8–
10,25,30
; higher scores in BO scales in particular were more commonly associated with the work environment

Table 2. Summary of studies using ProQOL tool, reporting all three sub scales.

Study Author (reference) Year Sample CS BO STS

10 Alkema et al. 2008 Other 40.5 23.8 17.5

11 Yoder 2010 Nurses 40.3 19.2 12.3
15 Potter et al. 2010 Nurses 38.3 21.5 15.2
17 Slocum-Gori et al. 2011 Other 43.9 20.8 18.6
19 Bhutani et al. 2012 Other 40.63 22.8 23.52
16 Circenis and Millere 2011 Nurses 37.42 23.5 19.59
3 Rossi et al. 2012 Other 32 21.15 10.2
18 Young et al. 2011 Nurses 36.6 24.82 21.88
22 Kim 2013 Nurses 40 24 23
24 Mizuno et al. 2013 Nurses 33.5 26.9 21.3
30 Hegney et al. 2014 Nurses 35.66 23.66 18.66
31 Hinderer et al. 2014 Nurses 37.96 20.56 13.94
34 Smart et al. 2014 Other 39.1 23.4 19
39 Hunsaker et al. 2015 Nurses 39.77 23.66 21.57
40 Kelly et al. 2015 Nurses 40.51 25.63 20.86
41 Kim et al. 2015 Nurses 32.8 29 27.3
42 Lee et al. 2015 Other 41.15 21.11 19.37
43 Mangoulia et al. 2015 Nurses 28.36 25.17 16.45
48 Kolthoff and Hickman 2016 Nurses 37.4 26.7 23
49 Weintraub et al. 2016 Other 37.5 18.4 17
50 Wu et al. 2016 Nurses 42.6 22.5 22.4
51 Beaumont et al. 2016 Other 41.6 22.9 21.8
52 Duarte and Pinto-Gouveia 2016 Nurses 40 26.6 25.7
54 Yu et al. 2016 Nurses 31.8 21.1 21.4
55 Muliira and Ssendikadiwa 2016 Other 19 36.9 22.9
56 Pruginin et al. 2016 Other 38.6 22.8 22.4
57 Cragun et al. 2016 Other 37.2 22.3 20.7
58 Craigie et al. 2016 Nurses 37.6 23.5 19.6
59 Allsbrook et al. 2016 Other 54.1 53 59.6
60 Tucker et al. 2017 Other 40 22.1 20.4
61 Wu et al. 2017 Other 39.2 25.3 22.6
62 Mooney et al. 2017 Nurses 38.7 25 20.5
63 Davidson et al. 2017 Other 40.8 21.7 21.8
64 Galiana et al. 2017 Other 41.1 15.6 12.4
65 Zajac et al. 2017 Other 40.8 22.5 24.2
66 Kelly and Lefton 2017 Nurses 38.8 23.5 22
67 Cohen et al. 2017 Other 43.1 13.2 12.6
68 Cetrano et al. 2017 Other 32.7 22.9 14.4
69 Kim et al. 2017 Nurses 32.1 28.7 25.6
71 Austin et al. 2017 Nurses and other 39.7 20.6 20.2
72 Trowbridge et al. 2017 Other 38.7 24.2 21.5
73 Klein et al. 2017 Other 35.2 27.3 26.1
74 Pfaff et al. 2017 Other 42.2 22.1 24.8
76 Duarte et al. 2017 Nurses 38 25.3 25.8

CS: Compassion Satisfaction; BO: Burnout; STS: Secondary Traumatic Stress.

Cavanagh et al. 19

such as location, shift type, or level of oversight responsibility.19,29,31,34 Factors associated with lower
scores, or improved CS, included positive work environments (including social support net-
works),25,28,50,53,66 work in private clinics,20 and support from management/leadership.35,37,39 Recent neg-
ative events were more likely overall to be associated with worse scores.31,36,63
Table 2 presents a summary of those studies that used the ProQOL tool as the primary measure of
compassion fatigue, and reported mean scores for all three sub scales (CS, BO, and STS/compassion
fatigue). The summary shows the scores between practitioner groups (nurses vs other health practitioners)
to be relatively consistent albeit, the scores for BO appeared to less variable across nurses. The results were
particularly interesting in that the scores for BO and STS/compassion fatigue were in the average to high
range, whereas the CS scores were not particularly “low”; this may raise the question as to the impact of
different domains on the overall effect on compassion fatigue.

Meta-analysis
In studies that used version 5 of the ProQOL (n ¼ 28), the pooled mean CS score was 41.8 (95% CI: 40.1–
43.5) (Figure 2), mean BO score was 28.4 (95% CI: 26.3–30.4) (Figure 3), and mean STS score was 25.8
(95% CI: 23.3–28.3) (Figure 4). This indicates average levels of BO and STS and average high levels of CS.
Heterogeneity was above 99.5% (p < 0.00001) on all version 5 subscale analyses. For the CS subscale, there
was an outlier value of 93.60; this study was conducted in a sample of 231 novice, pediatric RNs.
In studies that used version 3 of the ProQOL (n ¼ 6), the pooled mean BO score was 18.2 (95% CI: 13.6–
22.9) (Supplemental Figure 1), mean CS was 43.3 (95% CI: 36.2–50.4) (Supplemental Figure 2), and mean
STS was 16.6 (95% CI: 14.4–18.8) (Supplemental Figure 3). These scores indicate average high levels of
BO, average levels of CS, and low levels of STS. There was significant heterogeneity in all analyses (I2 >
89%, p < 0.00001).

Discussion
The results of our systematic review and meta-analysis demonstrate that across diverse healthcare practi-
tioner groups, most studies report results BO and compassion fatigue subscale results that fall into the
average risk groups within the ProQOL scoring matrix. The “BO” subscale had the highest mean value (and
therefore potentially most significant impact on compassion fatigue) across studies. This was also the
subscale that had the highest number of studies with heterogeneous results with 10 studies significantly
above the pooled mean. The synthesis of existing work as presented here demonstrates an association
between personal factors such as an existing diagnosis of anxiety or depression,7,8 and prior negative life
events3 and increased levels of compassion fatigue. Although the subscale scores for BO and compassion
fatigue were correlated, it was not as clear how factors related to the work environment such as shift time
and length,9,11,19,20 and the type of caring work being done,17 explicitly impacted compassion fatigue levels.
The pooled results of this meta-analysis are important for future research as reasonable summary estimates
for comparison. The high proportion of studies that were either exclusively in nursing professionals, or
included nurses in a diverse group of healthcare professionals reflects the important role of nurses in our
healthcare system, and their proportionally large cohort of healthcare providers. The similarity in results
across healthcare providers suggests that comparisons within and between health professionals are possible
and for non-nursing healthcare professionals, comparisons are still possible and relevant.
The finding that compassion fatigue exists across diverse practitioner groups can have a serious impact
on professional practice and workforce. Compassion fatigue, BO, post-traumatic stress disorder, and other
types of psychological/emotional distress are often associated and sometimes conflated. It is important to
differentiate these conceptually, as well as to recognize overlap as approaches to mitigating and treating the
20 Nursing Ethics XX(X)

Figure 2. ProQOL-V pooled mean scores by subscale: Compassion Satisfaction.

issues differ. Compassion fatigue is triggered by the continual use of empathy and emotional energy,
previous exposure to trauma, prolonged exposure to secondary trauma (a consequence of being witness
to the trauma of others, and being in a position of having to care for those who are suffering, rather than
being the primary subject of the trauma themselves), and the work environment.16 BO is defined as a
“psychological syndrome that involves a prolonged response to stressors in the workplace.” Specifically,
BO involves the chronic strain that results from an incongruence, or misfit, between the worker and the job.7
Cavanagh et al. 21

Figure 3. ProQOL-V pooled mean scores by subscale: Burnout.

BO is primarily triggered by work-related and organizational characteristics, with some predisposing

personal characteristics.16 Considering that BO develops in large part as a result of stressors that are
organizational and structural (i.e. related to working environment), changes in the work environment may
potentially mitigate negative impacts to patients, and be relatively easier to implement. However, BO may
also be related to personal characteristics or a disrupted care provider that is more difficult to address such as
inter-personal conflicts in or outside the work place, or personal financial stresses.
22 Nursing Ethics XX(X)

Figure 4. ProQOL-V pooled mean scores by subscale: Secondary Traumatic Stress.

Although BO and other distress may affect providers, compassion fatigue may more severely affects
patients as it is the direct effect of a healthcare provider’s diminished capacity to care that results from repeated
exposure to the suffering of their patients, as well as from the knowledge of their patient’s traumatic experi-
ences.1 Nimmo and Huggard,1 in a review of compassion fatigue in physicians, report that issues are “often
reflected in outcomes of emotional distress, pain, and suffering, and may manifest in increased rates of
absenteeism, reduced service quality, low levels of efficiency, and high attrition rates and eventually,
Cavanagh et al. 23

workforce dropout.” This evidence of compassion fatigue and the effect on providers and patient care raises an
important question on what strategies and programs health systems should consider to prevent or mitigate its
effect. As mentioned in a number of the studies included in this review, there are programs in existence, which
include general wellness programs that encourage self-care, and increased social and managerial support.
Education to practitioners, from students to the most experienced, about the existence and impact of compas-
sion fatigue could assist with identification of the condition, and also mitigate stigma in practitioners about the
prevalence and impact of psychological distress in health professions. Prior to the implementation of any
treatment or mitigation program, organizations should attempt to understand issues that affect their practi-
tioners, and implement programs tailored to meet specific mitigating causes and concerns. This may be one
specific benefit of a measurement tool such as the ProQOL as the effect of different domains can be measured,
and more focused interventions considered.

Strengths and limitations

Our study used robust methods and analysis. A limitation of cross-sectional studies in general is the inability
to infer causality from observed associations, and this includes in the synthesis of data in meta-analytic
techniques. There was significant heterogeneity in the pooled mean subscale scores on all versions of the
ProQOL; however, the 95% CIs were narrow indicating precision in the reported estimates.

Conclusion
Compassion fatigue exists across diverse practitioner groups and specialties and can be successfully mea-
sured using the ProQOL. Compassion fatigue’s relationship to demographic, personal, and professional
characteristics is unclear, as demonstrated by the variability in studies reviewed. Future research should be
directed to identifying specific triggers as root causes of compassion fatigue, impact of support programs for
providers, and developing education programs to mitigate the prevalence and severity of compassion
fatigue.

Conflict of interest
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or
publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs
Nicola Cavanagh https://orcid.org/0000-0001-5573-8764
Christopher James Doig https://orcid.org/0000-0002-8576-9139

Supplemental material
Supplemental material for this article is available online.

References
1. Nimmo A and Huggard P. A systematic review of compassion fatigue, vicarious trauma, and secondary traumatic
stress in physicians. AJDTS 2013; 1: 37–44.
2. Dominguez-Gomez E and Rutledge D. Prevalence of secondary traumatic stress among emergency nurses. J Emerg
Nurs 2008; 34: 199–204.
24 Nursing Ethics XX(X)

3. Rossi A, Cetrano G, Pertile R, et al. Burnout, compassion fatigue, and compassion satisfaction among staff in
community based mental health services. Psychiatry Res 2012; 200(2–3): 933–938.
4. Stam BH. The concise ProQOL manual, 2010, www.proqol.org
5. Stroup DF, Berlin JA, Morton SC, et al. Meta-analysis of observational studies in epidemiology: a proposal for
reporting. JAMA 2000; 283(15): 2008–2012.
6. Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols
(PRISMA-P) 2015 statement. Syst Rev 2015; 4: 1.
7. Abendroth M and Flannery J. Predicting the risk of compassion fatigue. J Hosp Palliat Nurs 2006; 8: 346–356.
8. Adams RE, Boscarino JA and Figley CR. Compassion fatigue and psychological distress among social workers: a
validation study. Am J Orthopsychiatry 2006; 76(1): 103–108.
9. Benoit LG, Veach PM and LeRoy BS. When you care enough to do your very best: genetic counselor experiences
of compassion fatigue. J Gen Couns 2007; 16: 299–312.
10. Alkema K, Linton JM and Davies R. A study of the relationship between self-care, compassion satisfaction,
compassion fatigue, and burnout among hospice professionals. J Soc Work End Life Palliat Care 2008; 4(2):
101–119.
11. Yoder EA. Compassion fatigue in nurses. Appl Nurs Res 2010; 23: 191–197.
12. Craig C and Sprang G. Compassion satisfaction, compassion fatigue, and burnout in a national sample of trauma
treatment therapists. Anxiety Stress Coping 2009; 23: 319–339.
13. Meadors P, Lamson A, Swanson M, et al. Secondary traumatization in pediatric healthcare providers: compassion
fatigue, burnout and secondary traumatic stress. Omega (Westport) 2009; 60(2): 103–128.
14. Hooper C, Craig J, Janvrin DR, et al. Compassion satisfaction, burnout and compassion fatigue among emergency
nurses compared with nurses in other selected inpatient specialties. J Emerg Nurs 2010; 36(5): 420–427.
15. Potter P, Deshields TL, Divanbeigi J, et al. Compassion fatigue and burnout: prevalence among oncology nurses.
Clin J Oncol Nurs 2010; 14: E56–E62.
16. Circenis K and Millere I. Compassion fatigue, burnout and contributory factors among nurses in Latvia. Procedia
Soc Behav Sci 2011; 159: 2042–2046.
17. Slocum-Gori S, Hemsworth D, Chan WW, et al. Understanding compassion satisfaction, compassion fatigue and
burnout: a survey of the hospice palliative care workforce. Palliat Med Care 2011; 27: 172–178.
18. Young JL, Derr DM, Cicchillo VJ, et al. Compassion satisfaction, burnout and secondary traumatic stress in heart
and vascular nurses. Crit Care Nurs Q 2011; 34(3): 227–234.
19. Bhutani J, Bhutani S, Balhara YP, et al. Compassion fatigue and burnout amongst clinicians: a medical exploratory
study. Indian J Psychol Med 2012; 34(4): 332–337.
20. Severn MS, Searchfield GD and Huggard P. Occupational stress amongst audiologists: compassion satisfaction,
compassion fatigue and burnout. Int J Audiol 2012; 51(1): 3–9.
21. El-Bar N, Levy A, Wald HS, et al. Compassion fatigue, burnout and compassion satisfaction among family
physicians in the Negev area-a cross sectional study. Isr J Health Policy Res 2013; 2(1): 31–38.
22. Kim S. Compassion fatigue in liver and kidney transplant nurse coordinators: a descriptive research study. Prog
Transplant 2013; 23(4): 329–335.
23. Michalec B, Diefenbeck C and Mahoney M. The calm before the storm? Burnout and compassion fatigue among
undergraduate nursing students. Nurse Educ Today 2013; 33(4): 314–320.
24. Mizuno M, Kinefuchi E, Kimura R, et al. Professional quality of life of Japanese nurses/midwives providing
abortion/childbirth care. Nurs Ethics 2013; 20(5): 539–550.
25. Sodeke-Gregson EA, Holttum S and Billings J. Compassion satisfaction, burnout and secondary traumatic stress in
UK therapists who work with adult trauma clients. Eur J Psychotraumatol 2013; 4, https://www-ncbi-nlm-nih-
gov.ezproxy.lib.ucalgary.ca/pmc/articles/PMC3877781/ (accessed 12 May 2017).
26. Whitebird RR, Asche SE, Thompson GL, et al. Stress, burnout, compassion fatigue and mental health in hospice
workers in Minnesota. J Palliat Med 2013; 16(12): 1534–1539.
Cavanagh et al. 25

27. Zeidner M, Hadar D, Matthews G, et al. Personal factors related to compassion fatigue in health professionals.
Anxiety Stress Coping 2013; 26(6): 595–609.
28. Ariapooran S. Compassion Fatigue and burnout in Iranian nurses: the role of perceived social support. Iran J Nurs
Midwifery Res 2014; 19(3): 279–284.
29. Bellolio MF, Cabrera D, Sadosty AT, et al. Compassion fatigue is similar in emergency medicine residents
compared to other medical and surgical specialties. West J Emerg Med 2014; 15(6): 629–635.
30. Hegney DG, Craigie M, Hemsworth D, et al. Compassion satisfaction, compassion fatigue, anxiety, depression and
stress in registered nurses in Australia: study 1 results. J Nurs Manag 2014; 22(4): 506–518.
31. Hinderer KA, VonRueden KT, Friedmann E, et al. Burnout, compassion fatigue, compassion satisfaction, and
secondary traumatic stress in trauma nurses. J Trauma Nurs 2014; 21(4): 160–169.
32. Martin LA, Debbink M, Hassinger J, et al. Abortion providers, stigma and professional quality of life. Contra-
ception 2014; 90(6): 581–587.
33. Mason VM, Leslie G, Clark K, et al. Compassion fatigue, moral distress, and work engagement in surgical intensive
care unit trauma nurses: a pilot study. Dimens Crit Care Nurs 2014; 33(4): 215–225.
34. Smart D, English A, James J, et al. Compassion fatigue and satisfaction: a cross-sectional survey among US
healthcare workers. Nurs Health Sci 2014; 16(1): 3–10.
35. Amin AA, Vankar JR, Nimbalkar SM, et al. Perceived stress and professional quality of life in neonatal intensive
care unit nurses in Gujarat, India. Indian J Pediatr 2015; 82(11): 1001–1005.
36. Berger J, Polivka B, Smoot EA, et al. Compassion fatigue in pediatric nurses. J Pediatr Nurs 2015; 30: e11–e17.
37. Branch C and Klinkenberg D. Compassion fatigue among pediatric healthcare providers. Am J Matern Child Nurs
2015; 40: 160–166.
38. Dasan S, Gohil P, Cornelius V, et al. Prevalence, causes and consequences of compassion satisfaction and compas-
sion fatigue in emergency care: a mixed-methods study of UK NHS consultants. Emerg Med J 2015; 32(8): 588–594.
39. Hunsaker S, Chen HC, Maughan D, et al. Factors that influence the development of compassion fatigue, burnout,
and compassion satisfaction in emergency department nurses. J Nurs Scholarsh 2015; 47(2): 186–194.
40. Kelly L, Runge J and Spencer C. Predictors of compassion fatigue and compassion satisfaction in acute care nurses.
J Nurs Scholarsh 2015; 47(6): 522–528.
41. Kim K, Han Y and Kim JS. Korean nurses’ ethical dilemmas, professional values and professional quality of life.
Nurs Ethics 2015; 22(4): 467–478.
42. Lee W, Veach PM, MacFarlane IM, et al. Who is at risk for compassion fatigue? An investigation of genetic
counselor demographics, anxiety, compassion satisfaction and burnout. J Genet Couns 2015; 24(2): 358–370.
43. Mangoulia P, Koukia E, Alevizopoulos G, et al. Prevalence of secondary traumatic stress among psychiatric nurses
in Greece. Arch Psychiatr Nurs 2015; 29(5): 333–338.
44. Meyer RML, Li A, Klaristenfeld J, et al. Pediatric novice nurses: examining compassion fatigue as a mediator between
stress exposure and compassion satisfaction, burnout and job satisfaction. J Pediatr Nurs 2015; 30(1): 174–183.
45. Osland EJ. An investigation into the professional quality of life of dieticians working in acute care caseloads: are we
doing enough to look after our own? J Hum Nutr Diet 2014; 28: 493–501.
46. Sacco TL, Ciurzynski SM, Harvey ME, et al. Compassion satisfaction and compassion fatigue among critical care
nurses. Crit Care Nurs 2015; 35: 32.
47. Harker R, Pidgeon AM, Klaassen F, et al. Exploring resilience and mindfulness as preventative factors for
psychological distress burnout and secondary traumatic stress among human service professionals. Work 2016;
54(3): 631–637.
48. Kolthoff JL and Hickman SE. Compassion fatigue among nurses working with older adults. Geriatr Nurs 2017;
38(2): 106–109.
49. Weintraub AS, Geithner EM, Stroustrup A, et al. Compassion fatigue, burnout and compassion satisfaction in
neonatologists in the US. J Perinatol 2016; 36(11): 1021–1026.
26 Nursing Ethics XX(X)

50. Wu S, Singh-Carlson S, Odell A, et al. Compassion fatigue, burnout, and compassion satisfaction among oncology
nurses in the United States and Canada. Oncol Nurs Forum 2016; 43(4): E161–E169.
51. Beaumont E, Durkin M, Hollins Martin CJ, et al. Compassion for others, self-compassion, quality of life and mental
well-being measures and their association with compassion fatigue and burnout in student midwives: a quantitative
survey. Midwifery 2016; 34: 239–244.
52. Duarte J and Pinto-Gouveia J. Effectiveness of a mindfulness-based intervention on oncology nurses’ burnout and
compassion fatigue symptoms: a non-randomized study. Int J Nurs Stud 2016; 64: 98–107.
53. Montross-Thomas LP, Scheiber C, Meier EA, et al. Personally meaningful rituals: a way to increase compassion
and decrease burnout among hospice staff and volunteers. J Palliat Med 2016; 19(10): 1043–1050.
54. Yu H, Jiang A and Shen J. Prevalence and predictors of compassion fatigue, burnout and compassion satisfaction
among oncology nurses: a cross-sectional survey. Int J Nurs Stud 2016; 57: 28–38.
55. Muliira RS and Ssendikadiwa VB. Professional quality of life and associated factors among Ugandan midwives
working in Mubende and Mityana rural districts. Matern Child Health J 2016; 20(3): 567–576.
56. Pruginin I, Segal-Engelchin D, Isralowitz R, et al. Shared war reality effects on the professional quality of life of
mental health professionals. Isr J Health Policy Res 2016; 5: 17.
57. Cragun JN, April MD and Thaxton RE. The impact of combat deployment on health care provider burnout in a
military emergency department: a cross-sectional professional quality of life scale V survey study. Mil Med 2016;
181(8): 730–734.
58. Craigie M, Osseiran-Moisson R, Hemsworth D, et al. The influence of trait-negative affect and compassion
satisfaction on compassion fatigue in Australian nurses. Psychol Trauma 2016; 8(1): 88–97.
59. Allsbrook K, Atzinger C, He H, et al. The relationship between the supervision role and compassion fatigue and
burnout in genetic counseling. J Genet Couns 2016; 25(6): 1286–1297.
60. Tucker T, Bouvette M, Daly S, et al. Finding the sweet spot: developing, implementing and evaluating a burn out
and compassion fatigue intervention for third year medical trainees. Eval Program Plann 2017; 65: 106–112.
61. Wu D, Gross B, Rittenhouse K, et al. A preliminary analysis of compassion fatigue in a surgeon population: are
female surgeons at heightened risk? Am Surg 2017; 83(11): 1302–1307.
62. Mooney C, Fetter K, Gross BW, et al. A preliminary analysis of compassion satisfaction and compassion fatigue
with considerations for nursing unit specialization and demographic factors. J Trauma Nurs 2017; 24(3): 158–163.
63. Davidson JE, Graham P, Montross-Thomas L, et al. Code lavender: cultivating intentional acts of kindness in
response to stressful work situations. Explore (NY) 2017; 13(3): 181–185.
64. Galiana L, Arena F, Oliver A, et al. Compassion satisfaction, compassion fatigue, and burnout in Spain and Brazil:
ProQOL validation and cross-cultural diagnosis. J Pain Symptom Manage 2017; 53(3): 598–604.
65. Zajac LM, Moran KJ and Groh CJ. Confronting compassion fatigue: assessment and intervention in inpatient
oncology. Clin J Oncol Nurs 2017; 21(4): 446–453.
66. Kelly LA and Lefton C. Effect of meaningful recognition on critical care nurses’ compassion fatigue. Am J Crit
Care 2017; 26(6): 438–444.
67. Cohen R, Leykin D, Golan-Hadari D, et al. Exposure to traumatic events at work, posttraumatic symptoms and
professional quality of life among midwives. Midwifery 2017; 50: 1–8.
68. Cetrano G, Tedeschi F, Rabbi L, et al. How are compassion fatigue, burnout, and compassion satisfaction affected by
quality of working life? Findings from a survey of mental health staff in Italy. BMC Health Serv Res 2017; 17(1): 755.
69. Kim YH, Kim SR, Kim YO, et al. Influence of type D personality on job stress and job satisfaction in clinical nurses: the
mediating effects of compassion fatigue, burnout, and compassion satisfaction. J Adv Nurs 2017; 73(4): 905–916.
70. Allen R, Watt F, Jansen B, et al. Minimizing compassion fatigue in obstetrics/gynecology doctors: exploring an
intervention for an occupational hazard. Australas Psychiatry 2017; 25(4): 403–406.
71. Austin CL, Saylor R and Finley PJ. Moral distress in physicians and nurses: impact on professional quality of life
and turnover. Psychol Trauma 2016; 9(4): 399–406.
Cavanagh et al. 27

72. Trowbridge K, Mische Lawson L, Andrews S, et al. Preliminary investigation of workplace-provided compressed
mindfulness-based stress reduction with pediatric medical social workers. Health Soc Work 2017; 42(4): 207–214.
73. Klein CJ, Riggenbach-Hays JJ, Sollenberger LM, et al. Quality of life and compassion satisfaction in clinicians: a
pilot intervention study for reducing compassion fatigue. Am J Hosp Palliat Med 2017; 35(6): 882–888.
74. Pfaff KA, Freeman-Gibb L, Patrick LJ, et al. Reducing the “cost of caring” in cancer care: evaluation of a pilot
inter-professional compassion fatigue resiliency programme. J Interprof Care 2017; 31(4): 512–519.
75. Hamid AA and Musa SA. The mediating effects of coping strategies on the relationship between secondary
traumatic stress and burnout in professional caregivers in the UAE. J Ment Health 2017; 26(1): 28–35.
76. Duarte J and Pinto-Gouveia J. The role of psychological factors in oncology nurses’ burnout and compassion
fatigue symptoms. Eur J Oncol Nurs 2017; 28: 114–121.