To validate the Compassion Fatigue Inventory in Spanish-speaking health personnel, its psychometric properties were tested, and the influencing factors were investigated.
Method
This was a cross-sectional validation study. Information was collected through a survey of 733 nurses, physician and medical residents using the Compassion fatigue Inventory (CFI), Secondary Traumatic Stress Questionnaire (CETS), The physician burnout syndrome scale (PhBS), the quality-of-life index (SQL -sp) and the reduced Moral Stress Scale (MMDHPr). The psychometric properties of the CFI were tested via exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and internal consistency analysis. The final version of the translated CFI was correlated with other scales.
Results
An EFA was performed to evaluate the translated CFI, resulting in 15 items in three dimensions, as reported in the literature: reduced compassion, social life, and workplace. The CFA showed good fit indices and psychometric values (Cronbach´s alpha = 0.87, Omega = 0.87, Comparative Fit Index = 0.99, Tucker Lewis = 0.99, root mean square error of approximation = 0.045, Standardized Root Mean Square Residual = 0.05). The CFI had a negative correlation with the SQL-sp and a positive correlation with the other subscales. Work experience is predictive of a small reduction in CFI scores.
Conclusion
The adaptation of the CFI in a sample of Latino health professionals shows satisfactory psychometric indices in the evaluation of compassion fatigue and can be proposed as a specific inventory to differentiate compassion fatigue from other occupational syndromes.
Hinweise
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Background
The occupational health of health workers is aimed at promoting and maintaining the highest degree of physical, mental, and social well-being [1]. This is influenced by the physical environment and safety climate, organizational policies, workers’ health status, work overload, work rhythms, work schedules, and degree of control over work. The evaluation and experience of workers, their subjective response (well-being), and access to work resources to obtain a sufficient quality of life contribute to their occupational health [2, 3]. Compassion is a cognitive, affective, and behavioral process that includes recognizing and understanding another's suffering, producing an emotional resonance of personal approach, and the motivation to act to alleviate it [4]. Work in the health field requires emotional labor facilitated by the expression of compassion, which facilitates work commitment and meaningfulness [5, 6]. Although compassion has been reported to be protective of the mental health of healthcare personnel [7], its expression in situations with high stress levels can lead to compassion fatigue [8]. Compassion at work acts as a psychological resource that, when depleted, produces difficulty in expressing empathy, generating an imbalance between effort and reward, aggravating the loss of resources on the part of health workers [9], which facilitates the appearance of other work syndromes and thoughts of abandonment of the profession [10].
The concept of compassion fatigue, initially coined by Joinson [11], was developed by Figley [12] to refer to the secondary traumatization suffered by health personnel who are in contact with patients in extreme suffering. Since its formulation, compassion fatigue has become one of the most frequently analyzed and evaluated syndromes in professions that require empathic interaction [13, 14]. Stamm [15] considered the loss of compassion satisfaction in helping patients, burnout, and secondary traumatic stress to be the central elements in compassion fatigue. This definition has been heavily criticized since it lacks a clear conceptual definition of compassion and a lack of measure of its loss [16, 17].
Anzeige
The theoretical model of compassion fatigue proposed by Coetzee and Laschinger [9] presents compassion as a psychological resource that requires feedback from the patient, the health care system, and coping by the professional himself to avoid exhaustion. According to the model, compassion fatigue occurs as a result of an unresolved stress-laden empathic process due to a lack of resources and positive feedback [8, 18]. This model is based on the theory of COR [19], allowing an understanding of the development of compassion fatigue as a process of work stress that results in a reduction in health professionals’ ability to express compassion in their relationships with patients and others. On the basis of this and the study of the experience of compassion fatigue in therapists by Norrman Harling and Högman [20], the working group of Eng and Nordström [21] developed a new instrument (the compassion fatigue inventory) that formulates compassion fatigue as the decreased expression of compassion by health professionals in their relationships with patients and others. Unlike other questionnaires, the Compassion Fatigue Inventory directly assesses compassion fatigue and allows its distinction from other occupational syndromes, such as secondary traumatic stress, professional fatigue, and conscience stress.
Secondary traumatic stress
Secondary traumatic stress (STS) is the process by which a health professional who cares for traumatized people also experiences traumatic responses [22]. The symptoms of STS are generally similar to those of direct trauma and consist mainly of re-experiencing, behavioral avoidance, and hyperarousal [23]. Other symptoms include anxiety, depression, sleep disturbances, maladaptive coping strategies, and negative emotions such as anger and feelings of worthlessness [24]. Moreno Jiménez, Morante Benadero [25] developed a processual model for STS in which organizational antecedents, level of satisfaction, personality, and consequences are considered. As an evaluation instrument, they proposed the Secondary Traumatic Stress Questionnaire (CETS), which was validated in Spanish-speaking firefighters and paramedics [26, 27].
Burnout
Maslach and Jackson [28] described burnout as a syndrome characterized by low personal achievement, emotional exhaustion, and depersonalization. Healthcare is facilitated by the organizational environment, which is modulated by factors such as the need to adapt to technological advances [29], the growing healthcare pressure that demands more administrative work [30], long work hours, a lack of sleep [31], social expectations about medical work, and civil responsibilities [32]. Recent trends in research on burnout treat it as a syndrome that expresses discomfort at work caused by a situation, a system, and a task that has become complex and difficult to resolve, especially in professions related to helping professions, such as healthcare [33]. Professional fatigue (burnout) is facilitated by the development of compassion fatigue [34, 35] and by the presence of moral stress [36, 37]. Its evaluation is performed via different instruments [38], the most important and prevalent of which is the Maslach Burnout Inventory [28]. In Spanish-speaking doctors, the Physician Burnout Syndrome Scale (PhBS) was developed as a specific medical tool validated in Hispanic and Mexican populations [39].
Moral stress
The term was coined by Jameton (1984) to capture the stress that health professionals, such as nurses, suffer from being unable to act in accordance with their professional values due to institutional obstacles. Moral stress is a relatively new construct, especially in workplace stress research [40].
Anzeige
The current organizational situation, the coronavirus disease 2019 (COVID-19) pandemic, and the increased sensitivity of professionals have increased in popularity in recent years. Among the structural causes are administrative demands derived from patient care, increased productivity, families' insistence on aggressive treatments, caring for too many patients, not having sufficient resources, continuing with aggressive treatments in terminal patients, and the obligation to exercise unnecessary medical orders [41]. Moral stress is strongly associated with the intention to leave work, which increases costs and decreases care within hospital systems [42, 43].
For its evaluation, the Moral Distress Scale [44] is used following the proposals of Hamric, Borchers [45], adapted by Epstein, Whitehead [46] with information from 52 countries, and the Measure of Moral Distress in Healthcare Professionals (MMDHP) is an adequate instrument. A Spanish validation of the instrument, which provides a unidimensional rating of moral stress, was used in this study [47, 48].
Aims and hypotheses
Compassion fatigue is a syndrome highly related to moral stress and burnout [10, 49], and it is also associated with the development of traumatic stress symptoms [50]. It is important to translate, apply, and validate an instrument such as the CFI that directly evaluates compassion fatigue and distinguishes it from similar constructs. In Latin America, there is a lack of validated instruments for the specific evaluation of compassion fatigue that discriminates it from other work syndromes. As a secondary objective, we sought to describe the work settings related to compassion fatigue and its associated factors in the study population.
Methods
Study design
This is a cross-sectional correlational instrumental study. A digital platform (Survey Monkey) was used to disseminate the CFI scale among Mexican health professionals working in public or private hospitals/clinics. The information was collected from February to April 2024 and reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting checklist [51].
Translation process and pilot study
Once the author's permission was obtained, two authors who are native spanish speakers fluent in english (AKG, RMML) translated the CFI questionnaire proposed by Eng [21] from english to spanish, and it was corrected by one author (BMJ). It was independently back-translated to english by two health psychology experts who are native english speakers and fluent in spanish. The original and back-translated versions were similar [69]. An initial pilot test was carried out on 60 subjects, which revealed that the CFI scale has good internal reliability (α = 0.91, Ω = 0.90).
Participants
From a total of 1070 people, a response rate of 68.5% was obtained who answered all the items. All the questions in each online questionnaire were mandatory, so no items were missed per instrument. Therefore, no substitutions were made. Among these 733 health professionals, 32.3% (237) were men, and 67.7% (496) were women. All of them work either in a private or public hospital or clinic in México.
Sampling procedure
The questionnaire and the informed consent letter were converted to electronic formats through the SurveyMonkey platform, where participants authorized the use of data and gave their electronic consent to respond to the electronic survey. Participants were recruited by convenience. A minimum of 15 subjects per question were included in the statistical analyses [52]. The application to participate and a link to the online survey were sent to several private groups of health professionals. It was distributed in groups of medical teachers to be distributed to medical residents. It was distributed to nursing groups locally and nationally.
Measures
To determine the sociodemographic characteristics and explore the objectives of the study, resources from the list proposed by Hobfoll [53] were included along with the demographic ones: sex, gender, age, job position, marital status, presence of marital problems, paternity/maternity, number of children, level of education, years of work experience, type of contract, work shift, usual place of work, presence of a second job, religion, and economic income.
Anzeige
Compassion fatigue inventory
The CFI proposed by Eng [26] has 16 items that measure the dimension of reduced compassion (item 1. My Will to help has declined), social life (item 10, I have noticed that my patience in my personal relationships has dwindled) and workplace (Item 14, I feel that my workplace provides care that is in accordance with my values) with Likert-type response options of 1 = “Does not fit at all,” 2 = “Fits poorly,” 3 = “Fits partially,” 4 = “Fits fairly well” and 5 = “Fits perfectly”. The method of grading requires that items 14, 15 and 16 be reversed in their score to later give a rating by dimension or total.
In this study, the final reliability values of the CFI for the sample of 733 subjects were an alpha of 0.88 and an omega of 0.87.
Secondary traumatic stress questionnaire
The CETS is a questionnaire with 14 items in Spanish that are graded on a Likert scale from 1 = totally disagree to 4 = totally agree. It measures three dimensions of secondary traumatic stress: emotional fatigue (4 items), secondary trauma (6 items), and belief shaking (4 items). It was developed by Moreno Jiménez, Morante Benadero [25] and validated in emergency medical professionals and firefighters in Mexico [27]. Written authorization was received from one of the authors for its use. In this study, an alpha value of 0.84 and an omega of 0.83 were obtained.
Physician burnout syndrome scale
The PhBS is a 12-item questionnaire in Spanish that is graded on a Likert scale from 1 = totally disagree to 4 = totally agree. It measures three dimensions of professional medical fatigue: exhaustion (4 items), disengagement (4 items) and loss of expectations (4 items). It was proposed and validated in Spanish-speaking medical professionals by Moreno-Jiménez, Barbaranelli [39]. We received written authorization from one of the authors for its use. In this study, an alpha and omega value of 0.89 was obtained.
The quality of life index
The SQL-sp is a questionnaire with 10 items in Spanish that evaluates the perception of physical well-being, psychological well-being, self-care and independent functioning, occupational functioning, interpersonal functioning, social/emotional support, community and service support, personal fulfillment, spiritual fulfillment and global perception of quality of life on a Likert scale from 1 = bad to 10 = excellent. This has been proposed and validated in Spanish by Mezzich, Ruipérez [54] and in Mexico by Meda-Lara, Yeo-Ayala [55]. Written permission was received from one of the authors for its use. In this study, an alpha and omega of 0.95 were obtained.
Measure of moral distress for health care professionals
The MMDHP is a questionnaire developed by Epstein, Whitehead [46] that consists of 27 items that have already been translated into Spanish and that provide a unidimensional response that correlates with the intention to leave the job [47, 48]. Owing to the length of the questionnaire and the battery applied, 10 items were selected on the basis of the theoretical model and the greater factor loading of the original scale. This assumes the theoretical and psychometric assumptions proposed by Epstein (2019) of the original questionnaire. Each item captures the intensity of perceived stress (0–4) and its frequency (0‒4). The moral stress score is calculated for each item by multiplying the dimension of the perceived stress level with the frequency of appearance and then adding all the items to obtain a global score ranging from 0–160. In this study, an alpha and omega of 0.89 were obtained.
Statistical analysis
Statistical analyses were performed in SPSS Statistics 29 and Rplus Studio vs 2024.04 with the Lavaan package. A confirmatory factor analysis was performed via Rplus Studio on the CETS, PhBSS, SQL-Sp, and MMDHP scales. In the respective analyses, the comparative fit index (CFI) and Tucker‒Lewis (TLI) indices were calculated considering levels greater than 0.9 as adequate. The standardized root mean mean residual (SRMR) and the root mean square error of approximation (RMSEA) were calculated, considering values less than 0.1 and 0.08, respectively, as appropriate [56]. To validate the CFI scale, two random groups were generated from the total sample. After verifying that there were no significant sociodemographic differences between the two groups, a polychoric correlation matrix was done for all the items in the CFI before EFA and CFA in each group. An exploratory factor analysis using the weighted least squares method was carried out in the first group (n = 365), and the number of factors to be used was obtained using a parallel analysis. A confirmatory factor analysis using the diagonally weighted least squares method was conducted in the second group (n = 368). A Pearson correlation was carried out between the CFI, EETS, EDPM, MMHDP and QLI-sp scales and their subdimensions.
Verifying the normality of the data distribution, the Student's t test and ANOVA were used for those who presented a normal distribution and Mann–Whitney and Kruskal Wallis for those who presented a non-normal distribution to analyze the CFI scores based on the data, demographics and work environments. Linear regression was performed via a bootstrap method with 5000 resamples and fixed random seeds.
Patient and public involvement
During the translation of the CFI test and the pilot study, medical residents and psychologists helped assess the appropriateness of the item’s phrasing and provided feedback. Health workers were given lectures on compassion fatigue prevention to promote the distribution of the CFI.
Results
Sample characteristics
The population consisted of 733 subjects, with a mean age of 35.79 ± 9 years (range 19–70 years) and a mean work experience of 9.89 ± 8.6 years. Medical residents were the profession with the highest response to the survey. Most of our sample worked in hospitalized settings (67%), followed by outpatient clinics (32%) on the matutine shift. With respect to marital status, singles prevailed (53.1%), followed by married individuals (30.6%); 87.7% reported having no problems in their romantic relationships. With respect to paternity/maternity, 332 (45.3%) reported having children or their partner had them. The average number of children who lived with health professionals was 2. With respect to the employment variables, 58% (n = 379) had a permanent contract, 217 were under a scholarship (33%), and 57 were under a temporary contract (8.7%). %); 80 participants reported practicing only in private practice. A total of 168 (22.9%) reported working at two sites. A total of 75.6% declared that they practice some religion. Detailed data can be consulted in Table 1.
Table 1
Demographic characteristics of the population
Dimension
Total = 733
Women = 496
Men = 237
Age. (y)
35.79 ± 9 y
37 ± 9 y
34 ± 9 y
Profession (work experience)
Nurse
256 35% (14 y)
218 (15 y)
38 (12 y)
Medical resident
257 35% (2 y)
137 (2 y)
120 (2 y)
Physician
184 25.1% (12 y)
116 (12 y)
68 (12 y)
Head of medical/nurse department
36 4.9%(19 y)
25 (20 y)
11 (19 y)
Workplace
Hospitalized
288 (39%)
200
88
Emergency department
92 (13%)
66
26
Critical care unit
33 (4%)
23
10
Operating room
79 (11%)
41
38
Outpatient clinic
150 (20%)
107
43
Other
91 (12%)
59
32
Work Shift
Matutine
472 (64%)
309
163
Vespertine
113 (15%)
74
39
Nocturnal
113 (15%)
91
22
Weekend
35 (5%)
22
13
Level of education
Bachelor´s degree
379 (52%)
265
114
Master/specialty
257 (35%)
173
84
PhD/subspeciality
97 (13%)
58
39
Monetary income (Us dollars)
2,492–12,463
479 (65.3%
329
150
12,464 – 24,926
163 (22.2%)
127
36
More than 24,927
91 (12.4%)
40
51
Most of our sample included nursing and medical residents who worked in hospitalization during the morning shift with a bachelor's degree and with a low income
Anzeige
Factor analysis of the compassion fatigue inventory
The sample was randomly divided into two groups: one with 365 participants and a second with 368 participants. Exploratory factor analysis (EFA) was carried out for the first group, and confirmatory factor analysis (CFA) was carried out for the second. The two subgroups used for the EFA and the AFC were equivalent in terms of age and sex.
The data from the first group showed an alpha and omega of 0.89. Floor and ceiling effects were assessed on the CFI; the percentages of respondents scoring at the minimum were 0.3% and at the maximum 0.3%. The descriptive statistics of the CFI were calculated, as shown in Table 2.
Table 2
Univariate descriptive statistics of CFI scores
Scale Mean if Item is deleted
Scale variance if item deleted
Corrected item-total correlation
Cronbach´s Alfa if item deleted
McDonald´s Omega if item deleted
CFIitem1
39.81
128.47
0.630
0.885
0.884
CFIitem2
40.19
128.54
0.671
0.884
0.883
CFIitem3
40.52
132.35
0.638
0.886
0.886
CFIitem4
39.78
126.05
0.697
0.883
0.881
CFIitem5
39.58
127.03
0.671
0.884
0.882
CFIitem6
40.19
130.54
0.629
0.886
0.885
CFIitem7
39.91
127.14
0.673
0.884
0.882
CFIitem8
40.17
129.11
0.659
0.885
0.883
CFIitem9
39.16
131.52
0.476
0.891
0.890
CFIItem10
39.36
126.3
0.666
0.884
0.882
CFIitem11
39.53
127.48
0.628
0.885
0.885
CFIitem12
39.42
125.77
0.651
0.884
0.883
CFIitem13
39.72
127.31
0.646
0.885
0.883
CFIitem14
39.56
140.96
0.192
0.900
0.902
CFIitem15
38.45
140.95
0.172
0.902
0.902
CFIitem16
39.48
139.06
0.245
0.899
0.901
How all the items contribute to the variance of the test is shown
The polychoric correlation matrix shows values between items between 0.03 and 0.79, being the lowest ones for items 14, 15, and 16. A Kaiser‒Meyer‒Olkin value of 0.92 and a Bartlett test value < 0.001 indicated an appropriate sample for performing an exploratory factor analysis. The weighted least squares method was used with an oblique rotation (oblimin). The parallel analysis results using 1000 iterations indicated that three factors should be retained. Factor 1 (reduced compassion) explained 30% of the variance, Factor 2 (social life) explained 23% of the variance, and Factor 3 (workplace) explained 8% of the variance, with a RMSR 0.03, which coincides with the proposed dimensional structure in the original instrument. Item 9 showed a factor loading outside of its original factor (Table 3).
Table 3
Factor loadings from the CFI exploratory factor analysis
Factor
1 Compassion reduction
2 Social life
3 Workplace
h2
CFI 1 My will to help has declined
(Mi voluntad de ayudar ha disminuido)
0.61
0.53
CFI 2 I have started to judge my patients in a way I would not want to
(He empezado a juzgar a mis pacientes de una manera que no me gusta)
0.89
0.76
CFI 3 I have started to feel a growing reluctance toward seeing my patients. (He empezado a sentir un creciente recelo a ver a mis pacientes)
0.94
0.79
CFI 4 I find it more difficult to respond to demanding patients in the way I would want to
(Actualmente me resulta más difícil responder a los pacientes exigentes de la manera que me gustaría)
0.74
0.67
CFI 5 I feel irritated when patients complain
(Me irrita (enoja) cuando los pacientes se quejan)
0.71
0.65
CFI 6 It is becoming increasingly harder for me to handle the complexity of clients with comorbidity
(Cada vez me resulta más difícil manejar la complejidad de los pacientes con varias enfermedades)
0.72
0.62
CFI 7 I have noticed that I distance myself from other peoples’ pain more often than before
(Me he dado cuenta de que me distancio del dolor de otras personas con más frecuencia que antes)
0.50
0.56
CFI 8 My work bores me more often than before
Mi trabajo me aburre más a menudo que antes
0.53
0.58
CFI 9 I have noticed that I try to stay engaged with my patients even though I do not have the energy for it
(Me he dado cuenta de que trato de mantenerme comprometido/a con mis pacientes a pesar de que no tengo la energía para ello)
0.56
0.37
CFI 10 I have noticed that my patience in my personal relationships has dwindled (He notado que mi paciencia en mis relaciones personales ha disminuido)
0.67
0.67
CFI 11 I have started to withdraw from social interaction
(He empezado a retirarme de la interacción social)
0.95
0.83
CFI 12 I feel that I do not have the same energy to engage in the problems of my close ones
(Siento que no tengo la misma energía para involucrarme en los problemas de mis seres queridos)
0.81
0.73
CFI 13 I have started to avoid spare time activities that are intellectually challenging
(He comenzado a evitar las actividades de tiempo libre que son intelectualmente desafiantes)
0.74
0.66
CFI 14 I feel that my workplace provides care that is in accordance with my values (reversed scoring)
(La atención medica en mi lugar de trabajo concuerda con mis valores)
0.67
0.50
CFI 15 I have enough resources at my workplace to provide my patients with the type of care they need (rev)
(Los insumos materiales de mi lugar de trabajo son suficientes para las necesidades de los pacientes)
0.56
0.32
CFI 16 I feel that there are clear rules and regulations for how I should work (rev)
(Siento que hay reglas y regulaciones claras sobre como debo trabajar)
0.69
0.51
Cronbach Alpha
0.90
0.86
0.61
McDonald’s Omega
0.90
0.87
0.62
A confirmatory factor analysis was first run using the diagonally weighted least squares method [70], and three factors with item factor loadings similar to those reported in the article by Eng et al. [21] were obtained. The structure proposed by the author was used, which includes item 9 in Factor 1. A total of 368 subjects were included. The CFA with its original structure (Model 1) showed a CFI = 0.98, TLI = 0.986, RMSEA = 0.069, and SRMR = 0.064.
Anzeige
By placing item 9 in the social life factor according to the results of the EFA (Model 2), a CFI = 0.994, a TLI = 0.993 RMSEA = 0.049, and an SRMR = 0.054 were obtained. By eliminating item 9 (Model 3, Fig. 1), we obtain an improvement in the CFI = 0.996, TLI = 0.994 RMSEA = 0.045, and SRMR = 0.05, so it can be eliminated, although the difference is small (Table 4).
Fig. 1
Confirmatory factor analysis of the CFI scale (15 items). Note: Item 9, which does not have adequate psychometric values, is eliminated. RC = reduced compassion, SL = social life, WK = workplace
Table 4
Confirmatory factor analysis of the CFI
Model
Cronbach´s Alpha
Omega
χ
sig
CFI
TLI
RMSEA
SRMR
Model 1
0.882
0.878
239.11
0.000
0.989
0.986
0.069
0.064
Model 2
0.882
0.878
164.52
0.000
0.994
0.993
0.049
0.054
Model 3
0.878
0.874
129.26
0.000
0.996
0.994
0.045
0.050
How the parametric values of the test improve when item 9 is eliminated
×
Confirmatory factor analysis of the secondary traumatic stress scale (EETS)
Given that the rest of the instruments applied in the set of the applied battery have been previously validated, we proceeded with them to the CFA. The initial CFA of the secondary traumatic stress scale with a three-factor structure via the maximum likelihood estimation method revealed that the original 14 items had inadequate values in the CFI tests = 0.87, TLI = 0.84, RMSEA = 0.10, and SRMR = 0.108. To improve the psychometric values, one item from each dimension was eliminated. Item 2, “I feel useless and disappointed in the work I do (me siento inútil y desilusionado frente al trabajo que desarrollo),” was eliminated from the emotional fatigue dimension. Item 9, “This job has taught me that life ends up giving you what you deserve (este trabajo me ha enseñado que la vida termina dándole a uno lo que se merece),” was eliminated from the belief shaking dimension. Item 1, “This job overcomes me emotionally (me supera emocionalmente este trabajo),” was eliminated from the secondary trauma dimension. The decision was made on the basis of the lowest contribution to the latent variable in the initial CFA. Leaving an 11-item scale that showed an alpha of 0.844, an omega of 0.831, χ = 171.65, df = 41, CFI 0.959, TLI 0.945, RMSEA 0.069 and SRMR 0.049, data that are acceptable according to the usual criteria. These resulted in a revised version of the EETS version, with 11 items remaining.
Confirmatory factor analysis of the Physician Burnout Syndrome Scale (PhBSS)
The initial CFA of the Medical Professional Burnout Scale, which is based on the authors' proposal [39] of a three-factor structure, was confirmed with the maximum likelihood estimation method, showing its original structure adequate values, χ = 236 df = 51 CFI = 0.95, TLI = 0.94, RMSEA = 0.072, and SRMR = 0.045. Thus, the original structure proposed with 12 items from the PhBs is used in this article.
Confirmatory factor analysis of the quality-of-life index (QLI-SP)
The CFA of the quality-of-life index with a single-factor solution using the maximum likelihood estimation method showed adequate psychometric values for the original 10 items in the tests of χ = 197, df = 35, CFI = 0.941 TLI = 0.92, R = 0.119 and SRMR = 0.034. The QLI-SP version with 10 items is thus confirmed.
Confirmatory factor analysis of the reduced Moral Stress Scale. (MMDHPr)
The CFA of the scale (MMDHPr), with 10 items in total and a unidimensional score [57], has been shown to be psychometrically satisfactory. The maximum likelihood estimation method revealed that the 10 items had a similar structure to that of the original proposal [46]; they obtained adequate values in the tests, and the global indicators were adjusted to χ = 108, df = 32 CFI 0.971 TLI 0.96 RMSEA 0.067 and an SRMR 0.038, so the reduced version was maintained in the global analyses.
Correlations between CFI, EETS, PhBS, QLI-SP, and MMDHPr
When performing the convergent validation, we found that the CFI scale (15 items) presents a positive correlation with the EETS with its three dimensions. The CFI scale (15 items) presents a positive correlation with the PhBS with its three dimensions. The CFI scale (15 items) presented a positive correlation with the MMDHPr (0.534 p < 0.001). When divergent validation is performed, the CFI scale (15 items) presents a negative correlation with the SQLI-sp (−0.596 p < 0.001). The correlations by dimension are shown in Table 5.
Table 5
Correlations by dimension between the different instruments
Hypothesis testing
The average CFI score was 2.59 ± 0.75, with a standard error of measurement of 0.24 and a minimal detectable change of 0.66.
With respect to sex, there were slight differences in the CFI score (Male M = 2.46 vs F 2.6, p 0.34). There were no differences in the CFI base in marital statuses (p = 0.73), in the type of work contract (p = 0.275), in the work shift (p 0.182), or in having children (p = 0.38). There was no correlation between age and the CFI score (r = −0.07, p = 0.05). People who reported that they professed a religion had lower CFI scores (2.56 ± 0.73 vs 2.74 ± 0.74, p = 0.003). Higher CFI scores were reported for physicians and health professionals who work in outpatient clinics, those with higher education, and those with higher monetary income, as shown in Table 6. The greater the work experience is, the slight the decrease in the CFI (B = −0.009 LLCI −0.015 ULCI −0.002ULL p = 0.007).
Table 6
Score comparisons between groups
N
CFI M
SD
Profession
Nurse
256
2.45
0.66
Medical resident
257
2.54
0.79
Physician
184
2.88
0.72
Head of medical/nurse department
36
2.6
0.74
F(13.921)
P < 0.001
η2 = 0.05
Workplace
Hospitalized
288
2.67
0.76
Emergency department
92
2.39
0.64
Critical care unit
33
2.7
0.74
Operating room
79
2.42
0.71
Outpatient clinic
150
2.75
0.77
Other
91
2.41
0.7
F (5,727) = 5.45
P < 0.001
η2 = 0.04
Work Shift
Matutine
472
2.61
0.75
Vespertine
113
2.58
0.7
Nocturnal
113
2.5
0.72
Weekend
35
2.8
0.8
F (3,729) = 1.62
P = 0.18
η2 = 0.00
Level of education
Bachelor´s degree
379
2.44
0.73
Master/specialty
257
2.73
0.72
PhD/subspeciality
97
2.85
0.72
F(2,730) = 18.74
P < 0.001 **
η2 = 0.05
Monetary income (dollar)
2,492–12,463
479
2.54
0.75
12,464–24,926
163
2.66
0.72
More than 24,927
91
2.76
0.76
F (2,730) = 4.03
P = 0.02 **
η2 = = 0.01
The Levene test was p < 0.05 in the profession category, so Welch's t test was used. [58] In the remaining studies, the Levene test was not p > 0.05, so ANOVA was used
Discussion
The central objective of this study was to validate the CFI scale in a Spanish-speaking health professional population, showing its convergent relationship with secondary traumatic stress, professional fatigue, and moral stress and its divergent relationship with quality of life.
This was done by examining the psychometric properties of the Hispanic adaptation of the CFI. The result of the EFA has shown a dimensional structure similar to the CFI scale proposed by Eng with good internal consistency (alpha = 0.89) in its 16 original items with three factors, namely, reduced compassion, social life, and workplace, which explain 62% of the variance, as proposed by the version of the original scale proposed by Eng, Nordström [21]. The psychometric results indicate that item 9, “I have noticed that I try to stay engaged with my patients even though I do not have the energy for it,” preferably saturates in the social life factor, which is explained in a Hispanic population by the social representation of being engaged within a marriage or personal relationship [59]. Its deletion is proposed since it does not affect the internal consistency and clarifies the structure of the inventory.
The reduced compassion dimension of the CFI includes 8 items (items 1 to 8) with adequate internal consistency (alpha = 0.89). The behaviors described in the items are contrary to what a compassionate response on the part of health personnel would show toward the patient [60]. The reduction in the willingness to help, judgment of the patient, irritability when complaining, difficulty in responding to problems posed by the complexity of the disease, insensitivity, and apathy when faced with pain are attitudes that indicate a reduction in the psychological resource of compassion [9]. This dimension was correlated with professional fatigue (burnout) in three dimensions (exhaustion, disengagement and loss of expectations), which is in accordance with what was reported in the literature, where a reduction in compassion led to an increase in the prevalence of burnout [61]. Reduced compassion is highly correlated with the emotional fatigue (0.614**) and belief shaking (0.573**) dimensions of the Secondary Traumatic Stress scale, both of which highlight the importance of the affective and cognitive interaction of secondary trauma [62]; however, it is correlated to a lesser extent (0.195**) with the dimension of secondary traumatic symptomatology, indicating that compassion fatigue and secondary traumatic stress are both negative syndromes of attention and care but differ from each other [50]. The correlation of reduced compassion with moral stress was moderate, which is in line with findings in the literature that correlate moral stress with loss of commitment [63]. This result is replicated in our study by the correlation between the loss of expectations of PhBs and the level of moral stress. Quality of life has a negative correlation with reduced compassion, which has been reported in different studies [60, 64].
The Social Life dimension of the CFI includes 4 items (items 10 to 13) with adequate internal consistency (alpha = 0.87) and a high correlation (0.639**) with the Reduced Compassion dimension of the CFI, which shows the strong relationship between the reduction in professional compassion and the affect of personal life [65], which is also affected by the development of burnout [66]. In the present study, this dimension was also positively correlated with professional fatigue and the secondary traumatic stress dimensions of emotional fatigue and belief shaking.
The workplace dimension of the CFI includes 3 items (items 14 to 16) with low internal consistency (alpha = 0.62) and low correlations except with moral stress, which has been reported as a condition for the development of compassion fatigue by favoring a decrease in the expression of empathy [67].
Based on the second objective of our study, to evaluate the prevalence and associated conditions, we found higher scores of compassion fatigue in physicians, which may be secondary to having less work experience than nurses and a greater workload.
In this study, people who reported professing a religion had a lower CFI score, which has been reported to be protective against the development of compassion fatigue. However, the cross-sectional design of the study does not allow for the establishment of a cause‒effect. Unlike other studies, we did not find that sex or other sociodemographic factors were correlated with CFI scores [68].
Limitations
Due to its cross-sectional design, this study prevents the establishment of a causal relationship. The study population is distributed in several hospital and clinic settings, making it difficult to establish a relationship between workplace settings and compassion fatigue scores. Finally, this Spanish version of CFI used a new theoretical model that allows to measure reduced compassion during clinical encounters. It is advisable to conduct specific studies to ensure its validity in a cultural context different from Mexico.
Implications for nursing education
Nurses and physicians working in hospital and clinic settings are exposed to trauma and suffering from patients. During their training, it is necessary to teach them about occupational diseases that they may experience during their practice and offer them validated instruments for self-monitoring to initiate and preserve recovery and self-care activities.
The findings of this study show that lack of work experience has been related to higher CFI test scores, so it is important to educate during the training years on protective factors such as financial education, management of work stress, promotion of an ethical work environment, and healthy lifestyles.
Conclusions
The Latin adaptation of the CFI instrument has high instrumental validity and is good at discriminating from other occupational syndromes in a Spanish-speaking population.
Item 9 is proposed to be eliminated because its cultural connotations are more closely related to personal life. Its deletion does not diminish criterion validity or affect the psychometric value of the test, which is why a 15-item CFI is proposed.
Compassion fatigue decreases with increasing work experience. In this sample, the work context with the highest scores for compassion fatigue was outpatient consultation followed by hospitalization, so it is important to promote interventions that facilitate recovery in the personnel assigned to this area.
In conclusion, the CFI is a good instrument for evaluating compassion fatigue, which is understood as a reduction in the expression of compassion by healthcare personnel when providing care to patients, with consequences at the personal and workplace levels.
Acknowledgements
We would like to thank Pedro Juarez Rodríguez for their support and kind advice on using Survey Monkey. To Esteban Gonzalez Diaz and Jason Miguel Aragon for their feedback while translating the instrument.
Clinical trial number
Not applicable.
Declarations
Ethics approval and consent to participate
Ethics approval of the protocol was obtained from the Local Ethics Committee No. 13018 of the 'Unidad Médica de Alta Especialidad, Hospital de Especialidades, Centro Médico Nacional de Occidente, Lic. Ignacio García Téllez,' which belongs to the social security network 'Instituto Mexicano del Seguro Social' (IMSS), with the government registration number COFEPRIS 17 CI 14 039 114, CONBIOETICA 14 CEI 20190123 and was registered as R-2024–1301-043. Electronic consent was obtained from all participants before starting the survey. No personal data that would allow identification were collected. Before signing the electronic consent form, they were informed of the handling of confidential data.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
