Prevalence and contextual factors associated with compassion fatigue among nurses in northern Uganda

Background Compassion fatigue is associated with negative consequences that undermine workplace performance. However, literature is scarce on compassion fatigue among nurses in the context of Uganda who are at higher risk of compassion fatigue owed to the nature of their occupation and the unfavorable work environment. We aimed to assess the prevalence and predictors of compassion fatigue among nurses in Northern Uganda. Methods We conducted a cross-sectional survey among 395 randomly selected nurses from two referral and four general hospitals in Northern Uganda. Data was collected using a self-administered questionnaire. Data analysis consisted of descriptive statistics, cross-tabulations, and logistic regression at a 95% level of significance in SPSS version 25. Results Of 395 nurses who took part in the study, 58.2% were female, 39.8% had a diploma, 47.1% were single, and 32.4% had worked for between 11–15 years. Close to 50% of the nurses experienced compassion fatigue. The predictors of compassion fatigue among the participants were; workplace bullying (AOR: 3.83; 95% CI: 1.99–7.36; p = <0.001) career enhancement opportunities (AOR: 9.16; 95% CI: 2.32–36.22; p = 0.002; and remuneration (AOR: 7.30; 95% CI: 2.73–19.49; p = <0.001). Conclusion More than 3 in 4 nurses in Northern Uganda experience compassion fatigue. The Ministry of Health together with other stakeholders should strive to increase career enhancement opportunities for nurses, improve nurses’ remuneration and improve the overall workplace environment to prevent compassion fatigue among nurses in the country.

Introduction compassion fatigue and work overload and years of working experience [26]. However, other studies show that nurses with more years of experience and work overload had less compassion fatigue [27,28].
The susceptibility of compassion fatigue among nurses may be even higher in northern Uganda not only due to shortage of nurses and overstretched health systems but also historical factors. The region was affected by a protracted brutal civil war that claimed tens of thousands of lives and many others displaced [29]. The war described as the "worst forgotten humanitarian crisis in the world" lasted from1986 to 2006. The Juba peace truce arranged in 2005 brought a shift in humanitarian efforts to economic recovery and health rehabilitation. The significant response to such events relied greatly on nurses. Admittedly the war ended but its impact on the health sector and individuals is evident [30]. Therefore, nurses are still working with critical patients. Nurses experiencing compassion fatigue have difficulties making appropriate clinical decisions and potentially are at risk of harming patients. Subsequently, scholars agree that compassion fatigue is an occupational hazard that professionals need to address [31]. Moreover, this topic has not been sufficiently studied especially in Uganda. The current study assessed the prevalence and associated factors of compassion fatigue in nurses in northern Uganda.

Study design and setting
A cross-sectional method study design was employed in one survey covering nurses working in northern Uganda. The study was conducted in two referral and four general hospitals in northern Uganda between 21 st February and 30 th March 2021. The region experienced a twodecade-long insurgency caused by the Lord's Resistance Army (LRA) rebels that claimed thousands of lives and left many people displaced and homeless. Subsequently, the healthcare system in the region is weak and in a state of recovery. This has created an unsuitable working environment for the health workforce particularly, the nurses who are tasked with looking after critically ill patients.

Participants and sample size
The study included all nurses registered with the Uganda Nurses and Midwives Council (UNMC) from different health centers in Northern Uganda. Nurses who were busy and or attending to critically ill patients at the time of data collection were excluded from the study. The sample size was calculated using a single population proportion formula [32]. The z-score (z) was 1.96, the margin of error (e) was 0.05, and the prevalence (p) for the known population was 0.5 since the proportion of nurses with compassion fatigue in Uganda is not known. The calculated sample size was 427 plus a 10% allowance for nonresponse. Eligible participants were randomly enrolled in the study until the desired sample was realized.

Data collection tool and procedures
We used a self-administered questionnaire to collect data from the nurses across the six health facilities. We used a validated revised Stamm's ProQOL V-5 tool to assess compassion fatigue among the nurses (r = 0.82) [10]. ProQOL V-5 has 30 items measuring compassion fatigue, compassion satisfaction, and burnout but only ten items measuring compassion fatigue were analyzed (S2 Appendix). The ProQOL is currently in its fifth revision of the original tool called Compassion Fatigue Self-Test Survey tool developed by Figley in 1995 [33] on a 5-point Likert scale of 1 to 5. The instrument has undergone a rigorous psychometric assessment to improve sub-scale validity and reliability [10]. The compassion fatigue sub-scale has an alpha scale reliability of 0.75 and inter-scale correlations indicate 5% shared variance (r = -0.14; coσ = 2%; n = 1187) [10]. Compassion fatigue in this study was binned as a dichotomy of "yes" and "no". We also captured demographic and workplace contextual factors conceptualized to be associated with compassion fatigue. The completion of the questionnaire by the study participants lasted for 20 to 30 minutes.

Data management and analysis
Data were analyzed using SPSS version 23.0. Every questionnaire was checked for completeness at the end of each interview. A data entry screen was created in SPSS version 23 with checks to ensure accuracy during entry. Data were scanned for out-of-range and missing values before commencing data analysis. Categorical variables were analyzed using descriptive statistics, cross-tabulation, and logistic regression at a 95% level of confidence. Variables with a p<0.05 at the binary logistic regression level were included in the multivariate logistic regression model. The model fitness was checked using the Hosmer-Lemeshow test and a backward conditional logistic regression method was used to determine variables that were independently associated with the outcome variable. The effects of the independent variables were expressed as odds ratio (OR) and associated 95% confidence intervals (CI). The outcome variable of the study was compassion fatigue transformed into a dichotomous variable of yes and no.

Ethical considerations
This study was approved by the Makerere University School of Medicine Research and Ethics Committee Institutional Review Board (UG-REC-027). Participants in this study were recruited based on written informed consent (S1 Appendix) and confidentiality was maintained throughout the entire research protocol process. The study was anonymous and participants had the right to withdraw at any time.

Results
Out of the 427 calculated sample size, only 409 responded to the questionnaire giving a response rate of 95.7%. However, 12 questionnaires were excluded due to missing information on compassion fatigue giving a total of 395 participants for the final analysis.

Prevalence of compassion fatigue among nurses in Northern Uganda (N = 395)
Based on the ProQOL manual guidelines, 49.11% (n = 194) had high levels of compassion fatigue, 29.62% (n = 117) reported average levels of compassion fatigue and 21.27% (n = 84) indicated that they had low levels of compassion fatigue (Table 2).

Workplace contextual factors associated with compassion fatigue among nurses in Northern Uganda (N = 395)
The majority of the nurses (56.2%) experienced workplace bullying, 55% were less satisfied with their career enhancement prospectus at their workplace while the majority (52%) were satisfied with their current remuneration ( Table 3).

Predictors of compassion fatigue among nurses in Northern Uganda
In the multivariate logistic regression model in Table 4, the factors that were independently associated with compassion fatigue among the study respondents were: workplace bullying (AOR: 3.83; 95% CI:

Discussion
We examined the prevalence and contextual factors associated with compassion fatigue among nurses in northern Uganda. Findings in our study demonstrated that close to 50% had high levels of compassion fatigue. Our findings suggest that the nurses are exposed to traumatic materials of their patients experiencing distress. The presence of small acts of empathy that takes on the meaning and mark patients' continuing memories of the overall experience of the health care received is enough to cause compassion fatigue. Additionally, the findings suggest that nurses may be lacking sufficient skills in coping with the traumatic experiences of their patients. The insufficient competencies may be compounded by poor working conditions and an unsupportive environment [34]. The adverse effects of compassion fatigue impend the provision of empathetic care to patients [35]. Overall, the results confirm results from previous studies which showed that working with patients who are in pain and witnessing traumatic labor impact the mental health of nurses [28,35]. Compared to previous studies, our findings reported higher levels of compassion fatigue than 44.8% in Greece, 38.9% in Poland, and 24.2% in China [36][37][38]. However, our results were lower than 59%, 52%, 75%, and 64% reported by nurses in Portugal, America, Scotland, and Jordan respectively [14-17, 39, 40]. This incongruence may be attributed to differences in sample sizes, departments, cultures, regions, resources, and working environments in different countries. Genders, marital status, and education were not significantly associated with compassion fatigue. These results are not surprising because previous studies reveal inconsistent findings. However, work experience was significantly associated with compassion fatigue. Our results are in tandem with previous research which shows no association between compassion fatigue and demographic factors [25,39,41,42]. Other studies, however, reveal a statistically significant association between compassion fatigue and demographic factors [38,43,44]. According to Hill and Stephens [45], compassion fatigue lowered with increasing working experience. In a systematic review of 71 studies, an association between socio-demographic characteristics and compassion fatigue was inconsistent [46]. Therefore the association between compassion fatigue and demographic factors remains inconclusive. Results indicate that nurses who experienced workplace bullying were more likely to experience compassion fatigue compared to their counterparts who did not experience workplace bullying. This result is not surprising because workplace bullying is more frequently reported among nurses than people in other professions [47] and has negative ramifications. This bullying usually takes numerous forms including job-related bullying, intimidation-related bullying, and personal bullying. The negative ramifications of workplace bullying include loss of confidence, reduced self-esteem, depression, and compassion fatigue [48] as reflected in our study. Our results are in line with the findings of Kim and colleagues [49] which showed that workplace bullying had a significant effect on compassion fatigue. Findings from Laeeque and colleagues' [50] study on compassion fatigue in nurses in Pakistan revealed similar results.
We observed in the current study that nurses who expressed satisfaction with their career enhancement opportunities in their current workplace were nine times more likely to experience compassion fatigue compared to their colleagues who were less satisfied with their career enhancement prospects. This is a surprising result because it is a well-known fact that lack of career enhancement opportunities leads to work-related stress [51] which may precipitate compassion fatigue. However, research indicates that the factors causing compassion fatigue in nurses include patient care, lack of skills, lack of communication, feeling the responsibility for the deadly ill, lack of communication, lack of social support, and increased intensity of patients [52,53]. Therefore, while little is known about the association between career enhancement opportunities and compassion fatigue, it is apparent that compassion fatigue is a feeling resulting from experiencing an event that includes trauma-related stress [54]. In contrast to our findings, Andriani et al., [41] found a significant association between compassion fatigue and pay satisfaction.
Nurses who reported that their current remuneration was satisfactory were seven times more likely to experience compassion fatigue compared to those who were less satisfied with their current pay. This was an unexpected result which nevertheless is understandable. First, compassion fatigue is an occupational hazard because simply being in nursing places one at risk [55] regardless of the remuneration. Second, compassion fatigue is a natural result of empathy. In the context of the current study, it seems that the nurses who felt that they were satisfied with their remunerations empathized with their patients, identified themselves with patients more easily, and were exposed to traumatic materials of the patients. According to Figely [54], the more the level of empathy the higher the risk for compassion fatigue. These nurses seem to experience compassion fatigue because of proximity to tragedy over time, they seem not to separate themselves from the source of work-related stress [56]. Third, these nurses seem to be overcommitted to their jobs and do not invest in their personal quality of life putting them at risk for compassion fatigue [57].

Limitations of the study
The results of this study aimed at providing clarity regarding the prevalence and the predictive factors of compassion. However, we are unable to infer causality from the observed association due to the cross-sectional design used in the study. The study also used self-report questionnaires with the possibility of recall bias in the results. Additional research is needed to expand and clarify the present results of compassion fatigue among nurses in Uganda.

Conclusion
More than 3 in 4 nurses in Northern Uganda experience compassion fatigue. Workplace bullying, career enhancement opportunities, remuneration, and working experience are predictors of compassion fatigue. The findings of this study give insights into what should be done to improve the workplace environment of nurses in Uganda. These results can be used by nursing leadership and other stakeholders to create an enabling environment for nurses to promote job satisfaction and avoid the brain drain of the skilled nursing workforce. More research is needed to understand how compassion fatigue is correlated with burnout and job satisfaction.