Read the Full Article
Hospitalized patients with cancer may undergo surgeries, receive complex treatments that cannot be administered in ambulatory settings, and receive care during terminal disease. In acute care settings, oncology nurses provide care to patients 24 hours a day. They control patients’ physical symptoms while supporting them emotionally, and routinely witness suffering and death. Boyle (2011) stated, “[Nurses] cannot leave the situation after bad news is shared or a death has occurred” (para. 13). At times, oncology caregivers also develop close relationships with patients. When patients die, staff may have a difficult time caring for other patients (Wenzel, Shaha, Klimmek, & Krumm, 2011). These situations place staff members at risk for developing compassion fatigue (CF).
Joinson (1992) first identified the phenomenon of CF in her study about emergency department nurses who were losing their nurturing abilities. The literature also shows that CF is present in oncology nurses. Bush (2009) stated that CF is “an emotional state with negative psychological and physical consequences that emanate from acute or prolonged caregiving of people stricken by intense trauma, suffering, or misfortune” (p. 28). Showalter (2010) explained that CF can “negatively impact the ability to provide services” and that caregivers may believe that their role is “hard labor rather than a labor of love” (p. 240). CF may negatively affect caregivers emotionally, socially, physically, spiritually, and intellectually. Some effects of CF are a lack of energy, reduced output, diminished performance, unresponsiveness, apathy, callousness, indifference, appetite changes, and sleep disturbances (Abendroth & Flannery, 2006; Coetzee & Klopper, 2010). CF may impair caregivers’ ability to provide high-quality care.
Organizational factors to reduce CF should focus on caring for the caregiver. However, a survey of Oncology Nursing Society (ONS) chapter leaders revealed that great variability exists in the number of healthcare organizations in the United States that provide interventions that address CF (range 0%–60%). Fifty percent of the respondents had pastoral care, and only 55% of organizations provided educational offerings on workplace- related coping (Aycock & Boyle, 2009). Staff education should include opportunities that focus on end-of-life care (Aycock & Boyle, 2009; Becze, 2012; Hildebrandt, 2012), such as the End-of-Life Nursing Education Consortium (ELNEC) project. When a patient dies, it is imperative to acknowledge staff grief (Hildebrandt, 2012). Providing a safe environment for nurses to discuss death with peers, nursing leaders, or pastors provides them with an opportunity to acknowledge the loss (Aycock & Boyle, 2009; Becze, 2012; Hildebrandt, 2012; Wenzel et al., 2011). In the literature, interventions to address CF include support groups (Absolon & Krueger, 2009; Fetter, 2012), journaling (Fetter, 2012; Macpherson, 2008; Reimer, 2013), and debriefing sessions (Keene, Hutton, Hall, & Rushton, 2010).
Impetus for Quality Improvement Project
After reviewing the results of Press Ganey scores at Karmanos Cancer Center in Detroit, Michigan, a National Cancer Institute (NCI)–designated comprehensive cancer center, the current authors hypothesized that staff may have been experiencing CF based on patients’ satisfaction scores with nursing care. Analysis of the scores from the first quarter of 2012 to the second quarter of 2014 revealed that patient satisfaction scores with nursing care on acute care units remained relatively stable; however, differences in quarterly mean scores for the medical (hematology/blood and marrow transplantation) and the blended medical-surgical (surgical and gynecologic oncology) units existed (see Table 1). The medical units historically sustained higher patient satisfaction scores.
Multiple variables (nurse–patient ratios, nursing hours per patient day, case-mix index, discharges, patient deaths, and medications per patient per day) were assessed to determine if they contributed to patient satisfaction outcomes. No significant correlations were found; however, the units with lower scores had more deaths. The death rate (per 1,000 patient days) was examined to determine if a relationship existed between unit death rate and patient satisfaction scores. The results showed a statistically significant negative correlation between the death rate and patient satisfaction levels with nursing care on the Press Ganey survey (see Table 2). As the death rate increased, the surviving patients’ perception of care decreased. It was speculated that staff who experienced more deaths on their unit were experiencing CF, which could have contributed to patient satisfaction outcomes. Support for this was endorsed by the results of an informal fatigue inpatient staff survey in 2014 that found staff requesting support after a patient death. This led to further inquiry. After reviewing the literature related to CF and discovering the correlation between death rates and patient satisfaction scores in this oncology setting, the current authors implemented a quality improvement project to support staff following patient death.
Morse’s (2001) praxis theory of suffering was used to better understand the feelings experienced by staff grieving the loss of a patient. Although the theory was designed to assist nurses as they care for patients (and family members), in this project, it was used for staff who may have been experiencing CF as they coped with prolonged patient suffering and death. Morse (2001) suggested that people experience two phases of suffering: enduring and emotional suffering. Suppression of emotions occurs during the enduring phase, when the person focuses on the present moment. People are aware that if they respond emotionally to a loss, they will be unable to function (Morse, 2011). Gerow et al. (2010) described this phenomenon as a “curtain of protection to mitigate the grieving process and allow them to provide supportive nursing care” (p. 124). If left unrecognized, it may lead to the development of CF. During the emotional suffering phase, people recognize the meaning of the loss (Morse, 2011). Bush (2009) spoke of the importance of staff seeking out others for support, stabilizing their feelings, and letting go of self-blame and guilt when such thoughts arise. This process provides comfort to people who are suffering. In the oncology setting, this can be achieved by assisting staff in resolving grief and loss that can occur by witnessing prolonged patient suffering and death. An adaptation of Morse’s (2011) framework was used to guide the development of the intervention (see Figure 1).
Project Aim and Objectives
The aim of this quality improvement project was to develop an intervention to support staff experiencing grief after patient death. The long-term goal was to decrease staff CF, if present, and increase patient satisfaction with nursing care. The clinical question was, “Is there a relationship between CF and patients’ perception of nursing care at this NCI-designated comprehensive cancer center?” The project objectives were to (a) assess the current level of CF of direct care providers in the inpatient oncology setting; (b) determine if the current level of CF for direct patient care staff differed among the medical and blended medical-surgical units, as well as between nurses and assistive personnel; (c) determine if the level of CF of direct patient care staff differed from baseline to the completion of the intervention at three months; and (d) identify if the intervention had an impact on patient satisfaction scores after controlling for the death rate.
A mixed-methods sequential design was used to determine if a grief support intervention positively affected staff CF and patient satisfaction scores. Quantitative data were collected on the pre- and postintervention surveys. The postintervention survey also contained a qualitative question regarding the intervention.
The Wayne State University and University of Detroit Mercy Institutional Review Boards granted concurrence of exemption status before study implementation.
Setting and Participants
The study was conducted on two medical and two blended medical-surgical units in an NCI-designated comprehensive cancer center located in the midwestern United States. The center has 112 licensed acute care beds and an average daily census of 75 patients. When the project began in March 2015, 117 RNs and 69 assistive personnel were eligible for participation (see Table 3).
CF was measured using the Professional Quality of Life (PROQOL) scale, version 5 (Stamm, 2009), which also measures compassion satisfaction (CS). It is the most commonly used tool to measure response to effects of working in stressful environments (Stamm, 2010). The PROQOL contains 30 Likert-type questions that assess situations and feelings that may have occurred in the past 30 days. The scale, which includes CS and CF subscales of burnout and secondary traumatic stress, includes 10 questions contributing to each score. Scores ranged from 1 (never) to 5 (very often). The balance between the CS and CF scores indicate an individual’s level of quality of life (scores less than 23 are considered low, scores from 23–41 are considered average, and scores that are 42 or greater are considered high) (Stamm, 2009). Optimal PROQOL scores are high CS and low burnout and secondary traumatic stress. The PROQOL was administered to participants pre- and postintervention. In this project, the Cronbach alphas were 0.86 preintervention and 0.86 postintervention for CS, 0.73 preintervention and 0.74 postintervention for burnout, and 0.67 preintervention and 0.78 postintervention for secondary traumatic stress.
In addition to the PROQOL, pre- and postintervention surveys (developed by the project director), including demographic information (age, gender, unit, and role), were administered. The preintervention survey also asked participants how many years they worked on the unit (mean years for Ns = 5.79, mean years for oncology nurses = 7.05), as well as their nursing profession and highest degree in nursing, if applicable. The highest nursing degree of 2 participants was a diploma, whereas 18 nurses had an associate degree, 47 nurses had a bachelor’s degree, and 1 participant had a master’s degree. The postintervention survey included a question regarding the completion of the preintervention survey and two questions that asked how helpful the debriefing session was in acknowledging grief and loss.
Patient satisfaction scores were measured using second quarter 2015 data from the Press Ganey survey, a widely recognized patient satisfaction survey whose results contain data from 10,000 healthcare facilities, including 69% of the Magnet hospitals and 82% of the best hospitals per U.S. News and World Report (Press Ganey, n.d.). The survey has high reliability. The Cronbach alpha was 0.97 (Press Ganey, 2010). Press Ganey data were obtained as an aggregate; therefore, no patient identifiers were present during data analysis.
In March 2015, flyers were posted on acute care units, inviting staff to the preintervention sessions. A presentation outlining the project, an information sheet (including risks and consent process), and a preintervention survey (i.e., demographic questions and the PROQOL scale) were available for staff during the sessions.
Unit-based debriefings began after each patient death and lasted three months. This time frame was chosen to correspond with quarterly patient satisfaction data. The chaplain and nursing supervisors, who received one-on-one training from the project director, facilitated the debriefings; at least one of them was onsite to support staff 24 hours a day, 7 days a week. Training included discussion of CF literature, their role in the project, and directions for conducting the sessions. The debriefing sessions consisted of three questions related to patient care, based on Koloroutis’s (2004) relationship-based care model, which was previously adopted by the hospital. Questions included the following:
• How did you help the patient/family through this transition?
• What example of colleague collaboration was most noteworthy in this patient experience?
• What impact will this patient’s death have on you?
The sessions occurred before the end of the shift on which the patient died or at shift change to allow more staff members to participate. Sessions were not mandatory; however, staff members were encouraged to attend. During debriefings, facilitators recorded the date, time, and length of the session, unit, and number of participants. Postintervention, flyers were posted notifying staff of sessions in which they could complete a survey.
The completion of surveys and attendance in debriefing sessions provided consent for participation in this project. Participation was voluntary, and the staff members participating had minimal risk (i.e., they may experience emotional discomfort while acknowledging feelings of grief and loss).
The unit with the highest percentage of completed pre- and postsurveys selected one of two incentives: a coffee brewing system or a supply of coffee.
Univariate descriptive statistics and frequencies were extracted with SPSS®, version 22, for demographics and pre- and postintervention PROQOL scores. Reliability analysis was performed on PROQOL scores pre- and postintervention. Independent t tests were used preintervention to determine if a difference in scores existed between groups, and postintervention to determine if a difference in scores existed between staff who did and did not participate in the intervention. Last, analysis of covariance (ANCOVA) was used to determine if a difference in Press Ganey scores for nursing care existed pre- and postintervention after controlling for the death rate. All data were entered in two separate data sets to ensure the accuracy of data entry and results.
Of the 186 employees eligible for participation in the preintervention survey, 107 (58%) returned surveys; of the 183 employees eligible for participation in the postintervention survey, 136 (74%) returned surveys. Seventy percent (n = 91) of those who completed a postintervention survey also completed a preintervention survey. Of those who took the postsurvey, 42 had participated in the intervention and 89 had not. Table 4 provides pre- and postintervention demographic results.
The preintervention PROQOL raw scores for direct care providers indicated an average level of CS (mean = 40.81, SD = 4.932), low burnout (mean = 22.5, SD = 4.505), and average secondary traumatic stress (mean = 24.17, SD = 4.051). No significant difference existed in burnout (t  = 0.128, p = 0.898) or secondary traumatic stress (t  = 0.743, p = 0.459) between the medical and the blended medical-surgical units. Likewise, no significant difference existed in burnout (t  = 0.121, p = 0.904) or secondary traumatic stress (t  = 1.595, p = 0.114) between RNs and assistive personnel (see Table 5).
Sixteen patients died during the second quarter of 2015. The number of deaths ranged from one to eight per unit. The chaplain and nursing supervisors conducted 15 debriefing sessions, respectively, lasting 3–30 minutes, with a mean of 9.87 minutes (SD = 6.643). The average number of participants in the sessions was 4.53 (SD = 2.167).
To assure confidentiality, the decision was made not to match participants; therefore, postintervention results included responses only from those who stated they had completed the preintervention survey (N = 91). Although CF decreased overall in both groups, no significant difference in burnout or secondary traumatic stress was observed for those who did and did not participate in the debriefings. An ad hoc analysis revealed that those who participated in the debriefings had significantly higher CS (mean = 43.39, SD = 4.1366) compared to those who did not (mean = 40.19, SD = 4.5998; t  = 3.221, p = 0.002). Participants rated the session(s) in acknowledging their grief and loss on a Likert-type scale ranging from 1 (very unhelpful) to 5 (very helpful). Most participants (60%) felt that the debriefings were helpful, with a mean score of 3.73 (SD = 1.025), indicating they were somewhat helpful. Comments were categorized into those who thought the debriefings were helpful (sharing of emotions; provided an outlet for closure to speak with others who felt similar) and unhelpful (did not know the patient; dignified death; the family was at peace).
The Press Ganey scores from the second quarter of 2015 and the same quarter from the previous three years were compared. In the second quarter of 2015, 32% of patients returned surveys, which was comparable to previous years (32%–26% response rate). Although all the patient satisfaction with nursing measures scores were higher in 2015 than in previous years, “skill of the nurse” was the only item that was significantly higher (p = 0.007) in 2015. ANCOVAs showed no significant differences for any of the nursing measures pre- and postintervention after controlling for the death rate.
This quality improvement project was designed to assess RNs and oncology care associates’ current level of CF and to identify if a grief support intervention after patient death would affect the surviving patients’ perception of care. Although staff did not appear to have CF at either data collection point, they found the debriefing sessions helpful and that the patients’ perception of care was higher compared to the same quarter in the previous three years. The results were not statistically significant, but to the current authors’ knowledge, this is the first study to examine the relationship between death rate and patient satisfaction with CF.
After the pilot period, the direct care staff members’ PROQOL scores were near ideal: high average CS (mean = 41.87, SD = 4.7055) and low burnout (mean = 21.61, SD = 4.7178) and secondary traumatic stress (mean = 22.18, SD = 4.6501). Although Stamm (2005) and Yoder (2010) used earlier versions of the PROQOL scale, their participants did not appear to have high CF either. Those studies and the current one had high reliability. One explanation for the reported low levels of CF risk in these studies is that the effects of CF may not be able to be measured on a scale. In the current study, people who choose to work at a comprehensive cancer center may feel called to care for patients and their families when they need it the most and, therefore, do not view their work as hard labor, as suggested by Showalter (2010).
Although the PROQOL scores of the staff were average at baseline, differences between the nursing groups existed. In the literature, risk factors for developing CF include being aged younger than 40 years, having fewer than 10 years of experience, and having a bachelor’s degree (McSteen, 2010; Potter et al., 2010). In the current study, when outcome variables were compared, nurses aged younger than 40 years had lower CS (mean= 39.31) (mean = 43.05, t  = –2.399, p = 0.018) and higher burnout (mean = 23.36) compared to nurses aged 40 years or older (mean = 20.32, t  = –3.186, p = 0.002), respectively. However, no significant difference existed in secondary traumatic stress. Nurses with fewer than 10 years of experience had significantly lower CS (mean = 39.59) than those with 10 or more years of nursing experience (mean = 44, t  = –3.254, p = 0.002), but no significant difference existed in burnout or secondary traumatic stress. In addition, no significant differences were found between nurses with an associate degree or diploma compared to those with a bachelor’s or master’s degree. The results of this quality improvement project support findings in the literature that risk factors for CF include younger age and less experience (McSteen, 2010; Potter et al., 2010).
Most staff members (n = 25) participating in the intervention reported that the debriefings were helpful in acknowledging grief and loss. These results are similar to those of Absolon and Krueger (2009), Fetter (2012), and Keene et al. (2010), whose participants found the support/bereavement interventions helpful in addressing CF, grief, or loss. Although most staff members reported that the debriefings were helpful, 30% selected the neutral option. Because about one-third chose this response, future studies should examine other potential options to evaluate the impact of the intervention.
Although patient satisfaction scores improved for all the nursing measures in 2015 compared to previous years, no significant differences were detected after controlling for the death rate. Although the nonsignificant relationship between death rate and patient satisfaction was disappointing, this quality improvement project explored underlying explanations for patient satisfaction scores that could have resulted from CF. Boyle (2011) stated, “Until the consequences and ramifications of compassion fatigue can be linked to more concrete outcomes, it will remain an elusive aspect of nurses’ work” (para. 23).
Participants were not matched at data collection points, so the current authors were unable to determine significant differences in staff members pre- and postintervention. This may have revealed additional differences in PROQOL scores that were not evident in these data. In addition, the number of deaths was variable, which limited the number of staff eligible to participate in debriefings. Seven facilitators provided the intervention, causing variation in delivery and duration of the debriefing sessions, and staff reported that the units were busy, which may have prevented others from participating in the debriefings. The current study also had a small sample size. Although the response rates were excellent, this project was conducted on only four units of one comprehensive cancer center.
Implications for Nursing Practice
The literature indicates that CF frequently occurs in the oncology setting, but only a few participants in the current study had heard of the phenomenon at project implementation. Nurses in similar oncology settings may benefit from CF education and interventions to address the phenomenon. This study also observed patient satisfaction data through a different lens. Too often, the face value of patient satisfaction scores is used for evaluation of care delivery without looking for underlying factors contributing to the results (e.g., death rate). In addition, this study sample included assistive personnel, who spend significant time caring for patients at the end of life. The current study showed no difference in CF between RNs and assistive personnel; therefore, it is essential that self-care and bereavement programs include them. Last, the nursing participants in this project were younger than the U.S. average age of nurses, which is 50 years (American Nurses Association, 2014). With the baby boomer generation starting to retire and the nursing job growth that may follow with that trend, nurses must mentor and support the next generation of oncology nurses.
PROQOL scores were average at baseline, which left a narrow margin for improvement; therefore, near ideal PROQOL scores postintervention were viewed as a success. Ad hoc analysis further substantiated the improvement, with significantly higher CS scores for debriefing participants compared to those who did not participate. The participation rate was a notable outcome, which may indicate that staff were interested in grief support. In addition, there was improvement in all nursing satisfaction measures postintervention. These findings are consistent with the current literature and add a viable alternative for addressing staff grief in the oncology setting.
The authors gratefully acknowledge Margaret Dimond, PhD, and Susanne Suchy, MSN, RN, AOCNS®, who served as content experts for this project.
About the Author(s)
Lisa M. Zajac, DNP, RN, ANP-BC, OCN®, is the director of clinical informatics at the Karmanos Cancer Center in Detroit, MI; and Katherine J. Moran, DNP, RN, CDE, FAADE, is an associate professor and Carla J. Groh, PhD, PMHNP, FAAN, is a professor, both in the College of Health Professions at the University of Detroit Mercy in Michigan. The authors take full responsibility for this content. Groh is supported by a grant from the Blue Cross Blue Shield Foundation. The article has been reviewed by independent peer reviewers to ensure that it is objective and free from bias. Zajac can be reached at email@example.com, with copy to CJONEditor@ons.org. (Submitted June 2016. Accepted December 19, 2016.)
Abendroth, M., & Flannery, J. (2006). Predicting the risk of compassion fatigue: A study of hospice nurses. Journal of Hospice and Palliative Nursing, 8, 346–356.
Absolon, P., & Krueger, C. (2009). Compassion fatigue nursing support group in ambulatory care gynecology/oncology nursing. Journal of Gynecologic Oncology Nursing, 19, 16–19.
American Nurses Association. (2014). The nursing workforce 2014: Growth, salaries, education, demographics and trends. Retrieved from http://bit.ly/1t8tLOp
Aycock, N., & Boyle, D. (2009). Interventions to manage compassion fatigue in oncology nursing. Clinical Journal of Oncology Nursing, 13, 183–191. doi:10.1188/09.CJON.183-191
Becze, E. (2012). To retain oncology nurses, offer grief resolution. ONS Connect, 27, 16–17.
Boyle, D.A. (2011). Countering compassion fatigue: A requisite nursing agenda. Online Journal of Issues in Nursing, 16, 2. doi:10.3912/OJIN.Vol16No01Man02
Bush, N.J. (2009). Compassion fatigue: Are you at risk? Oncology Nursing Forum, 36, 24–28. doi:10.1188/09.ONF.24-28
Coetzee, S.K., & Klopper, H.C. (2010). Compassion fatigue within nursing practice: A concept analysis. Nursing and Health Sciences, 12, 235–243. doi:10.1111/j.1442-2018.2010.00526.x
Fetter, K.L. (2012). We grieve too: One inpatient oncology unit’s interventions for recognizing and combating compassion fatigue. Clinical Journal of Oncology Nursing, 16, 559–561. doi:10.1188/12.CJON.559-561
Gerow, L., Conejo, P., Alonzo, A., Davis, N., Rodgers, S., & Domian, E.W. (2010). Creating a curtain of protection: Nurses’ experiences of grief following patient death. Journal of Nursing Scholarship, 42, 122–129. doi:10.1111/j.1547-5069.2010.01343.x
Hildebrandt, L. (2012). Providing grief resolution as an oncology nurse retention strategy: A literature review. Clinical Journal of Oncology Nursing, 16, 601–606. doi:10.1188/12.CJON.601-606
Joinson, C. (1992). Coping with compassion fatigue. Nursing, 22(4), 116.
Keene, E.A., Hutton, N., Hall, B., & Rushton, C. (2010). Bereavement debriefing sessions: An intervention to support health care professionals in managing their grief after the death of a patient. Pediatric Nursing, 36, 185–189.
Koloroutis, M. (Ed.). (2004). Relationship-based care: A model for transforming practice. Minneapolis, MN: Creative Health Care Management.
Macpherson, C.F. (2008). Peer-supported storytelling for grieving pediatric oncology nurses. Journal of Pediatric Oncology Nursing, 25, 148–163. doi:10.1177/1043454208317236
McSteen, K.L. (2010). Compassion fatigue in oncology nursing: A witness to suffering. Retrieved from http://media.oncologynurseadvisor.com/documents/32/ona_ce_1010-1012_7955...
Morse, J.M. (2001). Toward a praxis theory of suffering. Advances in Nursing Science, 24, 47–59.
Morse, J.M. (2011). The praxis theory of suffering. In J.B. Butts, & K.L. Rich (Eds), Philosophies and theories for advanced nursing practice (pp. 569–597). Sudbury, MA: Jones and Bartlett Learning.
Potter, P., Deshields, T., Divanbeigi, J., Berger, J., Cipriano, D., Norris, L., & Olsen, S. (2010). Compassion fatigue and burnout: Prevalence among oncology nurses [Online exclusive]. Clinical Journal of Oncology Nursing, 14, E56–E62. doi:10.1188/10.CJON.E56-E62
Press Ganey. (2010). [Inpatient (in) survey psychometrics report]. Unpublished raw data.
Press Ganey. (n.d.). Explaining your Press Ganey data to others. Retrieved from https://helpandtraining.pressganey.com/resources/hospital-benchmarks
Reimer, N. (2013). Creating moments that matter: Strategies to combat compassion fatigue. Clinical Journal of Oncology Nursing, 17, 581–582. doi:10.1188/13.CJON.581-582
Showalter, S.E. (2010). Compassion fatigue: What is it? Why does it matter? Recognizing the symptoms, acknowledging the impact, developing tools to prevent compassion fatigue, and strengthen the professional already suffering from the effects. American Journal of Hospice and Palliative Care, 27, 239–242. doi:10.1177/1049909109354096
Stamm, B.H. (2005). The ProQOL manual. Retrieved from http://www.compassionfatigue.org/pages/ProQOLManualOct05.pdf
Stamm, B.H. (2009). Professional Quality of Life scale (PROQOL): Compassion satisfaction and compassion fatigue (PROQOL) version 5 (2009). Retrieved from http://www.proqol.org/uploads/ProQOL_5_English_Self-Score_3-2012.pdf
Stamm, B.H. (2010). The concise ProQOL manual (2nd ed.). Retrieved from http://www.proqol.org/uploads/ProQOL_Concise_2ndEd_12-2010.pdf
Wenzel, J., Shaha, M., Klimmek, R., & Krumm, S. (2011). Working through grief and loss: Oncology nurses’ perspectives on professional bereavement [Online exclusive]. Oncology Nursing Forum, 38, E272–E282. doi:10.1188/11.ONF.E272-E282
Yoder, E.A. (2010). Compassion fatigue in nurses. Applied Nursing Research, 23, 191–197. doi:10.1016/j.apnr.2008.09.003