INTRODUCTION
The term “burnout” describes the end-result of a significant process of emotional attrition, and is typically characterized by the core symptoms of physical and emotional exhaustion, a sense of depersonalization, and of diminished personal achievement (Maslach, Reference Maslach1982; Maslach & Leiter, Reference Maslach and Leiter1997).
Emotional exhaustion reflects the stress dimension of burnout, and usually drives health professionals towards depersonalization, i.e., the attempt to distance themselves emotionally from their patients, who are no longer seen as unique and engaging people. The sense of personal achievement indicates the capacity to see oneself as a competent and successful professional. Burnout typically involves high levels of emotional exhaustion and a marked sense of depersonalization, together with low levels of personal achievement (Maslach et al., Reference Maslach, Schaufeli and Leiter2001).
Various studies have provided evidence concerning the possible antecedents and consequences of burnout (Maslach et al., Reference Maslach, Schaufeli and Leiter2001) both in healthcare professionals (Papadatou et al., Reference Papadatou, Anagnostopoulos and Monos1994; Vachon, Reference Vachon1995; Rokach, Reference Rokach2005; Trufelli et al., Reference Trufelli, Bensi and Garcia2008) and in other professional domains (Byrne, Reference Byrne1994; Friedman, Reference Friedman2000). Among health workers, oncology professionals (including oncology nurses) typically show significant levels of burnout (McVicar, Reference McVicar2003; Medland et al., Reference Medland, Howard-Ruben and Whitaker2004; Sherman et al., Reference Sherman, Edwards and Simonton2006).
Different clinical settings can provide assistance to oncology patients and their families. Hospital care units traditionally aim at successfully curing oncology patients by providing them with the best technical intervention (“high tech” interventions), whereas palliative care units mainly aim at taking care of terminally ill oncology patients by means of supportive assistance and spiritual care (“high touch” interventions). The different missions of the two clinical settings imply that both the nature of the interventions provided and the characteristics of the professional–patient relationship developed differ significantly in hospital units compared to hospices. It is therefore reasonable to expect different levels of burnout symptoms in the health professionals working in the two clinical settings. However, to date, little is known about the direct comparison of the level of burnout symptoms experienced by specialist nurses working with oncology patients in hospital or palliative-care units. To the best of our knowledge, the only study directly tackling this issue (Bram & Katz, Reference Bram and Katz1989) compared two small groups of oncology nurses working in oncology units or palliative care units: interestingly, they found that working in palliative care units was associated with lower levels of burnout symptoms, despite the considerable physical and emotional effort required by working daily with terminally-ill patients, and repeatedly facing their death.
To date, although the role played by both work setting and individual personal characteristics in determining the presence of burnout symptoms is now established (Vachon, Reference Vachon1995; Maslach et al., Reference Maslach, Schaufeli and Leiter2001; Vachon & Huggard, Reference Vachon, Huggard, Ferrell and Coyle2010), the differential contributions of situational factors (e.g., clinical setting) and individual factors (e.g., demographic and personality characteristics) in triggering burnout symptoms in specialist oncology nurses still remains to be clarified. The present study therefore aimed to clarify the differential contributions of a situational factor (clinical setting) and four individual factors (work experience, anxiety, depression, and relational style) in triggering burnout symptoms in two independent groups of specialist oncology nurses working in two different settings (hospital units and hospices). In keeping with the perspective that emphasizes a major role played by situational factors in determining the level of burnout, we hypothesized that “clinical setting” would be a stronger predictor of burnout than the other factors, with nurses working in hospices showing lower levels of burnout symptoms.
METHOD
Participants
The study involved two independent groups of Italian specialist oncology nurses working in two different clinical settings: (1) The “San Luigi Gonzaga” University Hospital of Orbassano, Italy; and (2) three hospices in Piedmont, Italy.
The first group of nurses (n = 59; 51 females and 8 males, age range 24–51, mean age 34.14 ± 7.3 years) worked in six hospital units characterized by a very high prevalence of oncology patients (Internal Medicine 1, Internal Medicine 2, Oncology Day-Hospital, Pulmonary Oncology Day-Hospital, Pneumology 1, and Pneumology 2).
The second group (n = 33; 32 females and 1 male, age range 24–59, mean age 37.27 ± 9.71 years) worked in three palliative care units in Piedmont, Italy: the Busca Hospice, the “F.A.R.O. Foundation” Hospice at the San Vito Hospital, and the Lanzo Hospice.
Inclusion criteria were as follows: participants must (1) work as attending specialist nurse responsible for the care and support of oncology patients; (2) be currently employed in the hospital or in one of the hospices involved in this study; (3) not be connected with the present study. The study was granted approval by the local Research Ethics Committee. Informed consent was obtained from all participants.
Procedure
After a plenary meeting to present the aims of the study and the clinical self-report measures used, an anonymous form was handed to each of the nurses who had agreed to take part in the present study. The form requested the nurse to provide demographic and job characteristics (gender, age, clinical setting, work experience) and contained three clinical self-report measures. The completed forms could be returned by hand or by mail to the Registrar of the Psychiatric Clinic of the “San Luigi Gonzaga” Hospital (Orbassano, Piedmont, Italy).
The three self-report measures administered were:
1 The Italian version of the Maslach Burnout Inventory (MBI), (Maslach & Jackson, Reference Maslach and Jackson1981a, Reference Maslach and Jackson1981b, adapted from Sirigatti & Stefanile, Reference Sirigatti and Stefanile1993), a 22-item measure assessing the symptoms of Emotional Exhaustion (EE, 9 items), Depersonalization (DP, 5 items), and reduced Personal Achievement (PA, 8 items). The Italian version of the MBI differs from the original as there is only one subscale for “feeling frequency,” and “intensity” is not tested, because of the high correlation between the two dimensions (Brusaferro et al., Reference Brusaferro, Agnoletto and Gubian2000). Each item is rated on a 7-point Likert scale ranging from 0 (never) to 6 (every day). Three separate subscales specifically measure EE, DP, and PA. Burnout is not conceptualized as a dichotomous variable (either present or absent), but as a continuous variable that can be quantified and that indicates the presence of low, moderate, or high levels of the symptoms experienced by the professional.
2 The Italian version of the Hospital Anxiety and Depression Scale (HADS) (Zigmond & Snaith, Reference Zigmond and Snaith1983, adapted from Costantini et al., Reference Costantini, Musso and Viterbori1999), a 14-item self-assessment scale that provides a valid and reliable measure of severity of two emotional disorders, anxiety and depression. Each item is rated on a 4-point scale ranging from 0 to 3.
3 The Italian version of the Attachment Style Questionnaire (ASQ) (Feeney et al., Reference Feeney, Noller, Hanrahan, Sperling and Berman1994, adapted from Fossati et al., Reference Fossati, Feeney and Donati2003), a 40-item self-report questionnaire designed to investigate adult relational style. The 40 items are subdivided into five scales: Confidence (8 items), Discomfort with Closeness (10 items), Need for Approval (7 items), Preoccupation with Relationships (8 items), and Relationships as Secondary (7 items). Each item is rated on a 6-point Likert scale ranging from 1 (totally disagree) to 6 (totally agree). The higher the score associated with a scale, the more evident the presence of that specific dimension (i.e. a high score on the Confidence scale implies a high level of confidence).
Statistical Analyses
Statistical analyses were performed using SPSS© (Statistical Package for the Social Sciences) version 15.0 for Windows©. Since graphical and statistical exploration of the data by means of box plots, histograms, Q-Q plots, and normality tests indicated an acceptable distribution, parametric tests were used. Analyses were run as follows: the demographic and clinical variables of the two groups of nurses were compared using unpaired t-tests. Second, burnout symptoms in the two groups of nurses were compared using unpaired t-tests, followed by a frequency distribution of participants' MBI subscale scores, according to the Italian classification. Lastly, multivariate regression analyses were run, in order to clarify the differential contribution of the predictors of interest to the outcome measures.
A p value < 0.05 was accepted as statistically significant.
RESULTS
The overall response rate was 92/112 (82.1%): 59/77 (76.6%) of hospital nurses responded, whereas 33/35 (94.3%) of hospice nurses responded. We were therefore able to reach the majority of the specialist nurses working in the clinical settings of interest. It is still important to mention that the response rate in our study was higher than those reported by other investigations (Brusaferro et al., Reference Brusaferro, Agnoletto and Gubian2000; Dorz et al., Reference Dorz, Novara and Sica2003).
The demographic and clinical characteristics of the two groups of participants are shown in Table 1, as well as their statistical comparisons via a series of t-tests. The two groups of specialist nurses (working in hospital units or hospices) were well matched for age and work experience, as well as for their level of anxiety and depression, as measured by the HADS. As far as the ASQ is concerned, two nurses working in the hospital units did not fill in this measure. The two groups of specialist nurses only differed significantly on the “Confidence” and “Relationships as Secondary” subscales, whereas the other three subscales (“Discomfort with Closeness,” “Need for Approval” and “Preoccupation with Relationships”) did not differ significantly between the two groups.
Table 1. Demographic characteristics, work experience and clinical assessment of participants
*p < 0.05; **p < 0.01 Discomfort = Discomfort with Closeness; NA = Need for Approval; PR = Preoccupation with Relationships; RS = Relationships as Secondary.
As far as the MBI is concerned, five nurses working in the hospital units did not fill in the measure. The two groups of specialist nurses significantly differed on all of the subscales (EE, DP, PA), with hospital-based nurses showing a higher level of EE and DP and a lower level of PA, than nurses working in hospices (Table 2).
Table 2. Comparison of burnout symptoms in the two groups
*p<0.001.
As burnout is conceptualized as a continuous variable indicating the presence of low, moderate, or high levels of symptoms, we report the classification of the MBI validated values in the Italian version of the tool (Table 3) and the distribution of the participants according to this classification (Table 4). Altogether, these results show the presence of significantly higher levels of burnout symptoms in the group of hospital-based nurses than in the group of hospice-based nurses.
Table 3. Classification of the MBI subscale values - Italian version (adapted from Brusaferro et al., 2000, p. 19)
Table 4. Distribution of the participants, according to the classification of MBI subscales
Furthermore, in order to clarify the differential contribution of the predictors of interest, multivariate regression analyses were run. The contribution of each individual predictor to the MBI subscale scores is shown in Table 5. The three statistical models yield a satisfactory proportion of variance explained, as shown by the adjusted R 2 associated with the models (0.419, 0.341, 0.229 for the EE, DP and PA subscales, respectively).
Table 5. Predictors of MBI subscales scores: Multivariate regression analyses (N = 87)
β = Standardized coefficients.
*p<0.05; **p < 0.01.
aCoded as: 0 = Hospital units; 1 = Hospice.
bHigher scores indicate greater discomfort.
cHigher scores indicate greater preoccupation with relationships.
According to the experimental hypothesis, the institutional factor (clinical setting) strongly emerged as a significant factor influencing all of the MBI subscale scores. In addition, the personal factor “Depression” significantly influenced the EE and DP subscales; the personal factor “Discomfort with Closeness” significantly influenced the EE subscale, whereas the personal factor “Preoccupation with Relationships” significantly influenced the DP subscale. An analysis of bivariate correlations among the significant predictors identified showed the presence of correlations ranging from very low to moderate (0.037 < r < 0.349), allowing us to rule out the presence of any marked overlap between them (data not shown). None of the other factors considered showed any significant influence on the MBI subscale scores.
DISCUSSION
In the present study, we investigated the differential contribution of institutional and situational factors in triggering burnout symptoms in oncology nurses working either in hospital units with a very high prevalence of oncology patients or in hospices. In keeping with the perspective that emphasizes a major role played by situational factors in determining the level of burnout, we hypothesized that the clinical setting in which nurses worked would have been a stronger predictor of burnout than the others, and that nurses working in the hospices would have shown a lower level of burnout symptoms.
Interestingly, in our study we were able to reach the vast majority of the specialist oncology nurses working in the two clinical settings of interests (global response rate of 82.1%), making it possible for us to be confident about the representativeness of the experimental samples involved. The two groups of nurses (working in hospital units or hospices) were well matched for demographic characteristics. In order to rule out the possibility that different levels of burnout in the two groups were actually triggered by different levels of anxiety and depression, we compared the two groups and showed the absence of a significant difference between the two groups in terms of anxiety and depression, as measured by the HADS. Regarding this, it is still important to note that the levels of both anxiety and depression in the two groups were well below the clinical borderline range of values, allowing us to exclude the presence of significant anxiety or depression in our samples on both statistical and clinical grounds. In terms of their relational style as measured by the ASQ, the two groups of nurses only differed significantly on the “Confidence” and “Relationships as Secondary” subscales, with nurses working in hospices showing a higher score on the former and a lower score on the latter, compared to nurses working in hospital units. The other three subscales (“Discomfort with Closeness,” “Need for Approval,” and “Preoccupation with Relationships”) did not differ significantly between groups.
Regarding the presence of burnout symptoms as measured by the MBI, as expected the two groups of specialist nurses significantly differed on all of the subscales scores (EE, DP, PA), with nurses working in hospices showing a lower level of EE and DP and a higher level of PA, compared to nurses working in hospital units. Furthermore, an analysis of the frequency distribution of the level of experienced burnout symptoms clearly showed that the majority of nurses working in hospital units presented with moderate to high levels of symptoms, whereas the vast majority of nurses working in hospices presented with low levels of symptoms. In order to provide further evidence supporting the significant contribution made by the factor “clinical setting” in determining burnout symptoms, multivariate regression analyses were applied and showed that it emerged as a significant factor influencing all of the MBI subscale scores. Therefore, working in hospices appeared to act as a protective factor, reducing the scores of EE and DP and increasing the score of PA.
Although the ASQ scales termed “Confidence” and “Relationships as Secondary” originally differed between the two groups of nurses, the multivariate regression models did not encompass them as significant predictors of burnout symptoms, indicating that they appear not to have played a significant role in determining the presence of burnout symptoms. Three other individual factors contributed instead to burnout symptoms: “Depression” significantly contributed to increasing the score of EE and reducing the score of PA; “Discomfort with Closeness” significantly contributed to increasing the score of EE; whereas “Preoccupation with Relationships” significantly contributed to increasing the score of DP.
According to the experimental hypothesis, our findings therefore support the view that the clinical setting in which nurses work is a strong predictor of experienced burnout symptoms, showing that specialist oncology nurses working in the demanding clinical setting of a hospice present with relatively low levels of burnout symptoms, compared to their colleagues working in hospital oncology units. Even if at this time the precise identification of the factors underlying this significant difference is not yet possible, it could be argued that assuming a perspective focused on taking care of terminally ill oncology patients (as nurses working in hospices do) instead of aiming at healing them could significantly help to reduce the gap between what should ideally be achieved (i.e. healing), and what it is actually possible to do (i.e. providing care). This might contribute to reducing the level of personal and professional frustration and to limiting the presence of significant experienced burnout symptoms among nurses working in hospices. Conversely, assuming a perspective focused on providing a successful cure (as nurses working in hospital units do) could significantly increase the gap between ideal and possible goals, therefore contributing to increasing their level of frustration and their experienced burnout symptoms. However, future studies should specifically address this point, in order to verify the plausibility of our suggestion.
The present study has some limitations. First, the sample size was not very large. Although we were able to reach the vast majority of nurses working in the clinical settings of interest, future studies should involve larger numbers of participants in order to provide more robust evidence. Furthermore, we did not investigate the role of other factors (e.g., marital status, children, leisure activities) that have been recognized as possible contributors to burnout symptoms.
CONCLUSION
As suggested by previous research and corroborated by the present study, the clinical setting in which specialist oncology nurses work plays a key role in triggering burnout symptoms. Our findings help to clarify the differential contribution of institutional and individual factors to burnout symptoms, and therefore our results fully support demands for programs and interventions focusing not only on individual characteristics but also on institutional aspects, in order to contain the experienced burnout symptoms among specialist oncology nurses.
