Professional quality of life, sleep disturbance and health among nurses: A mediation analysis

Abstract Aims This study aimed to examine sleep disturbance as a mediator of the relationship between professional quality of life (compassion satisfaction, burnout, secondary traumatic stress) and health (physical and mental health) in nurses. Design Descriptive, cross‐sectional study. Methods Three hundred eighteen Registered Nurses completed a web‐based survey at the National Institutes of Health Clinical Center in the United States. Mediation analyses were conducted to test hypothesized relationships. Results Nurses with higher levels of compassion satisfaction reported lower levels of sleep disturbance and better physical/mental health. Burnout and secondary traumatic stress were negatively associated with physical/mental health and positively associated with sleep disturbance. Sleep disturbance fully or partially mediated the relationships between professional quality of life and physical/mental health among nurses.


| INTRODUC TI ON
Working as a nurse has many rewards, but the physical demands, irregular schedules, long work hours, and regular exposure to suffering and death can influence nurses' professional quality of life (Flarity et al., 2013;Hinderer et al., 2014;Khamisa et al., 2015;Kim et al., 2015). Professional quality of life has emerged as a growing issue of interest among healthcare professionals (Kim et al., 2015).
Professional quality of life refers to the positive (compassion satisfaction) and negative (compassion fatigue) aspects associated with working as a professional provider of care. Compassion satisfaction is the positive feeling gained from helping or caring for others and performing work-related tasks well. Compassion fatigue refers to the negative consequences of working in a helping profession and is comprised of two aspects, burnout and secondary traumatic stress. Burnout is defined as the negative emotional reaction to external stressors associated within one's work environment (e.g., long working hours or heavy workloads). Secondary traumatic stress refers to the negative emotions and behaviours associated with exposure to work-related traumatic stressful events (Stamm, 2010). Professional quality of life not only can impact nurses' job performance (Stamm, 2010), but it can also directly influence their physical and mental health (Eanes, 2015;Fu et al., 2018;Khamisa et al., 2015). The relationship between workplace stress in nurses and sleep disturbances is well documented, and impaired sleep is considered a critically important issue facing nurses today (Caruso et al., 2019). There is good evidence that disturbed sleep contributes to significant physical and mental health issues for nurses (Caruso et al., 2019;Eanes, 2015;Kunzweiler et al., 2016;Perry et al., 2015;Sun et al., 2019). Nurses who are in poor physical and mental health as a result of workplace stress and workplacerelated sleep disturbances may represent a threat to patient safety (Caruso et al., 2019), as health problems in nurses can influence work efficiency and contribute to medical errors (Eanes, 2015;Zhang et al., 2017). Yet to date, few studies have examined the interrelationships between professional quality of life, sleep disturbance and health in nurses. Therefore, the aim of this study is to examine the mediating role of sleep disturbance on the relationship between professional quality of life and physical and mental health in nurses.

| Professional quality of life and physical and mental health in nurses
Healthcare professionals in general are known to experience burnout and secondary traumatic stress, but nurses are particularly at risk (Kim et al., 2015). Numerous studies have examined the incidence of compassion satisfaction, burnout and secondary traumatic stress in nurses (Yilmaz & Üstün, 2018), but far fewer have examined the relationship between professional quality of life with physical and mental health in nurses. While compassion satisfaction appears to be protective of physical and mental health among nurses (Eanes, 2015;Fu et al., 2018;Khamisa et al., 2015), compassion fatigue is associated with declines in nurses' health and quality of life (Fu et al., 2018). Nurses with high levels of compassion fatigue reported experiencing a variety of physical and psychological symptoms including headaches, gastric disturbances, depressive symptoms, anxiety and sleep disturbances (Fu et al., 2018;Khamisa et al., 2015). Khamisa et al. (2015) examined the relationships between work-related stress, burnout, and job satisfaction with physical and mental health in 895 South African nurses and found that burnout was associated with deteriorations physical and mental health; in particular, burnout was the most influential factor predicting anxiety and insomnia.

| Mediating role of sleep disturbance in the professional quality of life on health
The relationship between sleep disturbance and poor health outcomes is well established. Sleep disturbance contributes to not only a variety of chronic disease such as obesity, diabetes, cardiovascular disease and certain cancers (Dong et al., 2017;Eanes, 2015;Khormizi et al., 2018;Kunzweiler et al., 2016;Sun et al., 2019), but also to mental health problems such as depression (Eanes, 2015;Perry et al., 2015;Sun et al., 2019;Zhang et al., 2017). Nurses may be less likely than the general population to receive an adequate amount of quality sleep (Caruso, 2014;Eanes, 2015;Geiger-Brown et al., 2012), likely because the nursing workplace often involves shiftwork. Shiftwork, particularly working nights or rotating shifts, is associated with circadian rhythm disruption that often leads to sleep deprivation, fatigue and diminished attention (Caruso, 2014;Imes & Chasens, 2019;Niu et al., 2011 rather, they tend to interact with other factors and conditions in multiple ways (Vitaliano et al., 2003). However, most research to date involving the health of nurses has focused on singular factors such as long work hours or burnout (Gómez-García et al., 2016;Stimpfel et al., 2012).

| Conceptual framework
This study is guided by the theoretical framework of Punnett et al. (2009), which suggests that the relationship between working conditions and employee health is multifaceted, with health behaviours such as sleep playing a role in the association between working conditions and health among employees ( Figure 1). For example, negative working conditions such as exposure to workplace F I G U R E 1 Theoretical model of sleep disturbance as the mediator between professional quality of life and health of nurses. Note: The dependent variable results from the direct effect of the independent variable (path c) as well as the mediating variable (path b). Variable mediating the independent variable can also be involved (path a) stress may contribute directly to an employee's physical and mental health. Workplace stress can also contribute to negative health behaviours such as poor sleep, which in turn may contribute to negative changes in physical and mental health. Likewise, positive working conditions, such as those whereby individuals have high levels of control and support, can have a protective effect on an individual's physical and mental health (Caruso et al., 2019;Punnett et al., 2009). However, these proposed pathways have not yet been

| Study design, participants and setting
An anonymous cross-sectional survey design was used to examine nurses' participation in health-promoting self-care activities.

| Sampling and survey procedure
The online survey was created using SurveyMonkey ® and delivered via email. The participants received a total of three emails during a 3week window. The initial email contained a letter from the Principal Investigator (PI) with a description of the study and a link to an anonymous online survey. One week after the first email was delivered, a reminder email was sent, followed a week later by a third final reminder and thank you. Announcements were also made at Nursing Practice Council, Leadership meetings and unit and clinic-based staff meetings, where large groups of nurses routinely gathered.

| Professional quality of life
The Professional Quality of Life Scale (ProQOL) (Stamm, 2010) measures general job satisfaction within the timeframe of the last 30 days. The 30-item ProQOL instrument includes three subscales: compassion satisfaction, burnout and STS. Each subscale consists of 10 questions, with each item on a 5-point Likert scale, ranging from 1 (never) to 5 (very often). Each of the three subscales is scored by summing the 10 items, after appropriately reverse scoring is completed.

| Physical and mental health
Physical and mental health were collected using the Patient-

Reported Outcomes Measurement Information System (PROMIS ® )
Global Health scale v1.0/1.1, which has exhibited good internal consistency reliability and strong construct validity across populations (Cella et al., 2010;Hays et al., 2009;Katzan & Lapin, 2018). The instrument consists of ten global health items that are used to calculate two component scores: global physical health and global mental health. Individual items for global physical health and global mental health are scored using a 5-point Likert scale that ranges from 1 to 5, with higher scores indicative of better physical and mental health.
The measures are standardized to a T-score metric, with a mean of 50 and standard deviation of 10 that is centred around the general  (Hays et al., 2015;HealthMeasures, 2020). In this study, the physical and mental health scales had Cronbach alpha coefficients of 0.704 and 0.850, respectively.

TA B L E 1 (Continued)
psychometric properties for the PROMIS ® sleep disturbance have been shown to be acceptable in nurses (Imes & Chasens, 2019;Kemper et al., 2015). In this study, the Cronbach alpha was 0.839.

| Statistical analysis
The descriptive statistics (mean and frequencies) of the demographics and workplace characteristics were calculated to describe the sample. The assumptions (e.g., normality) of outcomes and the potential mediators were assessed prior to modeling. The Harman's single-factor test was conducted to detect common method bias (Podsakoff et al., 2003). The results indicated that the common method variance was not a serious threat in this study. The correlations among ProQOL, sleep disturbance and health were computed.
The mediation effect with path analysis was conducted to assess the effects of ProQOL on nurses' health and how the effects were mediated by sleep disturbance. Based on existing evidence, age and gender were identified a priori as covariates and were controlled for in all the models. Multiple fit indices were used to determine whether the model was fit to the data: RMSEA < 0.08, CFI ≥ 0.95 and SRMR ≤ 0.08 (Hooper et al., 2008). The significance of the mediation effects was assessed using the non-parametric, bias-corrected 95% bootstrapped confidence interval (BCI) with 5,000 bootstrap replications (Mackinnon et al., 2004;Preacher & Hayes, 2008). The indirect effect is considered as significant if the zero is not included in the 95% BCI. All mediation analyses were conducted in Mplus Version 7.2 (Muthén andMuthén, (1998-2012)). Missing data in all models were estimated with the full information maximum likelihood (FIML) method used by Mplus. A p < .05 was considered significant.

| Ethics
The NIH Clinical Center Office of Human Subjects Research Protections approved this study (OHSR #13263). Informed consent was implied if participants completed the online survey.

| Correlations between ProQOL, sleep disturbance, and health
A number of statistically significant correlations were identified in the preliminary analysis among ProQOL, sleep disturbance, and physical and mental health (Table 2).

| Model 1a
The total effect of compassion satisfaction on physical health was was not statistically significant after adjusting for sleep disturbance.
Thus, the association between compassion satisfaction and physical health was fully mediated by sleep disturbance.

| Model 1b
The total effect of burnout on physical health was statistically significant (β = −0.370, p < .001). The direct effect of burnout

| Model 1c
The total effect of secondary traumatic stress on physical health was The results indicated that sleep disturbance partially mediated the relationship between secondary traumatic stress and physical health.

| Associations between ProQOL, sleep disturbance and mental health
Results from the mediation analyses that were conducted to identify direct and indirect effects of ProQOL, compassion satisfaction (Model 2a), burnout (Model 2b), secondary traumatic stress (Model 2c) and sleep disturbance on nurses' mental health are shown in Table 4.

| Model 2a
The

| Model 2b
The total effect of burnout on mental health was statistically significant (β = −0.629, p < .001). Burnout was directly associated with the mediating variable of sleep disturbance (β = 0.362, p < .001). Sleep disturbance was significantly negatively associated with mental health (β = −0.244, p < .001). In addition, burnout was directly associated with mental health (β = −0.540, p < .001). The indirect pathways between burnout and mental health through sleep disturbance was statistically significant (β = −0.088, BCI = −0.110, −0.044). The results indicated that sleep disturbance partially mediated the relationship between burnout and mental health.

| Model 2c
The total effect of secondary traumatic stress on mental health was

| D ISCUSS I ON
This study aimed to investigate the direct impact of professional quality of life on sleep disturbance and health, and the mediating role of sleep disturbance in the direct relationship among nurses.
Our study findings revealed that nurses experiencing higher levels of compassion satisfaction were more likely to have better physical and mental health, while nurses with higher levels of burnout and secondary traumatic stress were more likely to experience poor physical and mental health. The association between professional quality of life and health in this population of nurses is consistent with previous research (Fu et al., 2018;Khamisa et al., 2015). Like the findings of Fu and colleagues (2015), all three dimensions of professional quality of life (compassion satisfaction, burnout, secondary traumatic stress) were significantly associated with physical and mental health in nurses. Like Khamisa et al. (2015), burnout was associated with worsened physical and mental health and insomnia, although our study went a step further and found that burnout, and most other aspects of professional quality of life, contribute both directly to physical and mental health, but also indirectly to physical and mental health via worsening sleep disturbances. The findings of this study may indicate that enhancing compassion satisfaction and reducing burnout and secondary traumatic stress may be particularly beneficial in improving the physical and mental health of nurses, especially those nurses who are experiencing sleep disturbances.
As anticipated, nurses in this study who reported sleep disturbances were more likely to exhibit poor physical and mental health.
The findings of this study thereby confirm previous studies that established an association between sleep disturbance and health among nurses (Dong et al., 2017;Eanes, 2015;Khormizi et al., 2018;Kunzweiler et al., 2016;Perry et al., 2015;Sun et al., 2019;Zhang et al., 2017). More importantly, the present study adds new evi-

| Limitations/future directions
This study provides innovative information about the mediating role of sleep disturbance in the association between professional quality of life and health in nurses, but some limitations should be considered when interpreting the findings. The primary limitation was the use of a cross-sectional design, which cannot provide information about causal relationships among variables. Studies utilizing longitudinal designs are needed in order to confirm causal relationships among professional quality of life, sleep disturbance, and physical and mental health of nurses. In addition, this study only recruited nurses from the NIH Clinical Center, a unique research hospital with a highly educated, all-RN nursing workforce. Thus, the findings of this study may not be generalizable to other nursing populations.
In the future, studies are needed that recruit nurses with different levels of education and who work in diverse medical settings in order to determine whether these findings can be replicated with other populations.

| CON CLUS ION
This study used a mediation analysis to contribute additional knowledge regarding the complexity of the interrelationships between professional quality of life, sleep disturbance and health in nurses.
This study confirmed the importance of professional quality of life to nurses' health and well-being that had been found in other studies, and the relatively large sample size allowed for the novel inclusion of a mediator, sleep disturbance, as an additional pathway whereby compassion satisfaction, burnout and secondary traumatic stress influence the health of nurses. The findings support Punnett's conceptual model (2009), by confirming that there is a direct pathway between nurses' professional quality of life and their physical and mental health, and that is pathway is mediated by sleep disturbance.
It is important for healthcare organizations to protect the health of nurses by developing and supporting programmes that promote healthy sleep and improve nurses' professional quality of life. Future research is needed to further understand the behavioural mechanisms for these relationships, explore other novel mediators and to evaluate the effects of interventions that improve professional quality of life, sleep quality and health among nurses.

CO N FLI C T O F I NTE R E S T
The author(s) declare that they have no conflict of interest.

E TH I C A L A PPROVA L
The Office of Human Subjects Research Protections approved this study at the National Institutes of Health, Clinical Center (OHSR#13263).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.