Introduction
Since the first known case of coronavirus disease 2019 (COVID-19) was reported in Wuhan, China, in December 2019, the disease has spread rapidly across the globe (Wang et al., 2020). In a race against time to combat the virus, the Chinese Health Authority dispatched more than 28,000 nurses from all over the country to Hubei Province to tackle the outbreak (State Council of the People's Republic of China, 2021; Zhou et al., 2020). Nurses have been at the forefront of this healthcare crisis (Fernandez et al., 2020). Facing the suddenness of this pandemic and the shortages of sufficient manpower and personal protective equipment, nurses have experienced unprecedented work-related stress and sleep problems (Tsay et al., 2020; Yang et al., 2020). In related studies conducted in China during the COVID-19 pandemic, the incidence of sleep disorders among first-line nurses has been reported to be as high as 52.8%-64.15% (J. J. Wu et al., 2020; Zhang et al., 2020).
Sleep is a necessary and positive cyclical process that involves a variety of physiological and psychological phenomena (Fang et al., 2019), and research has been conducted to investigate the effects of disturbed sleep on health and medical safety (Lin et al., 2018). During the pandemic, poor sleep may not only suppress the immune system of nurses and increase their vulnerability to inflammation and infection (Ishikura et al., 2021; Khan & Aouad, 2017) but also threaten patient safety and quality of care, increasing the rates of misdiagnoses and adverse drug events (Gander et al., 2019). Therefore, good sleep is vital to maintaining the ability of nurses to care for patients with infectious diseases. It is highly important to study the mechanisms affecting sleep quality in nurses during epidemic and pandemic events.
The findings of a substantial body of studies suggest that work-related stressors increase the risk of sleep disorders in healthcare professionals (X. Deng et al., 2020; W. Lu et al., 2020), among which role stressors have been identified as one of the most critical (Alfes et al., 2018). However, previous studies of role stressors have focused primarily on role conflict and role ambiguity and ignored the negative effects of role overload (RO; Olivares-Faundez et al., 2014).
RO is defined as a role stressor perceived by an individual who has too little time or energy to meet role expectations (Morter, 2010). RO is often conceptualized as the antecedent of negative psychological stress experience (Iwasaki et al., 2018; Shupe et al., 2015; Taouk et al., 2018). During the pandemic, nursing work has become more complicated and demanding, with nurses expected to assume an increasing number of roles. Thus, RO has become an additional stressor for healthcare professionals during the pandemic response (Shacham et al., 2020). Although studies have found RO to have various, adverse effects on physical and mental health (Andrews & Kacmar, 2014; Duxbury et al., 2014), there remains a paucity of research on the correlation between RO and sleep quality, especially among Chinese first-line nurses in the context of the current pandemic.
To prevent the negative impact of RO on nurses, it is important to formulate effective coping strategies to manage role stressors. Mindfulness is the awareness that emerges through focusing attention on purpose nonjudgmentally in the present moment and on the unfolding of experience from moment to moment (Bishop et al., 2010). Studies have found that mindfulness also targets multiple cognitive and emotional processes that decrease ruminative thoughts, diminish emotional reactivity, and promote the impartial reappraisal of salient experiences, which together may facilitate sleep (Charbonneau, 2019; Ong et al., 2012). Considering the prevalence of RO and the severity of sleep disorders during the pandemic, this study was designed to explore whether mindfulness may also alleviate role-overload-related sleep disturbance. To our knowledge, scholarly research into the relationship among mindfulness, RO, and sleep disorders is still in its early stage and is worth further study.
Therefore, the primary aims of this study were to examine the association between RO and sleep quality in first-line nurses involved in combating the COVID-19 pandemic in Hubei Province and to explore the moderating effects of mindfulness.
Methods
Study Design and Participants
A cross-sectional survey design was adopted in this study, which was conducted from March 20 to April 5, 2020. The participants were recruited from five medical teams who had traveled from Fujian Province to the epicenter of the outbreak in China (Wuhan City) to treat patients with COVID-19. The inclusion criteria were as follows: (a) directly participating in first-line clinical nursing work and (b) willing to participate. Otherwise qualified nurses were excluded if they were primarily engaged in nursing management or logistics. Three hundred fifty-seven individuals were invited to participate. All of the participants participated voluntarily without compensation or reward and completed an anonymous questionnaire. Participants were informed about the study aims and importance. The research design was approved by the ethics committee of Fujian Provincial Hospital (K2020-03-01).
Sampling
The recommended sample size was calculated as 348 based on the formula: N = (Z[alpha]/2)2 x p x (1 - p)/d2. At a 95% CI of Z[alpha]/2 = 1.96, the estimated acceptable margin of error was d = 0.05 and the proportion with sleep disturbance (p) was 34.8%. This proportion was based on a systematic review and meta-analysis of sleep disturbances among nurses responsible for providing care to patients with COVID-19 in which the prevalence of sleep disturbance was 34.8% (Salari et al., 2020). The analysis portion of this study used data from 357 participants.
Data Collection
After contacting the nursing leaders of the five medical teams to introduce the purpose of the study and policy on data anonymity, a cross-sectional survey design involving a web-based questionnaire was employed to collect the data. A website link and a quick response code for the questionnaire were sent to the nursing leaders, who were asked to forward them to the nurses. The survey was completed during the medical staffs' rest period within a 2-week period (late March to early April). The questionnaire required 10-15 minutes to complete on a computer or smartphone. After retrieving the questionnaire results and excluding questionnaires with illogical answers as invalid, data in the 357 valid questionnaires (effective response rate: 92%) were collected for analysis.
Measures
Sleep quality
Sleep quality was assessed using the Pittsburgh Sleep Quality Index (PSQI), which is an instrument that measures individuals' subjective sleep quality over the prior month. This scale was developed by Buysse et al. (1989) and later translated into a Chinese version by Liu et al. (1996). The scale contains 19 items and the following seven factors: subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medications, and daytime dysfunction. Each factor is scored between 0 and 3, with >= 2 interpreted as "positive," meaning that the respondent's sleep quality for the associated factor is poor. The sum of the seven factor scores is the total PSQI score, with a total possible range of 0-21. Higher total scores are associated with worse sleep quality, with scores > 7 indicating impaired sleep. The scale was shown to have good reliability, with a Cronbach's [alpha] coefficient of .842 (T. Y. Lu et al., 2014).
Mindfulness
Mindfulness was assessed using the Five-Facet Mindfulness Questionnaire (FFMQ). This scale was compiled by Baer et al. (2006) in 2006 and translated into Chinese by Y. Q. Deng et al. (2011). The 39-item scale includes the following five subscales: observing, describing, acting with awareness, nonjudging, and nonreactivity. Participants rate each item using a 5-point Likert scale ranging from 1 = never or very rarely true to 5 = very often or always true, with higher scores associated with higher levels of mindfulness. In this study, the Cronbach's [alpha] coefficients for the five subscales were as follows: .91 (nonreactivity), .92 (observing), .91 (acting with awareness), .91 (nonjudging), and .94 (describing).
Role overload
RO was assessed using the Role Overload Scale (ROS), which was first revised by Peterson et al. (1995) based on conditions across 21 countries. The ROS is a mature and widely used tool for measuring RO in China. This five-item scale scores each item using a 5-point Likert-type scale (1 = strongly disagree to 5 = strongly agree). Items include "Too many things I need to do at work" and "Too many things my superior expects me to accomplish." The ROS has been shown to have good reliability, with a Cronbach's [alpha] coefficient of .891.
Covariates
On the basis of a prior literature review and expert consultation (Kim-Godwin et al., 2021; Sagherian et al., 2020; Zhang et al., 2020), the demographic covariates measured in this study included gender (male = 1, female = 2), years in practice (< 5 = 1, 5-10 = 2, > 10 = 3), marital status (single = 1, married = 2), age, department, monthly income, professional title, and educational level.
Statistical Analysis
IBM SPSS Version 24.0 (IBM Inc., Armonk, NY, USA) and the PROCESS Procedure Version 3.5 in SPSS were used for statistical analyses. Continuous variables were expressed as mean and standard deviation, and categorical variables were expressed as frequency and percentage. Confirmatory factor analysis was used to explore the validity and reliability of FFMQ, ROS, and PSQI. Correlation analyses were conducted to assess relationships among the variables.
Hierarchical regression was used to analyze the data. The variable centralization method was used to construct the interaction terms between independent variables and moderating variables and, respectively, control variables, independent variables, moderating variables. These interaction terms between the independent variables and the moderating variables were subsequently introduced into the regression equation.
Results
Confirmatory Factor Analysis and the Common Method Bias Test
Confirmatory factor analysis was used to test the validity of FFMQ. The results showed that the fitting indexes were up to or close to the standard, with [chi]2 = 2104, df = 692, [chi]2/df = 3.04, root mean square error of approximation = .079 < .08, incremental fit index = 0.85, Tucker-Lewis index = .84, and comparative fit index = .85. From the perspective of confirmatory factor analysis, the validity of the mindfulness questionnaire was good.
Harman's single-factor test is a well-known test to assess common method bias. By conducting an unrotated exploratory factor analysis of all the items involved in the variables in this study, we found that the first factor explained less than half (36.61%) of the variation (Zhang et al., 2020). Therefore, the common method bias effect was nonexistent/insignificant in this study.
Descriptive Characteristics
The descriptive characteristics of the demographic variables are presented in Table 1. The 357 nurses who participated in this study had a mean age of 31.15 years (SD = 5.57). Most were women (86.0%, n = 307), had bachelor's degree or higher (72.3%, n = 258), and had worked as a nurse for 5-10 years (53.0%, n = 189).
Correlations Among Mindfulness, Role Overload, and Sleep Quality
The correlations among reliability, validity, and bivariate correlations and RO, mindfulness, and sleep quality are presented in Table 2. The Cronbach's alpha coefficients of the variables were all > .80, average variance extracted was above .50, and composite reliability was above .70, indicating that these latent variables had good internal consistency, convergence validity, and composite reliability, respectively.
Sleep quality was significantly and positively correlated with RO (r = .44, p < .001) and negatively correlated with the five factors of mindfulness, in which the correlation coefficients between sleep quality and observing, describing, nonjudging, nonreactivity, and acting with awareness were - .28, - .32, - .30, - .42, and - .16 (p < .001), respectively. Furthermore, RO was found to be negatively correlated with the five factors of mindfulness, in which the correlation coefficients between RO and observing, describing, nonjudging, nonreactivity, and acting with awareness were - .32, - .30, - .26, - .20, and - .16 (p < .001), respectively. In brief, RO was negatively correlated with mindfulness and positively correlated with poor sleep quality, whereas mindfulness was negatively correlated with poor sleep quality (p < .05).
Moderating Effect of Mindfulness on the Relationship Between Role Overload and Sleep Quality
The study examined the moderating effect of mindfulness on the relationship between RO and sleep quality. In the first step, the control variables were added to obtain Model 1. In the second step, independent variable RO was added to obtain Model 2. With the addition of the independent variable, the effect was shown to be significant. In the third step, Model 3 was obtained by adding the moderating variable of mindfulness level. In the fourth step, Model 4 was obtained by adding interaction terms between the independent and moderating variables. The results are shown in Table 3.
As shown in Table 3, RO affected sleep quality positively ([beta] = 0.45, p < .01). The interaction between RO and mindfulness level had a significantly negative effect on sleep quality (B = - 1.08, p = .006). These results suggest that mindfulness has a negative moderation effect on the relationship between RO and poor sleep quality; that is, as mindfulness increases, the positive effect of RO on poor sleep quality weakens.
To more vividly show the moderating effect of mindfulness level on the relationship between RO and sleep disorders among nurses, this article drew a moderating effect graph based on the method of Aiken and West (1991), with the results shown in Figure 1. As illustrated in Figure 1, low levels of mindfulness resulted in a relatively strong impact of RO on sleep quality, whereas it was high when the levels led to a significantly lower impact of RO on sleep quality. Moreover, this positive relationship steadily weakened as level of mindfulness increased. The incidence of sleep disorder was lower in participants with low ROs, whereas in participants with high ROs, higher mindfulness levels were associated with lower levels of sleep disorder severity.
Discussion
The results of this study showed RO as having a significant and positive impact on sleep quality, with RO severity negatively associated with sleep quality. This is similar to the results of Iwasaki et al. (2018). The unexpected outbreak of COVID-19 has greatly disrupted nurses' pace of life and work. Many nurses left their familiar working environment to serve at the epicenter of the outbreak in China. Because of the lack of human and material resources and the implementation of new policies and procedures (Y. Wu et al., 2020), nurses are required to perform various nonnursing roles such as infection prevention, control and delivery of supplies, and cleaning. When nurses lack the necessary time, energy, knowledge, or skills, they try to reallocate or redistribute their resources to reduce stress and achieve role balance. If this fails, the pressure generated by the RO can impair health (Taouk et al., 2018). Nurses with high RO often exhibit states such as emotional disconnection and incompetence (Fisher, 2014), which may eventually result in severe insomnia (Nixon et al., 2011). Therefore, the results of this study showed RO, as a typical role stressor, to be positively associated with sleep disturbance.
The results also found that mindfulness may help alleviate sleep disorders caused by RO, indicating that mindfulness has a significant moderating effect on sleep quality. When faced with the same degree of RO, nurses who show high levels of mindfulness may express RO in less harmful ways than those with low mindfulness levels. This result is consistent with a previous study that identified the value of mindfulness in coping with role stressors by reducing the detrimental effects on emotions (Long & Christian, 2015). In addition, a previous study found the mindfulness intervention to have short-term and medium-term positive effects on compassion satisfaction, secondary traumatic stress, and mental health in nurses (Fu et al., 2021).
On the basis of the Job Demand-Resources model, mindfulness is conceptualized as a personal resource that alters the stress process, which reduces the negative appraisals of stress in favor of more-adaptive cognitive strategies (Grover et al., 2017; Guidetti et al., 2019). This study proved that nurses with a high level of mindfulness are able to correctly assess their own experience and their poor work environment and to not make negative evaluations of the RO caused by the pandemic. They often devote attention to current nursing activities and reasonable work demand and maintain a state of acceptance.
The results of this study have several practical implications. On the basis of the findings, to reduce the impact of RO on sleep quality in nurses, strategies at two levels may be pursued.
First, at the organizational level, scientific and effective RO management is necessary. During pandemics, nursing managers should ensure the vertical uniformity and horizontal boundaries of orders to avoid conflicting task requirements. Concurrently, the job description and system norms of the nursing position should be improved to allow nurses to clearly understand their role expectations, thus alleviating role-overload-related sleep disorders.
Second, the ability of nurses to cope with RO should be improved through mindfulness training. Mindfulness was originally developed as an intervention technique. During pandemics, incorporating mindfulness training in the workplace is good for health and may help nurses build personal resources, focus on the present, and try to look at the pandemic and RO in a nonjudgmental and objective manner to avoid emotional exhaustion and sleep disturbances (Hulsheger et al., 2018).
Several limitations of this study should be considered. First, because of the cross-sectional design used, it is impossible to definitively prove causal relationships between the variables. Second, RO is generally divided into overloads respectively related to quality and quantity. Studies on the mechanisms underlying the effect of RO on sleep quality in nurses have not explored the mechanisms of the two types of RO separately. It may be assumed that the effects of the two types of RO on sleep are different. Hence, further studies should be performed in this area. Third, we did not investigate the role of emotional distress on the association among sleep disorders, mindfulness, and RO. Emotional distress may be an important confounding factor that should be evaluated using more comprehensive measures. Future work may be conducted to investigate whether mindfulness reduces both emotional distress and RO in nurses to reduce the risk of sleep disorders. Finally, this study only used the PSQI to evaluate sleep and thus relied on participant self-reporting. Future studies should use objective methods such as polysomnography to evaluate sleep quality.
Conclusions
Sleep disturbances among frontline nurses are very common and worthy of concern. Few studies have focused on RO, which may contribute to sleep disturbance. This study found a significant relationship between RO and sleep disturbance. Our results imply that the high risk of experiencing sleep disturbance among nurses may be alleviated by reducing their perceived RO. The moderating role of mindfulness found in this study in the relationship between RO and sleep quality provides new insights to help improve sleep quality in nurses with high RO. Organizational strategies to provide mindfulness training and decrease RO among nurses may reduce sleep disturbances during the COVID-19 pandemic.
Acknowledgments
We thank the Fujian Medical University Educational Reform Fund (Grant Number 2019HL015, 2019) and Fujian Provincial Health Commission Science and Technology Program (Grant Number 2020GGB008, 2020) for their funding support.
Author Contributions
Study conception and design: NL, PZ
Data collection: NL, MC
Data analysis and interpretation: XC, LC
Drafting of the article: NL
Critical revision of the article: NL, PZ, HL
References