Factors associated with insomnia among frontline nurses during COVID-19: a cross-sectional survey study.

BACKGROUND: Research predominantly suggests that nurses are at high risk of developing psychopathology. The empirical data show that the occurrence rate of problem-related sleep quality among clinical nurses is high. Therefore, this study was conducted to address the lack of information on the relationship between the coronavirus disease 2019 (COVID-19) pandemic and insomnia.
METHODS: A convenience sample of nurses (n = 680) completed an online survey that included the Insomnia severity index, the COVID-19-related psychological distress scale, the general health questionnaire, neuroticism, dysfunctional beliefs, attitudes about sleep scale, and difficulties in emotion regulation scale.
RESULTS: The results showed that 35.8% (n = 253) of nurses were classified as individuals with moderate to severe clinical insomnia. The results showed that the psychological distress generated by COVID-19 predicted insomnia (β = .47, SE = 0.02, P < .001, t = 13.27, 95% CI 0.31-0.46). Additionally, the association is mediated by psychopathology vulnerabilities, emotion dysregulation, dysfunctional beliefs about sleep, and neuroticism. Moreover, female nurses exhibited higher levels of insomnia (Cohen's d = .37), neuroticism (Cohen's d = 30), psychopathology vulnerability (Cohen's d = .26), and COVID-19-related psychological distress (Cohen's d = .23).
CONCLUSION: The present study's findings help to explain how pandemic consequences can be associated with insomnia. Additionally, the findings make a significant contribution to better understanding the role of neuroticism, emotion dysregulation, beliefs, and psychopathology vulnerability in the development of insomnia among nurses. The findings suggest the potential influence of cognitive behavioral therapy for insomnia (CBT-I) and transdiagnostic integrated therapies that could be incorporated into therapeutic programs designed to develop as a way of inhibiting or preventing insomnia among clinical nurses.

Background

The outbreak of coronavirus disease 2019 (COVID-19) is more than a medical disaster. Similar to biological and natural disasters, COVID-19 can increase the risk of psychopathology vulnerabilities and posttraumatic stress disorder [1], particularly among vulnerable groups. Healthcare staff are a professional group that is exposed to multiple traumatic conditions. Compared with the general population, healthcare staff are remarkably (three times) more likely to be infected by a coronavirus [2]. Additionally, research predominantly suggests that HCWs are at high risk of developing psychopathology as first-line warriors against the pandemic [3]. Early study by the World Health Organization (WHO) authorities forewarns the possibility of an increase in mental health problems among nursing staff. Medical personnel in public health situations suffer not only physical and psychological stress but also the threat of prolonged exposure to infectious disease sources [4]. Following the COVID-19 epidemic, the demand for medical personnel surged considerably. The lack of healthcare workers has led to a dramatic increase in work overload for healthcare workers fighting COVID-19, as well as a great deal of anxieties and psychological distress among healthcare workers [5]. Exposure to a stressful situation, including actual or threatened death, can lead to sleep disorders. The qualitative and cross-sectional data indicate that the occurrence rate of sleep disturbance among nurses is high [6].

Sleep disorders

Sleep disorders are associated with considerable morbidity and functional impairment [7]. Insomnia is a prevalent sleep disturbance that refers to sleep symptoms such as difficulty initiating or staying asleep and irregular wake-sleep patterns and early morning awakening with nonresumption sleep [8]. Insomnia is a significant public health problem around the globe, with an estimated prevalence of 10–50% in adults [9]. Insomnia is recognized as a causal factor of mental health problems and the core symptoms in psychiatric disorders, particularly depression and anxiety disorders. Clinical staff, especially nurses, tend to be among the most affected by sleep disorders [10]. For nurses, sleep disturbance negatively affects quality of life [11], daily cognitive performance [12], work productivity, and patients’ treatment processes [13]. By identifying the insomnia risk factors that contribute to a nurse’s occupational and individual well-being, nursing management is better able to address the nurse’s needs to maintain positive well-being. This in turn will decrease burnout and increase the retention of experienced nurses, which will raise the quality of patient care.

Insomnia risk factors

According to neurobiological and psychological views, individual, behavioral [14], cognitive [15], and emotional [16] variables have been implicated in the onset of the maintenance of insomnia. Despite the conceptual frameworks used to describe the pathophysiology of Insomnia (e.g., neurobiological, behavioral), Insomnia etiology is largely understood from a diathesis-stress perspective. The 3P model (predisposing, precipitating, and perpetuating), suggested by Spielman et al. [17], describes how the interaction of different factors contributes to the initiation and persistence of acute insomnia. According to the 3P model, existing predisposing factors (e.g., neuroticism) make individuals more vulnerable to insomnia than others. Stressful experience is the most common precipitating factor of insomnia. Individuals with sleep problems recall the traumatic situation related to the onset of insomnia. It is posited that the joint effects of stressful life events [18, 19] and cognitive-emotional factors are central to the etiology of insomnia [20, 21].

For example, individuals with predisposing traits, when faced with uncommon or unexpected events, experience stress-related insomnia symptoms. The predisposing factors interact with precipitating factors (e.g., traumatic events), resulting in chronic insomnia. The chronic course of insomnia is maintained in acute insomnia by perpetuating cognitions (e.g., dysfunctional belief about sleep) and sleep-disrupting behaviors (e.g., emotion dysregulation) [22] after the severity of the precipitating stressor has lessened.

Purpose of the current study

The WHO has underscored the excessive burden on nurses and called for immediate action to prevent a severe impact on the physical and mental health of nurses [23]. Identifying factors and mechanisms that contribute to the development and persistence of insomnia should be a priority to identify better strategies that improve prevention and treat insomnia and its comorbid conditions. Sleep problems appear to have been common during the COVID-19 pandemic. Compared to individuals without insomnia, the negative impacts of the pandemic are more severe in people with insomnia [24]. In addition to high prevalence, insomnia increases the risk of medical and psychological illnesses and negatively impacts cognitive and job performance [25]. While research predominantly concentrates on sociodemographic factors and the prevalence of insomnia, research on the relationship between factors associated with insomnia is extremely scarce [26] There is evidence that sleep problems are associated with higher levels of psychological distress [27].

Therefore, the present study was conducted to address the lack of information on the relationship between COVID-19 psychological distress and insomnia. The study also explored whether COVID-19-related psychological distress directly predicted insomnia symptoms. It was hypothesized that psychological distress related to COVID-19 directly predicted insomnia symptoms. Additionally, it was hypothesized that dysfunctional beliefs about sleep, neuroticism, difficulties with emotion regulation, and psychopathology vulnerability would mediate the association between COVID-19-related psychological distress and insomnia symptoms. More specifically, it was expected that COVID-19 related to psychological distress indirectly influences insomnia through dysfunctional beliefs about sleep, neuroticism, difficulties with emotion regulation, and psychopathology vulnerability.

Method

Sample

The study sample comprised (N = 680) nurses (55.7%, n = 323 females) aged 23–45 enrolled via an online survey. The participants’ mean age was 32.85 years (SD = 7.09). Eligibility criteria included being (i) aged 18 years or above, (ii) able to read and complete an online consent form and survey, and (iii) not having COVID-19 or chronic disease (e.g., diabetes, cancer). Table 1 presents the overview of the sample’s demographic characteristics and group differences. The study applied a survey method for data collection, and the subjects obtained informed consent before enrolment. The study was found to be in accordance with the Helsinki Declaration, evaluated by the Research Ethics Committees of Lorestan University of Medical Sciences.

Table 1

Demographic characteristics of the sample (N = 680)

Item	Value		Test	P Value
Categorical variables
Gender, n (%)
Men	257 (44.3)		χ2 = 7.510	.006
Women	323 (55.7)
ISI, n (%)
ISI < 8	206 (30.3)
8 ≤ ISI < 15	231 (33.9)		χ2 = 16	<.001
15 ≤ ISI < 22	159 (23.4)
22 ≤ ISI	84 (12.4)
Experience level, n (%)
Years ≤5	250 (37.1)		χ2 = 14.5	<.001
5 < years	343 (62.9)
Continuous variables M (SD)
Age (years)	32.85	7.09	t(1, 678) = −1.73	.08
Insomnia	10.09	5.55	t(1, 678) = 4.50	<.001
Emotion Dysregulation	31.87	7.95	t(1, 678) = −2.35	.02
Dysfunctional Belief	42.55	9.41	t(1, 678) = 1.65	.1
Neuroticism	23.20	7.32	t(1, 678) = 3.58	<.001
Psychopathology Vulnerability	33.44	12.71	t(1, 678) = 3.12	.002
COVID-19 related psychological distress	11.87	7.28	t(1, 678) = 2.75	.01

.Note: n = frequency; M = mean; SD = standard deviation

t = independent t test to compare gender status, negative sign = Males higher score

Measures

Insomnia symptoms

The Insomnia Severity Index (ISI: [28]; Persian version: [29]), a self-report brief screening tool, was employed to assess insomnia severity over the last month. The subjects rated a seven-item scale on a five-point Likert scale from zero = no problem to four = severe pain, with a total score ranging from 0 to 28. In the current study, a total ISI score lower than 8 and a total ISI score of 8 ≤ ISI < 15 were considered to indicate no clinically significant insomnia and subthreshold insomnia, respectively. ISI total scores of 15 to 21 and 22 to 28 were considered moderate clinical insomnia and severe clinical insomnia, respectively. A high internal consistency coefficient was obtained for the current sample (Cronbach α = .88).

Dysfunctional beliefs

Dysfunctional beliefs and attitudes about sleep scale (DBAS: [30]; Persian version: [31]) was employed to measure unrealistic expectations and dysfunctional beliefs about sleep. Nurses rated each item on eleven Likert scales from zero (strongly disagree) to ten (strongly agree). A higher score indicates a higher level of dysfunctional beliefs about sleep. This measure had high internal consistency in our sample (α = .83).

Personality trait

The neuroticism subscale of the Big Five Inventory (BFI-44: [32]; Persian version: [33]) was used to assess neuroticism. Nurses rated eight items on a five-point scale of 1 (strongly disagree) to 5 (strongly agree), with a total score of 8–40. A higher score was interpreted as a greater level of neuroticism. The scale reliability was very good in the present study (α = .80).

Psychopathology vulnerability

The General Health Questionnaire (GHQ-28: [34]; Persian version: [35]) was used to measure the severity of psychopathology in four domains (i.e., depression, anxiety, social dysfunction, and somatic symptoms) during the past few weeks in nonpsychiatric settings. Nurses rated on a four-point scale ranging (0, 1, 2, 3). Higher scores indicate greater levels of psychopathology vulnerability. The scale reliability was very good in the present study (α = .83).

Emotion dysregulation

The Difficulties in Emotion Regulation Scale (DERS: [36]; Persian version: [37]) was employed to measure emotion dysregulation. The nurses rated 18 items on a five-point scale from 1 (almost never) to 5 (almost always), with total scores ranging from 18 to 90. Higher scores were considered to indicate greater levels of difficulty in emotion regulation. The scale demonstrated perfect internal consistency in the present study (α = 0.85).

Psychological distress

The COVID-19-Related Psychological Distress Scale (CORPDS) [38] is a specific self-report scale that assesses psychological distress in the context of COVID-19. Nurses scored the fourteen items on a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Higher scores demonstrate severe significant distress. This instrument had acceptable internal consistency (α = 0.87).

Demographic variables

Participants were asked their age, gender, and experience. However, socioeconomic status (SES) information was not requested. Participants were also asked whether they had COVID-19 or chronic disease, with the following questions: (ii) Are you currently being monitored or treated for a physical illness (e.g., diabetes, cancer) or mental health problems (e.g., depression), and (ii) Are you currently quarantined or suspected of being infected by COVID-19?

Procedure

Participant recruitment

The study was conducted during the COVID-19 pandemic via convenience sampling; hence, all data were collected online because face-to-face data collection was not possible. The participants were recruited from January to September 2021 using an online platform. The recruitment process included advertising the study via social media platforms (Instagram, WhatsApp) and online professional health staff forums with links to the survey. Additionally, the forum members were requested to share the study invitation letter with other colleagues. Once the participants clicked on the distributed survey link, an informed consent page was opened. Informed consent was agreed upon before proceeding to the survey. Once the nurses obtained informed consent, they accessed the survey. Informed consent included information about the study’s objectives and confidentiality.

Data analysis

SPSS (version 25, SPSS Inc., Chicago, IL) and AMOS (version 24, IBM) were utilized to test hypotheses with a two-tailed alpha level of 0.05 to determine statistical significance. The values for kurtosis and skewness (values within <|1|) were examined to test the assumption of a normal distribution. Chi-squares and independent t tests were performed to investigate the differences between groups (females and males). The variance inflation factor (VIF values < 5) was considered as the absence of multicollinearity [39]. Pearson correlation coefficients (r) were evaluated to investigate the relationship between associated variables.

Multiple mediation analysis was conducted to investigate whether the four mediators would mediate the association between COVID-19-related psychological distress (independent variable) and insomnia symptoms (outcome variable) using structural equation modeling (SEM). The guidelines suggested by Hu and Bentler [39] were considered in evaluating the goodness of fit of the model (comparative fit index (CFI) and Tucker–Lewis index (TLI) > .95; standardized root mean square residual (SRMR) < .06), PCLOSE >.05, and root mean square error of approximation (RMSEA) < .06). An indirect effect was significant if the calculated 95% bias-corrected confidence intervals (CIs) were not included zero. The 95% CI was generated by the bias-corrected method for the point estimate with 5000 bootstrap samples [40]. The guideline proposed by Cohen [41] was considered to evaluate the effect size values (Cohen f 2 ≥ 0.15: medium effect size; Cohen f 2 ≥ 0.35 large effect sizes). A priori power analysis was carried out using G-Power. The calculation indicated the desired sample size of 647, considering a small effect size for multiple linear regression, a power of 0.80, an alpha of .05, and five predictors [42].

Results

Descriptive statistics

Of the 720 distributed web-based surveys, a total of 680 useable surveys were included in the final analysis. The data are reported as the means (M) and standard deviations (SD) for continuous variables and as frequencies (n) or percentages for categorical variables in Table 1. The generated VIF and the calculated kurtosis and skewness values showed no violation of the multicollinearity issue and the normality (see Table 2). Compared with males, females obtained higher scores for insomnia (t [678] = 4.50, P < .001, Cohen’s d = .37), neuroticism (t [678] = 3.58, P < .001, Cohen’s d = .30), COVID-19-related psychological distress (t [678] = 2.75, P = .01, Cohen’s d = .23), and psychopathology vulnerability (t [678] = 3.12, P = .002, Cohen’s d = .26). Males had significantly higher scores in emotion dysregulation (t [678] = − 2.35, P = .02, Cohen’s d = .20). The results also showed that 35.8% (n = 253) of nurses obtained an ISI score ≥ 15, which was classified as moderate severe clinical insomnia. The correlation analyses are reported in Table 2. Experience was negatively correlated with insomnia (r = −.27; P < .001), dysfunctional belief (r = −.11; P = .002), and neuroticism (r = −.19; P < .001).

Table 2

Correlation Matrix and item analysis

Variable	1	2	3	4	5	6	Sk	Kur	VIF
1-Insomnia	.1.00						−.79	− 1.98	1.25
2-Emotion Dysregulation	.45**	1.00					−.57	−1.10	2.65
3-Dysfunctional Belief	.36**	.29**	1.00				−.71	1.29	1.57
4-Neuroticism	.61**	.24**	.16**	1.00			−1.12	.71	1.95
5-Psychopathology Vulnerability	.57**	.39**	.34**	.27**	1.00		−.52	−.44	2.23
6-CORPD	.47**	.30**	.37**	.19**	.45**	1.00	−.78	.95	2.51

Note: VIF variance inflation factor, CORPD COVID-19-Related Psychological distress, SK skewness, Kur Kurtosis; **Correlation significant at the P < .001 level (two-tailed)

Multiple mediation analysis

The predictive model is illustrated in Fig. 1. The analysis produced excellent fit for the model (χ2/df = 1.42, CFI = .996, TLI = .995, SRMR = .02, RMSEA = 0.03, 90% CI 0.02–0.05, R-squared = 0.73). While the total standardized effect of COVID-19-related psychological distress on insomnia was statistically significant (β = .47, SE = 0.03, P < .001, t = 13.77, 95% CI 0.39–0.54), with a medium to large effect size (Cohen f  = 0.29, P < .001), the analysis showed that COVID-19 related to psychological distress directly predicted insomnia, with a small effect size (Cohen f  = 0.06, P < .001). Additionally, COVID-19-related psychological distress indirectly predicted insomnia through dysfunctional beliefs about sleep (β = .03, SE = 0.01, P = .001, t = 3.29, 95% CI 0.01–0.05), neuroticism (β = .09, SE = 0.02, P < .001, t = 4.39, 95% CI 0.05–0.13), difficulties with emotion regulation (β = .04, SE = 0.01, P < .001, t = 4.10, 95% CI 0.02–0.06), and psychopathology vulnerability (β = .16, SE = 0.02, P < .001, t = 9.12, 95% CI 0.13–0.19). The overall indirect effect of all mediators was also statistically significant (β = .32, SE = 0.03, P < .001, t = 10.69, 95% CI 0.26–0.38). The standardized effects and effect sizes of the model paths are reported in Table 3.

Fig. 1

Parallel Mediation Model. Note: Path coefficient = standardized coefficient (t-value)

Table 3

The standardized effects and effect sizes of the model path

Path	Betaa	SE	Pvalue	t value	95%CIs b	Cohen f 2
Direct path
DBAS - > Insomnia	0.10	0.02	.001	3.57	0.04–0.13	0.02
GHQ - > Insomnia	0.35	0.03	<.001	12.39	0.30–0.41	0.33
DER - > Insomnia	0.14	0.03	<.001	5.17	0.09–0.19	0.06
Neuroticism - > Insomnia	0.51	0.02	<.001	18.50	0.46–0.55	0.52
CORPD- > DBAS	0.36	0.04	<.001	8.47	0.27–0.44	0.15
CORPD- > GHQ	0.45	0.04	<.001	11.15	0.37–0.52	0.25
CORPD- > DER	0.30	0.05	<.001	6.54	0.21–0.38	0.10
CORPD- > Insomnia	0.15	0.04	<.001	5.64	0.10–0.20	0.06
CORPD- > Neuroticism	0.19	0.04	<.001	4.38	0.10–0.26	0.04

Note: a Beta standardized path coefficient, CORPD OVID-19-Related Psychological Distress, DER Difficulties in Emotion Regulation, GHQ The General Health Questionnaire, DBAS Dysfunctional beliefs and attitudes about sleep, b CIs confidence intervals

Discussion

While nurses are affected by the pandemic, there is currently a lack of information regarding the occurrence of psychopathological symptoms in nursing staff. It was hypothesized that psychological distress related to COVID-19 directly predicted insomnia symptoms. It was also assumed that dysfunctional beliefs about sleep, neuroticism, emotion dysregulation, and psychopathology vulnerability would mediate the association between psychological distress related to COVID-19 and insomnia. Consistent with expectations and previous studies [43-45], the results indicated that COVID-19 psychological distress is associated with a greater level of insomnia, with a large effect size. The association is partially mediated by the associated variables.

COVID-19 psychological distress predicts higher levels of neuroticism. Neuroticism is a salient risk factor for Insomnia. The findings provide evidence that neurotic nurses exhibit higher levels of insomnia [43, 46]. The vulnerability or risk model hypothesizes that pre-existing personality traits (e.g., neuroticism) predispose an individual to develop sleep disorders [47]. The study’s findings extend prior findings that neuroticism strongly predicts lower subjective sleep quality [48]. Additionally, individuals high in neuroticism exhibit lower sleep quality and frequently report unwanted wakefulness after sleep onset [49]. Vulnerability to psychological disorders is key to understanding differential stress responses to severe life events. Depressive symptoms are one of the most common causes of sleep disorders, with more than 40% of insomnia cases caused by depressive symptoms [50]. According to the prevalence rates of insomnia, anxiety, and depression among healthcare workers, more than one-fifth of healthcare workers suffer from anxiety and depression, and over two-fifths suffer from sleeplessness [51].

Regarding the association between difficulties in emotion regulation and insomnia, growing evidence supports that the link is bidirectional [52, 53]. Compared with good sleepers, individuals with insomnia significantly utilize maladaptive emotion regulation strategies such as suppression and avoidance [54]. Neuroimaging studies also support the association between emotion regulations and sleep quality [55]. The prefrontal cortex (i.e., principal brain region responsible for cognitive control emotion regulation) is sensitive to sleep deprivation [56]. Compared to healthy individuals, experimental research has revealed that individuals with problem-related sleep present impaired functional connectivity between the prefrontal cortex and amygdala [57].

Unrealistic and erroneous cognitions significantly predicted insomnia, with moderate effect sizes. The study’s findings highlight the role of dysfunctional beliefs and attitudes about the nature of sleep, which may lead to a change in the course of the chronic form of insomnia to a persistent form [58]. In line with previous literature and cognitive models of insomnia [59], this perpetuating factor disrupts sleep in two ways. It triggers arousal and distress, increasing selective attention, and monitoring toward sleep-related threat cues, resulting in an overestimation of sleep deficits. Secondly, negative cognitions provoke safety behaviors, which develop dysfunctional beliefs about sleep [60, 61].

The study’s findings reveal the small associations between experience and insomnia. Additionally, higher experience levels were negatively associated with insomnia, emotion dysregulation, and neuroticism. Nurses with more experience may potentially adjust to the stressful working environment and regulate to the demanding working environment. Consistent with previous research, significant gender differences were found. In line with a broad range of COVID-19 literature, during the pandemic, females report higher levels of psychopathology vulnerabilities and are more at risk for problem-related sleep disorders than males [62, 63].

Practical implication

The findings also hold practical application in the design of primary and secondary prevention interventions. In terms of primary interventions, administration, policymakers, and healthcare providers must plan and implement appropriate programs and interventions to support nurses in preventing sleep issues. The etiology of insomnia is multidimensional; the treatment is also a multicomponent treatment. The American Academy of Physicians recommended CBT-I as a first-line therapy for insomnia, which is a multicomponent nonpharmacological treatment [64]. There is also extensive research demonstrating the effectiveness of CBT-I in the context of comorbid conditions. The CBT-I covers a broad range of psychological aspects ranging from changing cognitive thoughts to the reduction of psychiatric symptoms [65, 66].

In terms of secondary intervention, personality traits have been considered stable and inflexible over time, but growing evidence suggests that neuroticism may be more malleable than believed. In line with previous research [67]. COVID-19 distress was associated with greater neuroticism levels [68]. Neuroticism is recognized as a key etiological mechanism that is shared by psychological disorders [69, 70]. Emotional dysregulation also plays a crucial role in the treatment of complex cases, diagnoses with comorbidities, and psychological risk factors [71]. These mechanisms can be crucial in initiating, increasing, or maintaining emotional disorders and insomnia. In terms of secondary prevention, targeting the disorders’ risk factors before the acute form of the disorder potentially improves prevention and treats insomnia and its coexisting conditions. Understanding the contribution or coexistence of these risk factors has implications for Insomnia assessment and treatment. Given the high prevalence of insomnia and its comorbidity with a mental health problem, disorder-specific protocols could be difficult to justify when the comorbidities are the norm and clinical reality is complex [72]. Transdiagnostic and integrated therapies target identified groups of underlying core processes and can serve as a promising intervention [73]. The transdiagnostic interventions (e.g., TranS-C) comprise core and optional modules, which provide the treatment sessions to be more personalized to the specific sleep problem [74].

Limitations

The present study suffers from limitations – most notably related to the participants’ enrolment and information collection. The current research was carried out during the COVID-19 pandemic. Therefore, to follow pandemic instruction, data are collected online instead of the traditional method. This meant that nurses without internet access could not participate. Subsequently, the collected data do not represent such groups’ considerations and influence the study’s generalizability. Additionally, the self-report data were subject to common method biases. Finally, the study was cross-sectional. Therefore, causality between the study’s variables cannot be determined.

Conclusions

Despite these limitations, the present study’s findings help to explain how pandemic consequences can be associated with insomnia. The present study contributes to understanding the role of predisposing, precipitating, and perpetuating factors in the development of insomnia among nurses. The findings make a significant contribution to the expanding literature on emotion dysregulation, beliefs, and psychopathology vulnerability in insomnia. The findings suggest the influence of CBT-I, transdiagnostic, and integrated therapies that could be incorporated into therapeutic programs designed to develop as a way of inhibiting or preventing insomnia among clinical nurses. Future studies would need longitudinal designs to determine true causality.

Additional file 1. STROBE Statement—checklist of items that should be included in reports of observational studies.