Characteristics of Sleep Disturbance and Comparison Across Three Waves of the COVID-19 Pandemic Among Healthcare Workers
Article information
Abstract
Objective
Healthcare workers (HCWs) suffered from a heavy mental health burden during the coronavirus disease-2019 (COVID-19) pandemic. We aimed to explore the differences in sleep disturbance in three waves of the COVID-19 pandemic in Taiwan among HCWs. Moreover, factors associated with sleep disturbances in the third wave were investigated.
Methods
This study, with three waves of cross-sectional surveys, recruited first-line and second-line HCWs. The level of sleep disturbance and related demographic variables were collected through self-report questionnaires. Differences in sleep disturbance across the three waves were compared with analysis of variance. Factors associated with the level of sleep disturbance were identified using univariate linear regression and further used for multivariate stepwise and bootstrap linear regression to identify the independent predictors.
Results
In total, 711, 560, and 747 HCWs were included in the first, second, and third waves, respectively. For first-line HCWs, sleep disturbance was significantly higher in the third wave than in the first wave. The level of sleep disturbance gradually increased across the three waves for all HCWs. In addition, sleep disturbance was associated with depression, posttraumatic stress disorder (PTSD) symptoms, anxiety about COVID-19, vaccine mistrust, and poorer physical and mental health among first-line HCWs. Among second-line HCWs, sleep disturbance was associated with younger age, depression, PTSD symptoms, lower preference for natural immunity, and poorer physical health.
Conclusion
The current study identified an increase in sleep disturbance and several predictors among HCWs. Further investigation is warranted to extend the application and generalizability of the current study.
INTRODUCTION
Coronavirus disease-2019 (COVID-19) greatly affected physical and mental health [1] and resulted in different kinds of psychological burdens for healthcare workers (HCWs). These psychological impacts on HCWs can be pronounced, including sleep disturbance, depression, anxiety, and posttraumatic stress disorder (PTSD) symptoms [2,3]. Moreover, a meta-analysis demonstrated pooled prevalence rates of anxiety, depression, PTSD, sleep disturbance, and psychological distress among hospital staff of 34.4%, 31.8%, 11.4%, 27.8%, and 46.1%, respectively [4]. In Taiwan, HCWs have also been reported to have PTSD symptoms, sleep problems, and psychological distress due to the COVID-19 pandemic [5,6]. Several reasons may explain the massive psychological distress for HCWs. During the pandemic, the unprecedented workload resulted in burn out for HCWs [5]. On the other hand, the high risk of COVID-19 and related complications [7] may also have contributed to distress.
In the current study, we focused on sleep problems among HCWs during the pandemic. Since the impact of the COVID-19 pandemic on sleep disturbance has been evaluated, factors associated with sleep problems attracted our interest. Female sex and working in allied health (nonphysician) predicted sleep disturbance and distress among HCWs during the pandemic [8]. Another study revealed that younger age, female sex, frontline status, fear or risk of COVID-19, and a lower level of social support were all associated with a greater risk of sleep problems for HCWs [9].
Although factors of sleep problems among HCWs have been preliminarily identified, several gaps in the literature deserve further investigation. First, although sleep disturbance during the COVID-19 pandemic was pronounced [10,11], the difference in sleep problems before and after the pandemic remain inconclusive. The prevalence of sleep problems for HCWs during the COVID-19 pandemic was reported to be higher than that in the pre-COVID-19 stage [12]. However, another study demonstrated the opposite finding, and the authors supposed that it might have resulted from the lockdown and decreased social activities at night during the pandemic [13]. This raises our interest in how COVID-19 affected sleep patterns for HCWs, especially during different waves of the pandemic. To date, most studies focusing on sleep problems among HCWs have been cross-sectional [4], and few studies have explored changes in sleep quality in different waves of the COVID-19 pandemic. Second, further factors associated with sleep disturbances have yet to be investigated, such as detailed coping strategies for COVID-19. Therefore, we aimed to investigate the differences in sleep disturbance in different waves of the COVID-19 pandemic in Taiwan among HCWs. We also comprehensively explored factors associated with sleep disturbances in different categories, such as psychological distress, quality of life and health, coping strategies for COVID-19, and attitudes toward COVID-19.
METHODS
Participants and ethics
The current study derived data from a series of studies to explore the psychological, social, and other dimensional impacts of COVID-19 on the public, HCWs, and patients with mental illness since 2020 in Kaohsiung Municipal Kai-Syuan Psychiatric Hospital (KSPH) and affiliated institutes. The recruitment and procedure details have been published elsewhere [14-17]. In brief, these studies were cross-sectional surveys with paper-and-pencil questionnaires, and research assistants individually explained the procedures to participants to complete the research questionnaires. In the current study, we included HCW data from the following three periods: May 9, 2020, to May 31, 2020; May 30, 2021, to June 30, 2021; and March 20, 2023, to August 20, 2023. The above periods were close to the three waves of the COVID-19 pandemic in Taiwan with the Alpha/Beta variants, Delta variant, and Omicron variant. The inclusion criteria for HCWs were as follows: 1) healthcare staff who worked at KSPH or affiliated institutes, 2) those aged more than 20 years, and 3) those who signed informed consent forms before completing the questionnaire. The surveys of the three waves were approved by the Institutional Review Board of KSPH (KSPH-2020-03; KSPH-2021-08; KSPH-2023-04).
Measures
Disaster-Related Psychological Screening Test
The Disaster-Related Psychological Screening Test (DRPST) was developed to rapidly screen for major depressive disorder or PTSD after a disaster, and it has good reliability and validity [18,19]. To estimate the level of depression, 3 items of the DRPST were chosen to measure the level of depressed mood, fatigue or loss of energy, and worthlessness that had persisted for more than 2 weeks in the last month. Each question was rated on a 2-point Likert scale, with scores ranging from 0 (no) to 1 (yes). A higher total score of the 3 items indicated a higher level of depression.
To identify the severity of PTSD symptoms, four questions were selected from the DRPST to estimate the status of hypervigilance, somatic symptoms, avoidance, and re-experience of COVID-19 that had persisted for more than 1 week in the past month. Each question was rated on a 5-point Likert scale, with scores ranging from 1 (not at all) to 5 (extreme). A higher total score indicated a higher level of psychological distress. Details of the full questionnaires are listed in Supplementary Table 1.
Sleep disturbance scales from the Pittsburgh Sleep Quality Index
The Pittsburgh Sleep Quality Index (PSQI) was initially developed to measure sleep quality in clinical populations with good validity and reliability [20]. In the current study, 4 items of the PSQI were used to measure the level of sleep disturbance, including difficulty falling asleep, waking during the night, subjective sleep quality, and enthusiasm in the past month (Supplementary Table 1). Each item was rated on a 4-point Likert scale, with scores ranging from 1 to 4. Higher total scores of the 4 items indicated a more severe sleep disturbance.
The 12-item Short Form Survey version 2
The 12-item Short Form Survey version 2 (SF-12v2) is based on scoring coefficients derived from version 1 of the 36-Item Short Form Health Survey, which was developed to rapidly estimate general health status and has been well validated [21]. The SF-12v2 [22] is one of the most commonly used health-related quality of life questionnaires, and it has become widely used in community-based health surveys and illness outcome assessments due to its brevity, good reliability and validity [23,24].
These items were graded on a 5-point Likert scale, with scores ranging from 1 (extremely) to 5 (not at all). A higher score represented a higher health quality and lower disability. This questionnaire contained two components: a mental component summary of Short Form-12 Items Health Survey (MCS) and a physical component summary of Short Form-12 Items Health Survey (PCS). The PCS and MCS represent the quality of physical and mental health, respectively.
Societal Influences Survey Questionnaire
The Societal Influences Survey Questionnaire (SISQ) was developed to measure the psychosocial impact on and change in lifestyle among individuals during the COVID-19 pandemic. With good validity and reliability, the 15-item SISQ contains five assessment categories, including social distancing, social anxiety, social desirability, social information, and social adaptation [25,26]. Each question is graded on a 4-point Likert scale, with scores ranging from 1 (never) to 4 (often). Higher total scores for each of the five categories indicated higher compliance with maintaining social distancing, higher levels of anxiety due to COVID-19, a greater desire to seek COVID-19-related information, and greater awareness of the progress of the COVID-19 pandemic overseas.
Lo’s Healthy and Happy Lifestyle Scale
Lo’s Healthy and Happy Lifestyle Scale (LHHLS) was constructed to estimate the wellbeing associated with mental health and lifestyle, and it has been reported to have good reliability and validity [27]. With 14 questions, the LHHLS is graded on a 5-point Likert scale, with responses ranging from a scale of 1 (never) to 5 (always). Participants were asked to rate themselves regarding their feelings of a happy and healthy lifestyle over the past 2 weeks. A higher score on the LHHLS indicates a higher level of wellbeing in health and lifestyle.
Vaccination Attitudes Examination Scale
The Vaccination Attitudes Examination Scale (VAX) was developed to measure the attitudes of participants regarding vaccination for COVID-19 [28]. The Chinese version of the VAX has also been reported to have acceptable reliability and validity [14]. The VAX contains 12 questions in four categories, including mistrust of vaccine benefits, worries about unforeseen future effects, concerns about commercial profiteering, and preference for natural immunity. Each item was rated on a 6-point Likert-type scale ranging from “strongly agree” to “strongly disagree.” A higher total score for mistrust of vaccine benefits indicated greater trust toward vaccine benefits, and this category indicated a score deduction category in total scores of the four categories. Higher total scores of the remaining three categories represented higher levels of worries about unforeseen future effects of vaccines (adverse events or side effects), disbelief in commercial benefits regarding the production of vaccines, and preference for natural immunity but not vaccination. Together, a higher total score on the VAX indicated stronger antivaccination attitudes and mistrust of authorities. The details of the VAX are listed in Supplementary Table 1.
Demographic characteristics
Data were recorded as continuous variables for the participants’ age, educational level (years), and level of the above measurements. Categorical variables included sex, marital status (with or without partner), religion (religious or not religious), history of psychological trauma (yes or no), smoking (yes or no), alcohol consumption (≥3 times per week or not), regular exercise (≥3 days per week or not), regular diet habits (three or four meals a day, ≥5 days per week or not), history of chronic medical disease (yes or no), history of COVID-19 infection (yes or no), and COVID-19 vaccination (yes or no). To be applied in the further analysis, marital status was transformed into a dichotomous variable as with a partner (married and cohabitating) or without a partner (single, widowed, and divorced). HCWs were divided into two groups, including first-line HCWs (physicians and nurses) and second-line HCWs (social workers, psychologists, occupational therapists, pharmacists, administrative staff, and other paramedical staff), for further analysis.
Statistical analysis
For the participants in the third wave, descriptive analysis was used to demonstrate the demographic variables in detail. They were divided into groups of first-line HCWs and secondline HCWs, and the two groups were compared. Pearson’s χ2 test was used to compare categorical variables, and the independent t-test was used for continuous variables. As data on the comparison of sleep disturbance between HCWs in the first and second waves have been published [17], we further added data from the third wave to compare the long-term differences in sleep disturbance among HCWs in the three waves. One-way between-group analysis of variance (ANOVA) was conducted to explore the difference between waves for first-line and second-line HCWs. The assumption of homogeneity of variance needed to be tested initially. Homogeneity of variance was assessed with Levene’s test. If Levene’s test yielded a p-value above or equal to 0.05, then the assumption of homogeneity of variance was not violated. The F statistic was applied, and post hoc comparisons were made with Bonferroni correction. A p-value of 0.05 was used to indicate significance in the post hoc comparison. If Levene’s test yielded a p-value below 0.05, it indicated that the assumption of homogeneity of variance was violated. Then, the Brown-Forsythe F statistic was used, and post hoc analysis was performed with Dunnett’s T3 test.
In addition, we performed further analysis for data in the third wave to identify the characteristics of sleep disturbance. We analyzed data only in the third wave for the subsequent analysis because there was an abundance of questionnaires in the survey of the third wave. Univariate and stepwise multiple linear regression were used to ascertain that the independent factors were associated with the level of sleep disturbance, which was estimated with the PSQI. The alpha level was set at 0.05. The normality of dependent variables was checked by the Kolmogorov–Smirnov test. Because the nonnormally distributed samples were identified with the significance of the test (p<0.001), bootstrapping multiple linear regression with 5,000 bootstrap samples was used to verify the results from the linear regression. In the bootstrapping method, the 95th percentile of the confidence interval was applied to determine statistical significance, which confirmed the stability of the regression coefficient and reduced the length of the confidence interval [29]. In addition, the number of bootstrap samples was set to 5,000 to obtain a sufficiently accurate 95th bootstrap percentile [30]. All data were processed using SPSS version 27.0 for Windows (IBM Corp., Armonk, NY, USA).
RESULTS
Summary of demographic analysis
In total, 340, 313, and 504 first-line HCWs were included in the first, second, and third waves, respectively. For second-line HCWs, 371, 247, and 243 subjects were included in the first, second, and third waves, respectively. Details of demographic information in the first and second waves have been reported [17]. In the third wave, age (39.24 vs. 41.41, p=0.011) and educational level (15.62 vs. 16.09, p=0.008) were significantly lower among first-line HCWs than among second-line HCWs. The proportion of regular exercise (52.2%. vs. 65.8%, p<0.001) and regular diet (70.0%. vs. 87.6%, p<0.001) were significantly lower among first-line HCWs than among second-line HCWs. However, proportion of history of COVID-19 infection were higher among first-line HCWs (57.3%. vs. 43.2%, p<0.001) than among second-line HCWs. In addition, first-line HCWs had a significantly higher level of sleep disturbance (6.14 vs. 5.46, p=0.007), concerns about commercial profiteering (8.54 vs. 7.79, p=0.005), and preference for natural immunity (10.01 vs. 9.18, p=0.002) than second-line HCWs. The remaining details are listed in Table 1. Moreover, the distribution of detailed marital status, source of psychological trauma, and chronic diseases among first-line and second-line HCWs are listed in Supplementary Tables 2-7.
Characteristics of demographic and quantitative variables for HCWs at third wave
One-way ANOVA across three waves
Table 2 indicates the differences in sleep disturbance across the three waves. Due to the violation of homogeneity, Brown-Forsythe statistics were applied. For first-line HCWs, there was a statistically significant difference in the level of sleep disturbance (F[2, 1111.41]=5.43, p=0.005). Post hoc comparisons estimated with Dunnett’s T3 test demonstrated that the level of sleep disturbance in the third wave was significantly higher than that in the first wave (6.14 vs. 5.41, p=0.004). The level of sleep disturbance in the second wave (5.86±3.08) did not differ significantly from that in the first and third waves. For second-line HCWs, no statistically significant difference in the level of sleep disturbance was identified across the three waves (F[2, 695.36]=2.93, p=0.054). However, a trend was identified in that the level of sleep disturbance in the third wave was non-significantly higher than that in the first wave (5.46 vs. 4.93, p=0.07). The remaining data are listed in Table 2.
Difference of sleep disturbance measured by PSQI in three waves of COVID-19 pandemic
Predictors for the level of sleep disturbance
Among first-line HCWs, the results of univariate linear regression showed that a higher level of sleep disturbance was significantly associated with a higher level of depression (β=0.42; p<0.001), PTSD symptoms (β=0.32; p<0.001), social anxiety (β=0.20; p<0.001), social information (β=0.12; p=0.005), female sex (β=0.11; p=0.013), psychological trauma (β=0.12; p=0.009), regular alcohol consumption (β=0.12; p=0.005), and no regular diet (β=-1.01; p=0.024). Sleep disturbance was also related to lower levels of vaccine trust (β=-0.16; p<0.001), wellbeing (β=-0.34; p<0.001), PCS score (β=-0.43; p<0.001), and MCS score (β=-0.46; p<0.001). After stepwise multivariate regression, only higher levels of depression and PTSD symptoms and lower MCS and PCS scores remained significant (Table 3). In addition, bootstrapping regression showed that sleep disturbance was associated with a higher level of depression, PTSD symptoms and social anxiety and lower MCS and PCS scores (Supplementary Table 8).
Predictors for level of sleep disturbance examined by univariate and stepwise multivariate linear regression among first-line HCWs in the third wave
For second-line HCWs, the results of univariate linear regression demonstrated that a higher level of sleep disturbance was significantly associated with a higher level of depression (β=0.40; p<0.001), a higher level of PTSD symptoms (β=0.22; p=0.001), a higher level of social anxiety (β=0.15; p=0.017), female sex (β=0.14; p=0.025), psychological trauma (β=0.17; p=0.007), no regular diet (β=-0.14; p=0.031), and younger age (β=-0.18; p=0.005). Sleep disturbance was also related to lower levels of vaccine trust (β=-0.21; p=0.001), preference for natural immunity (β=-0.23; p<0.001), social desirability (β=-0.13; p=0.043), well-being (β=-0.34; p<0.001), PCS (β=-0.38; p<0.001), and MCS score (β=-0.39; p<0.001). After verification with stepwise multivariate regression, younger age, a higher level of depression, higher level of PTSD symptoms, lower level of preference for natural immunity, and lower PCS score remained significant (Table 4). The bootstrapping regression showed that sleep disturbance was associated with a higher level of depression, a lower level of preference for natural immunity, and a lower PCS score (Supplementary Table 9).
Predictors for level of sleep disturbance examined by univariate and stepwise multivariate linear regression among second-line HCWs in the third wave
DISCUSSION
Main findings of the current study
To our knowledge, this is the first study to compare the difference in the level of sleep disturbance across three waves of the COVID-19 pandemic and comprehensively assess the multidimensional factors associated with sleep disturbance among HCWs. For first-line HCWs, sleep disturbance was significantly higher in the third wave than in the first wave. For second-line HCWs, a nonsignificant trend was also identified in which sleep disturbance was more severe in the third wave than in the first wave. In summary, the level of sleep disturbance gradually increased across three waves of the pandemic among both first-line and second-line HCWs. In the third wave, sleep disturbance was associated with a higher level of depression, a higher level of PTSD symptoms, a higher level of anxiety about COVID-19, a lower level of vaccine trust, and poorer physical and mental health among first-line HCWs. For second-line HCWs in the third wave, sleep disturbance was associated with younger age, higher levels of depression, higher levels of PTSD symptoms, lower levels of preference for natural immunity, and poorer physical health.
Changes in sleep quality in different COVID-19 waves
We found a novel upward trend in sleep disturbance across three waves of the pandemic, where the difference between the first and third waves reached statistical significance for first-line HCWs. Regarding the change in sleep disturbance within different waves, a previous study reported a nonsignificant trend of increased levels of sleep disturbance from the first to the second wave of the COVID-19 pandemic for HCWs in Taiwan [31]. Our previous work also reported similar findings for first-line and second-line HCWs between the first and second waves [17]. Based on previous evidence, we further identified that sleep disturbance gradually increased. The increased level of sleep disturbance may have resulted from cumulative stressors starting in the first wave of the COVID-19 pandemic. The cumulative risk model conceptualizing stress in developmental psychopathology indicates that individuals who experience multiple stressors are at greater risk for psychopathology than those who experience a single stressor [32]. As the COVID-19 pandemic progressed, the endless workload for COVID-19 treatment, changes in infection control policies, vaccination campaigns, and alterations of original treatment for mental illness may have contributed to great suffering for HCWs, especially first-line HCWs.
On the other hand, several difference in the comparison of demographic information in the third wave may potentially explain the difference of sleep disturbance between first-line and second-line HCWs, such as difference of proportion of COVID-19 infection. Sleep disturbance is pronounced after COVID-19 infection [33], which may be mediated by inflammatory change [34]. We reported a relatively higher proportion of COVID-19 infection among first-line HCWs, which may predominantly affect level of sleep disturbance for first-line HCWs. Furthermore, irregular meals were reported to be associated with sleep disturbance [35], while exercise has moderately positive effects on sleep quality [36]. These evidences are comparable to our findings, where irregular exercise and diet habits may have poor impact on sleep for the first-line HCWs.
Factors associated with the level of sleep disturbance
We reported that sleep disturbance was associated with higher levels of depression and PTSD symptoms and lower levels of subjective wellbeing, mental health and physical health among first-line and second-line HCWs. Previous studies have demonstrated a high prevalence of depression, psychological trauma, and sleep disturbance during the COVID-19 pandemic [37]. Moreover, a comprehensive meta-analysis reported that sleep disturbance was associated with depression and psychological distress among HCWs during the COVID-19 pandemic [38]. Our study echoes previous studies, indicating the association between sleep disturbance and poor mental health. HCWs under the stress of the COVID-19 pandemic faced enormous work stress and potential trauma exposure, leading to sleep problems.
Regarding the association between sleep disturbance and coping strategies measured by the SISQ, social anxiety was identified as a factor. A Taiwanese study reported that excessive worry about COVID-19 was associated with lower psychological wellbeing [39]. Another study reported that sleep disturbance was associated with worries about being infected and fear of COVID-19 [38]. The findings in this study are comparable to the above evidence, indicating a significant association between concerns over COVID-19 and sleep problems. On the other hand, social information was related to sleep disturbance among first-line HCWs, even though it was shown to be a protective factor to actively cope with COVID-19 through obtaining related knowledge [25]. Nevertheless, searching for information may be related to a psychological burden. Excessive media exposure to COVID-19-related news triggers anxiety and stress responses among the general public [40]. Another nationwide survey in Thailand also indicated that individuals with greater exposure to COVID-19-related information had a higher risk of sleep disturbance than those with lower exposure [41]. In addition to enhancing active coping for COVID-19 among HCWs, our study showed that “too much of a good thing” may have been a factor in the emergence of sleep disturbance.
We found that a lower level of vaccine trust was associated with sleep disturbance among first-line HCWs. An epidemiological survey in Denmark showed that vaccine willingness was slightly lower among patients with mental illness than in the general population [42]. Another study revealed that people who hesitated to get vaccinated had slightly higher levels of psychological distress symptoms than non-hesitant individuals [43]. Our study among HCWs is comparable to previous studies. Unexpectedly, a lower level of sleep disturbance was found to be associated with a higher level of preference for natural immunity among second-line HCWs, indicating an antivaccine attitude. As no previous evidence had reported similar findings, we have several assumptions. First, the lower mortality rate of the Omicron variant [44] in the recruitment period may have contributed to this unexpected finding. The lower mortality rate of the current variant may have reduced willingness to vaccinate. Second, we second-line HCWs may have had insufficient knowledge of the risk of the Omicron variant in comparison with first-line HCWs. Nevertheless, further studies are warranted to explore the etiologies behind this.
We found that female sex was significantly associated with a higher level of sleep disturbance. Females have been reported to be at risk of sleep disturbance during the COVID-19 pandemic both in the public [45] and among HCWs [38]. Compared to males, females are socialized to express their emotions more strongly and to have negative views of their health, leading to psychological burdens [46]. Hence, the sex difference in sleep disturbance is supported by previous evidence. On the other hand, MCS was negatively associated with sleep disturbance among first-line HCWs, while this association did not remain significance among second-line HCWs after multiple adjustment. Since MCS was negatively associated with sleep disturbance among second-line HCWs in the univariate regression, we supposed that the divergence might result from the statistic power of sample for second-line HCWs. Finally, we found that younger age was associated with sleep disturbance among second-line HCWs. Previous research presented similar findings in general populations [11], and it may result from the vulnerability of mental health among younger individuals during COVID-19 pandemic [47]. Again, this association was not found in the first-line HCWs. We supposed that it might be contributed by relatively younger age of first-line HCWs than second-line HCWs.
Limitations
We reported several limitations in the current study. First, the multiple comparison of three waves with a cross-sectional design may not have controlled for confounding factors as in the cohort follow-up study. The cross-sectional design of this study limited causal inference for further interpretation. Second, we did not identify the HCWs with shift work due to the limitation of study protocol, where night shift might confound the interpretation of our results. Finally, a single-center study may limit the generalizability and applicability to other HCWs in other nations. Moreover, findings of our study were derived from psychiatric hospital, and clinicians should make interpretation carefully regarding our results.
Conclusions
We identified a gradual increase in sleep disturbance among HCWs across three waves of the COVID-19 pandemic, where the difference in sleep disturbance between the first and third waves reached significance among first-line HCWs. Moreover, we found several predictors of the level of sleep disturbance among HCWs during the COVID-19 pandemic. The clinical implications of the current study demonstrate the urgent need for further intervention for sleep disturbance among HCWs, especially first-line HCWs. Authorities should consider the mental health burden of HCWs when developing or adjusting medical routine or infection control policies during the COVID-19 pandemic. In addition, regular screening for mental health with appropriate questionnaires may be crucial for HCWs to identify sleep disturbances or other metal health problems early. On the other hand, the association between sleep disturbance and vaccine mistrust or excessive worry about COVID-19 also highlights the importance of timely interventions for sleep disturbance. Further studies with longitudinal follow-up can be helpful to better understand the etiologies of sleep disturbance during the COVID-19 pandemic or massive biological disasters.
Supplementary Materials
The Supplement is available with this article at https://doi.org/10.30773/pi.2024.0006.
Selected items of measures in this study
Distribution of marital status among first-line healthcare workers at third wave (N=504)
Distribution of events for psychological trauma among first-line healthcare workers at third wave (N=504)
Distribution of chronic disease (medical) among first-line healthcare workers at third wave (N=504)
Distribution of marital status among second-line healthcare workers at third wave (N=243)
Distribution of events for psychological trauma among second-line healthcare workers at third wave (N=243)
Distribution of chronic disease (medical) among second-line healthcare workers at third wave (N=243)
Predictors for level of sleep disturbance verified with multivariate linear regression with 5,000 bootstrapping samples among first-line healthcare workers in the third wave
Predictors for level of sleep disturbance verified with multivariate linear regression with 5,000 bootstrapping samples among second-line healthcare workers in the third wave
Notes
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Author Contributions
Conceptualization: Dian-Jeng Li, Joh-Jong Huang, Su-Ting Hsu, Frank Huang-Chih Chou, Hui-Ching Wu. Data curation: Su-Ting Hsu. Formal analysis: Dian-Jeng Li, Su-Ting Hsu, Frank Huang-Chih Chou, Hui-Ching Wu. Funding acquisition: Frank Huang-Chih Chou. Investigation: Dian-Jeng Li, Joh-Jong Huang, Kuan-Ying Hsieh, Guei-Ging Lin, Pei-Jhen Wu, Chin-Lien Liu. Methodology: Dian-Jeng Li, Joh-Jong Huang, Kuan-Ying Hsieh, Frank Huang-Chih Chou. Project administration: Frank Huang-Chih Chou. Resources: Frank Huang-Chih Chou, Hui-Ching Wu. Software: Dian-Jeng Li. Supervision: Frank Huang-Chih Chou, Hui-Ching Wu. Validation: Frank Huang-Chih Chou. Visualization: Dian-Jeng Li, Joh-Jong Huang. Writing—original draft: Dian-Jeng Li, Joh-Jong Huang. Writing—review & editing: Frank Huang-Chih Chou, Hui-Ching Wu.
Funding Statement
This study is supported by grants from the National Science and Technology Council, Taiwan (MOST 112-2314-B-280-001).
Acknowledgements
We thank all of the participants in this study and support from the Department of Health, Kaohsiung City and Kaohsiung Municipal Kai-Syuan Psychiatric Hospital.