1 INTRODUCTION

The newly emerged SARS-CoV-2 virus, first discovered in China at the end of 2019 (Lu et al., 2020), had spread around the globe within the first few months of 2020, and instantly posed a great challenge for countries, the society, and people alike. Job loss, home schooling, and social isolation, due to the lockdown restrictions as well as the risk of infecting oneself and loved ones, are examples of what people were confronted with during the beginning of the COVID-19 pandemic in several countries. For many, the COVID-19 pandemic constituted an ongoing source of worrying and stress, which may have led to impaired well-being and sleeping problems, as studies have suggested (American Psychiatric Association, 2020; Rajkumar, 2020; Sønderskov et al., 2020; Umucu & Lee, 2020; Zacher & Rudolph, 2020). Stressful live events usually trigger the use of certain coping strategies to reduce stress and enhance well-being and sleep (Lazarus & Folkman, 1984). However, not all coping strategies are equally efficient in doing so, and some may even have adverse effects. Therefore, in this study, we examined the relationships between worrying, well-being, sleep, and coping strategies, and further investigated the protective role of COVID-19-related coping efforts on the relationship between COVID-19-related worrying, emotional well-being, and sleep.

1.1 Worrying about COVID-19 and its effects on well-being and sleep

The ongoing COVID-19 pandemic has profoundly changed the lives of people worldwide. Suddenly people had to worry about the things they previously considered naturally given. During the early stages of the COVID-19 pandemic, people worried about the possibility of becoming infected with the SARS-CoV-2 virus, their future, and the social or psychological effects of lockdown measures. Among these concerns, the most prominent according to general surveys (American Psychiatric Association, 2020; Qiu et al., 2020; Statista, 2020) were worrying for oneself and even more so for loved ones about contracting COVID-19 and worrying about a possible negative impact on the future economy. Worrying can be defined as an uncontrollable chain of thoughts and images accompanied by negative feelings (Andrews & Borkovec, 1988; Verkuil et al., 2007). In accordance with the transactional model of stress and coping from Lazarus and Folkman (1984), which represents the theoretical framework of this study, worrying about COVID-19 can be considered a form of primary stress appraisal. Primary appraisal stands for the first evaluation of a stressor, represented here by the COVID-19 pandemic; it can be defined as either positive, irrelevant, or dangerous. According to the transactional model, stress occurs when an external or internal demand exceeds a person's individual resources to handle a specific stressor. We assume that the more the people are worried about the COVID-19 pandemic, the more likely they are to regard the COVID-19 as a potentially dangerous stressor, and experience lower well-being. A recent study from Zacher and Rudolph (2020) supports this assumption. The authors found that during the period between March and May in 2020, positive affect levels and life satisfaction decreased in the German population. In addition, Sønderskov et al. (2020) reported lower well-being scores during the COVID-19 pandemic in a Danish sample compared to a sample from 2016, and Sibley et al. (2020) found that people during the lockdown in New Zealand reported higher levels of mental distress.

Stress increases the risk for various health issues (Andrews & Borkovec, 1988; Borkovec et al., 1983; Brosschot & Van Der Doef, 2006), and it also has negative effects on a person's overall well-being (Ganster & Rosen, 2013; Lazarus & Folkman, 1984). Worrying specifically was found in previous research to be associated with higher levels of negative affect and lower levels of positive affect (McLaughlin et al., 2007; Paolini et al., 2006). Also, a recent study by Umucu and Lee (2020) found that COVID-19 stress was associated with decreased well-being. Furthermore, Zacher and Rudolph (2020) also reported higher levels of threat appraisal to be related to lower positive affect, higher negative affect, and lower life satisfaction. Based on these findings and the transactional model of stress and coping (see below), we assume that people who show higher levels of worrying about the COVID-19 pandemic also report lower levels of positive affect and higher levels of negative affect.

Another important indicator of a person's overall health is sleep, since many studies have found positive relationships between sleep and health (Benham & Charak, 2019; Fernandez-Mendoza et al., 2020). However, despite there being many studies stating that stress can have negative effects on sleep (Liu et al., 2016; Sadeh et al., 2004; Yang et al., 2018), literature regarding the association between worrying and sleep is still scarce. Some studies have reported sleep disturbance due to worrying to be a prominent issue. In a study by Marques et al. (2016), 33.2% of the study sample reported frequent sleep disturbance due to worry. In a study by Dregan et al. (2013), 37.9% of the participants who reported having sleep problems named worrying as the main reason. Furthermore, people who attributed sleep disturbance to worrying were also more stressed in general (Kelly, 2003; Marques et al., 2016). Studies on the COVID-19 pandemic and its effects on sleep reported on the one hand positive changes such as later sleep onset-wakeup times, reduced social jetlag, reduced sleep restriction, and longer sleep duration, and on the other hand negative changes such as more sleep disturbances, and a decrease in sleep quality (Blume et al., 2020; Hetkamp et al., 2020; Marelli et al., 2020; Sinha et al., 2020). Lockdown restrictions seemed to have both positive and negative effects on sleep in the general population, yet in a study by Mandelkorn et al. (2020), more than half of the participants (58%) reported that they were unsatisfied with their sleep during the early stages of the pandemic. Kocevska et al. (2020) also found both negative and positive changes in sleep quality. Research investigating the associations between stress, worrying, and sleep during the COVID-19 pandemic is still scarce. Xiao et al. (2020) observed worse sleep quality in the Chinese medical staff who reported more stress. Huang and Zhao (2020) also found that health care workers, who represent a highly stressed occupational group during the pandemic, to report lower sleep quality compared to others, and Kocevska et al. (2020) found that worrying about COVID-19 is associated with worse sleep quality. Fortunately, people are not completely helpless during challenging times because they usually do have ways to cope with situations like the COVID-19 pandemic and its negative effects on well-being and sleep.

1.2 Coping with the COVID-19 pandemic, well-being, and sleep

Since the COVID-19 pandemic represents a source of great stress, people had to find ways to deal with new challenges and negative feelings. However, there are many ways to cope with stressful live events. Coping means any cognitive or behavioural effort to master external or internal challenges, when personal resources are viewed as insufficient (Folkman & Lazarus, 1988). Research has shown that which strategy is best varies in different contexts and therefore strategies cannot generally be differentiated as good and bad (Folkman et al., 1986). Concerning COVID-19, many different coping strategies, such as seeking social support from friends and family, accepting the situation, reacting with humour, avoiding information, or buying stocks are imaginable. The coping strategies most often used to deal with the COVID-19 pandemic have not been sufficiently investigated. Folkman and Lazarus (1980) originally differentiated problem-focused and emotion-focused strategies. In recent research, many researchers have attempted to structure different coping strategies into broader concepts, such as approach and avoidance, or cognitive and behavioural coping strategies (Skinner et al., 2003). In 1997, Carver developed the COPE inventory, which assesses 14 different coping strategies and is currently one of the most used coping questionnaires (Kato, 2015). Although such a detailed differentiation clearly has its advantages, it also makes research on coping very heterogeneous, which is why we focus on four broader concepts of coping strategies that are based on the COPE inventory before (Baumstarck et al., 2017; Zacher & Rudolph, 2020). The four coping styles are problem-focused, meaning-focused, social, and avoidance coping.

Problem-focused coping is directly aimed at the stressor (Carver, 2011), and is positively associated with stress-related growth, quality of life, and positive affect (Göral et al., 2006; Moskowitz et al., 2009; Shermeyer et al., 2019; Smith et al., 2008). Recent studies from Umucu and Lee (2020) and Zacher and Rudolph (2020) confirm these findings for COVID-19-specific problem-focused coping. The association between problem-focused coping and negative affect is ambiguous. Some studies found a negative association (Moskowitz et al., 2009; Shermeyer et al., 2019), while others were not able to confirm these results (Smith et al., 2008; Zacher & Rudolph, 2020). Meaning-focused coping refers to searching for meaning in adversity (Carver, 2011). It is associated with better quality of life, higher levels of positive affect, and lower levels of negative affect (Hofstetter et al., 2005; Moskowitz et al., 2009; Pogrebtsova et al., 2018). Meaning-focused coping regarding COVID-19 was also positively related to general well-being and positive affect (Umucu & Lee, 2020; Zacher & Rudolph, 2020). Social coping refers to seeking social support, and shows no association with positive affect, and a positive association with negative affect in previous research (Moskowitz et al., 2009; Smith et al., 2008). Studies investigating social coping during the COVID-19 pandemic were, for the most part, able to confirm these results (Umucu & Lee, 2020; Zacher & Rudolph, 2020). Through the use of avoidance coping people attempt to escape a current stressful situation (Carver, 2011). To date, avoidance coping seems to be a rather dysfunctional coping style, since it shows negative associations with stress-related growth and positive affect as well as positive associations with depression, anger, and negative affect (Moskowitz et al., 2009; Smith et al., 2008). Zacher and Rudolph (2020) confirmed a positive relationship between COVID-19-related avoidance strategies and negative affect.

Research on coping and sleep is still scarce, especially considering the effects of COVID-19. Morin et al. (2003) found no differences in the usage of problem-focused coping in people with insomnia and good sleepers, although others have reported positive effects of problem-focused coping on sleep duration and quality (Faber & Schlarb, 2016; Morin et al., 2003). Meaning-focused coping was found to be positively related to sleep quality by Hofstetter et al. (2005). To the best of our knowledge, no study has investigated the association between social coping, especially during the COVID-19 pandemic, and sleep. Avoidance coping was related to sleep disturbances, an indicator of sleep quality, in one study (Hoyt et al., 2009) and not related to sleep quality in another study (Hofstetter et al., 2005). Findings regarding the association between avoidance coping and sleep quality are thus so far inconclusive. Sleep duration was not related to avoidance coping in a study by Sadeh et al. (2004). Even though many studies have directly linked different coping strategies with good or bad outcomes for well-being and sleep, it is important to investigate whether coping strategies might efficiently buffer associations between stressors and well-being or sleep outcomes (see Dardas & Ahmad, 2015).

1.3 Coping as a moderator between COVID-19 worry, well-being, and sleep

So far, studies concerning COVID-19-specific coping and well-being during the early stages of the COVID-19 pandemic focused on the associations between stress appraisal, the use of coping strategies, and aspects of well-being. However, to the best of our knowledge, no study has tested possible moderation effects of COVID-19-specific coping on the relationship between COVID-19-related stress or worry and aspects of well-being. It is possible, however, that certain coping strategies could buffer or even enhance the relationships between COVID-19 worry, well-being, and sleep, and therefore may be considered as effective or ineffective coping strategies. For this reason, we investigated moderation effects of the aforementioned four coping strategies in relation to COVID-19 worry, well-being and sleep.

1.4 Aims of the study

Previous research linked greater worrying and stress with worse outcomes for well-being and sleep in general as well as for well-being during the early stages of the COVID-19 pandemic. However, the relationship between stress or worrying due to COVID-19 and sleep has not been sufficiently examined so far. Therefore, the first aim of this study was to investigate whether well-being (positive and negative affect), sleep quality and duration, and changes in sleep quality and duration due to the COVID-19 pandemic were associated with COVID-19 worry during the early stages of the COVID-19 outbreak in Germany in 2020. Based on the transactional model of stress and coping and the results of previous research, we expect higher levels of worrying about COVID-19 to be associated with greater negative affect, lower positive affect, and worse sleep parameters (poor sleep quality, short sleep duration, negative changes in sleep quality, and duration).

The second aim of the study was to examine which COVID-19-specific coping strategies were used most often during the early stages of the COVID-19 pandemic. The third aim was to investigate how COVID-19-specific coping relates to well-being and sleep. Based on previous findings, we expect positive relations for problem- and meaning-focused coping with well-being and sleep. In addition, we expect social coping to be positively related to negative affect and to be unrelated to positive affect. Regarding the association between social coping and sleep, no predictions could be postulated. We expect greater avoidance coping to be associated with lower well-being. Even though research about avoidance coping and subjective sleep quality is scarce and inconclusive, we assume avoidance coping might further be related to worse sleep quality and a negative change in sleep quality due to the COVID-19 pandemic, since it was found to be related to sleep disturbances, an aspect of sleep quality, in a study before and also shows negative relationships with various subjective well-being measures. No associations between avoidance coping and sleep duration or changes in sleep duration due to the COVID-19 pandemic are expected.

The fourth aim of the study was to explore whether the four coping strategies examined were useful in dealing with COVID-19. Therefore, the possible moderating effects of the coping strategies on the relationship between COVID-19 worries and well-being (positive and negative affect) as well as on the relationship between COVID-19 worries and sleep (sleep quality and duration) will be investigated.

2 METHODS

The study was approved by the institutional Ethics Committee and all participants gave informed consent. Data collection took place from 1 to 19 April 2020, during which time lockdown restrictions were in force in Germany. The nationwide lockdown began on 22 March and lasted until 3 May. Schools, stores (with the exception of supermarkets, drug stores, and pharmacies), as well as many nonessential companies were closed. Employees were asked to work at home. Travelling was restricted and international borders were closed. Furthermore, a contact ban was imposed, which only committed contact to only one other person apart from one's own household outdoors.

2.1 Participants

A total of 692 participants completed the online questionnaire, that was distributed via social media and mailing lists in Germany. Three outliers and 24 underaged participants were excluded. The remaining sample consisted of 665 participants, whose age ranged from 18 to 73, with a mean age of 36 (SD = 14). Of the total study sample, 53.8% were female (one person reported as diverse); 47.5% reported having a university degree; 19.6% completed vocational training; 57.7% were currently employed, and worked for more than 20 h a week, 14.3% worked less than 20 h due to the COVID-19 pandemic, 8.6% reported being unemployed or working less than 20 h, 16.6% were students and 2.9% reported being retired. Participants' professions represented a variety of different work sectors (e.g., finance, administration, healthcare, security, education, science, or building industry). None of the participants were tested positive for COVID-19 infection; 42 participants reported having or had COVID-19 like symptoms, but not a test result. Seventy-seven percent of the study sample reported good or very good sleep quality during the beginning of the COVID-19 pandemic, and 78.8% of the study sample slept on average seven or more hours a night. Most of the study sample reported no changes in sleep quality (65.2%) or duration (52.9%). A decline in sleep quality was reported by 23.3%, and an improvement in sleep quality was reported by 11.6% of the study sample. Similarly, sleep duration decreased for 15.3% of the participants and increased for 31.8%.

2.2 Measurement instruments

The online questionnaire included questions about demographics, current employment status, and the following measures.

2.2.1 Worry about COVID-19

At the time of the inquiry no published questionnaire assessing COVID-19 worries was available. On the basis of a study in which items about worrying were adapted to the Ebola crisis in 2014 (Thompson et al., 2017), we created six items to assess three important aspects of worrying about COVID-19, which have already been described. Two items measured worrying about a possible infection (I am scared of getting infected by the COVID-19 pathogen, the possibility of contracting COVID-19 is bothering me), two items measured worrying about the future (I worry about my future because of the current COVID-19 pandemic, the possible consequences due to the COVID-19 pandemic are stressing me), and two items measured worrying about lifestyle limitations and burdens (I feel constricted in my way of life, my life changed significantly due to the COVID-19 pandemic). Participants were asked to rate on a 5-point Likert scale how well every item applied to them (1 = not at all, 5 = exactly). To test dimensionality, we performed a confirmatory factor analysis with three components (consisting of two items as described above) and one higher order factor using LISREL. This second-order factor model provided a good fit with χ2 (6) = 10.50, p = 0.11 (RMSEA = 0.033; CFI = 0.998; RFI = 0.988). Second-order factor loadings (0.92, 0.51, 0.46) were significant (p´s < 0.001). Because we were interested in a global measure of COVID-19 worry, one mean score of all six items was computed (α = 0.75).

2.2.2 Positive and negative affect

To assess the positive and negative affect, the Positive and Negative Affect Schedule (PANAS) was used (Breyer & Bluemke, 2016; Watson et al., 1988). The PANAS measures positive and negative affect with 10 items each, which can be summarized into one mean score each for positive and negative affect. Every item represents an adjective, and the participants rate how intensely they felt this specific emotion or feeling during the past two weeks. Items are answered on a scale from 1 (not at all) to 5 (extremely). The reliability and validity of the PANAS are satisfactory (Breyer & Bluemke, 2016). The reliabilities of positive and negative affect for this study were satisfactory (α = 0.85; 0.82, respectively).

2.2.3 Sleep duration and subjective sleep quality

Sleep duration and subjective sleep quality were measured using items of the Pittsburg Sleep Quality Index (PSQI; Backhaus & Riemann, 1996; Buysse et al., 1989). The PSQI is a widely used instrument to measure sleep quality—consisting of seven subscales—which offers good validity and reliability (Backhaus & Riemann, 1996). For sleep duration, the participants rated how long, on average, they slept in the past two weeks. Unlike in the original questionnaire, we decided to use sleep duration as a continuous variable to not lose variance. Subjective sleep quality was measured with the ‘sleep quality’-subscale, asking participants to rate their overall sleep quality for the past two weeks. We used an eight-point format, and added two additional items (on most days in the last two weeks I slept well; my sleep in the last two weeks was not restorative). All the three items were correlated (r = 0.76 and above), and therefore, aggregated into one mean score, with higher scores indicating better subjective sleep quality. The reliability of sleep quality for this study was good (α = 0.93).

2.2.4 Change in sleep quality and sleep duration

We used two single-item measures of sleep quality and duration from the PSQI (Buysse et al., 1989) and adapted the response format to assess subjective change due to the COVID-19 pandemic. Participants were asked to rate on a 5-point-Likert scale their sleep quality (Compared to the time before the COVID-19 pandemic, I now sleep 1 = a lot worse, 5 = a lot better) and their sleep duration (Compared to …. I now sleep 1 = much shorter, 5 = much longer).

2.2.5 COVID-19-specific coping

Coping strategies were measured with the Brief-COPE from Carver (1997), which was adapted to the consequences of the COVID-19 pandemic. The Brief-COPE is a short form of the COPE inventory, one of the most frequently used coping questionnaires (Kato, 2015), and includes 14 different coping strategies, each measured by two items (Carver, 1997). Two items assessing self-blame were not included, due to inadequacy in the context of the COVID-19 pandemic. Two items measuring religious coping were also excluded, as they did not fit to our proposed higher-order coping strategies. All the 12 measured coping strategies of the Brief-COPE were reduced into four higher order coping strategies, similar to those used by Baumstarck et al. (2017), Litman (2006), and Zacher and Rudolph (2020). Acceptance, positive reframing, and humour items were subsumed under the construct of meaning-focused coping (α = 0.68). Problem-focused coping included active coping and planning items (α = 0.75). Self-distraction, behavioural disengagement, denial, and substance use items were summarized as avoidance coping (α = 0.53), and instrumental support, emotional support, and venting items as social coping (α = 0.77).

2.2.6 Control variables

Peoples' preferred sleep time and sleep quality are strongly influenced by individuals' chronotype. Late chronotypes tend to have later sleep onset and offset times and report a lower sleep quality compared to early chronotypes (Roenneberg & Merrow, 2007; Roeser, Meule, et al., 2012). Late chronotypes also show lower levels of well-being and higher levels of stress compared to early chronotypes (Buschkens et al., 2010; Roeser, Obergfell, et al., 2012). Therefore, we included chronotype as a control variable. Chronotype was measured using the Composite Scale of Morningness (CSM; Randler, 2014; Smith et al., 1989), which is a widely used chronotype self-reporting questionnaire to assess daily preferences for activity. The CSM consists of 13 items, which can be aggregated into one global score. Higher values indicate a greater tendency toward morningness. Values range from 22 to 55. According to Randler (2008), the CSM scale offers good psychometric properties. Reliability was α = 0.90.

In addition, we included age, gender, and education as control variables because they showed significant correlations with the relevant variables (see Table 2). Education was operationalized as a dichotomous variable that differentiated between participants with a university degree and those with less education.

2.3 Statistical analysis

The first and third aim of the study were analysed based on bivariate correlations with a significance level of p < 0.05. The second aim of the study was to investigate which coping strategies were used most often during the start of the COVID-19 pandemic and was tested via a repeated measures ANOVA with Bonferroni-adjusted post-hoc analyses with a significance level of p < 0.05. Moderations effects were investigated using multiple regression analyses. In the first step worrying about COVID-19, control variables (chronotype, age, gender, education), and all four coping strategies served as predictors for the measures of well-being (positive and negative affect) and different sleep characteristics (sleep quality and duration). In the second step interactions between worrying about COVID-19 and the four coping strategies were included. Since response formats differed between variables, all predictors were transformed into z-scores to avoid problems of multicollinearity (Cohen et al., 2003). We expected chronotype to be associated with worrying, coping strategies, sleep, and well-being, and therefore included chronotype as a control variable. All calculations were conducted using IBM SPSS Statistics, version 26, and Microsoft Excel 2016. Descriptive statistics of all relevant variables are shown in Table 1.

TABLE 1. Descriptive statistics M SD Min Max Skewness Kurtosis COVID-19 worry 2.63 0.79 1.00 5.00 0.28 −0.27 Positive affect 2.88 0.69 1.00 5.00 0.15 −0.09 Negative affect 2.12 0.67 1.00 4.40 0.48 −0.12 Sleep quality 5.79 1.86 1.00 8.00 −0.72 −0.49 Change in sleep quality 2.86 0.71 1.00 5.00 −0.11 1.52 Sleep duration 7.49 1.12 4.00 12.00 −0.04 1.02 Change in sleep duration 3.21 0.83 1.00 5.00 0.05 0.31 Chronotype 37.55 7.25 16.00 55.00 −0.25 −0.21 Problem-focused coping 4.31 1.04 1.00 6.00 −0.65 0.25 Meaning-focused coping 4.41 0.88 1.83 6.00 −0.37 −0.37 Social coping 2.70 1.02 1.00 6.00 0.36 −0.51 Avoidance coping 2.27 0.63 1.00 4.75 0.61 0.87 3 RESULTS

Based on the four aims of this study, the results are presented in four sections. First, correlations of worrying about COVID-19 with positive and negative affect and sleep variables are reported. Second, we explore the usage of the four different investigated coping-strategies during the early stages of the COVID-19 pandemic. Third, how these coping strategies relate to well-being and sleep are described. Finally, we report whether these coping-strategies act as moderators for the relationship between COVID-19 worry with positive and negative affect and sleep.

3.1 Worrying about COVID-19, well-being, and sleep

All bivariate correlations are shown in Table 2. As expected, worrying about COVID-19 showed a positive correlation with negative affect, and negative correlations with positive affect and all sleep variables. The higher the reported worry about COVID-19, the higher the negative affect, and the lower the positive affect. Highly worried participants also reported lower sleep quality, shorter sleep duration, and a negative change in sleep quality and duration compared to weeks before the emergence of COVID-19. The correlations differ in size, but the pattern of the results confirmed our hypotheses.

TABLE 2. Bivariate correlations 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 1 COVID-19 worry 1 – – – – – – – – – – – – – – 2 Positive affect −0.18*** 1 – – – – – – – – – – – – – 3 Negative affect 0.55*** −0.16*** 1 – – – – – – – – – – – – 4 Sleep quality −0.30*** 0.36*** −0.36*** 1 – – – – – – – – – – – 5 Change in SQ −0.27*** 0.23*** −0.28*** 0.58*** 1 – – – – – – – – – – 6 Sleep duration −0.08* 0.02 −0.04 0.29*** 0.20*** 1 – – – – – – – – – 7 Change in SD −0.10* 0.09* −0.12** 0.31***