The relationship between stress, sleep
related worry and sleep impairment in
individuals with insomnia
K.B. van Zanten
University of Amsterdam
Department of Clinical Psychology
Master’s thesis by K.B. van Zanten
Student number: 5879264
Supervision by dr. J. Lancee
Words: 4812
Abstract
The cognitive model of Harvey (2002) proposes that sleep related worries and stress have a direct relationship with sleep impairment in people with insomnia. Nevertheless, not much research has been done investigating the relationship between daily stress, worry and sleeping impairment. In this study, fifty-two participants filled out questionnaires regarding worrying, sleep, depression and anxiety symptoms. In addition, this study measured worry twice a day. Participants filled out a morning and evening diary for seven days in a row. In the morning diary sleep and sleep related worries were measured and in the evening diary sleep, daytime worries and daily stress were measured. Six multilevel analyses were carried out with sleep efficiency, total sleep time, total awake time, sleep onset latency, wake after sleep onset and terminal wakefulness as the dependent variables. The independent variables were daytime stress, daytime worries and nighttime worries. Based on the results, only nighttime worries were related with sleep impairment (p<.05). Stress and daytime worries had no significant relation with sleep impairment in almost all of the analyses. These results contradict the model of Harvey (2002) and suggest that only nighttime worrying is the main factor in maintaining insomnia and not daytime worrying or stress, as the model suggests. Implementing these results, can make CBT-I more effective.
Introduction
Insomnia is a common disorder which effects many people and has a great influence on their lives. Nine to fifteen percent of the general population meets one of the insomnia criteria of the DSM-IV and eight to eighteen percent suffers from the daytime consequences of insomnia. Six percent of the general population meets the criteria for the diagnosis of Insomnia (Ohayon, 2002). According to the DSM-IV, this diagnosis consists of the following problems: concentration, fatigue, and impaired cognitive functioning caused by sleeping problems for at least three months (American Psychiatric Association, 2013).
Insomnia has great consequences for an individual and for the society at large. Individuals with insomnia suffer from mood and anxiety disorders, reduced productivity, and show higher absence from work compared to healthy sleepers and as consequence is associated with high societal costs (Daley, Morin, LeBlanc, Grégoire, Savard & Baillargeon, 2009). The resulting excessive sleepiness is associated with major depressive disorder, anxiety disorder, presence of heart disease, alcohol dependency and a greater likelihood for having a car accident (Ohayon, 2012). It is evident that sleep impairment can have great psychological and physiological consequences.
Because of these serious consequences, it is in the public interest to see how insomnia can be treated effectively. Cognitive Behavior Therapy (CBT-I) for Insomnia is proven to be the best treatment for insomnia (Jacobs, Pace-Schott, Stickgold, & Otto, 2004; Riemann & Perlis, 2009), which focuses on behavior and cognitions about sleep (Morin, LeBlanc, Daley, Gregoire, Mérette, 2006). Jacobs, Pace-Schott, Stickgold, and Otto (2004) found that CBT-I and pharmacotherapy are equally effective. However some individuals do not benefit from CBT-I (Morin & Benca, 2012). Nevertheless, CBT-I is preferred over biological therapy (i.e. pharmacotherapy, bright light therapy) (Milner, & Belicki, 2010). Further researching insomnia and its components is required, to make CBT-I more effective. More specifically, it is important to investigate the factors that aggravate and maintain insomnia, so that they can be implemented in CBT-I treatment.
The model of Harvey (2002) is a commonly used model, when analyzing the factors that are associated with sleep impairment, see Figure 1. According to this model, negative cognitive activities causes arousal and distress, which cause selective attention and monitoring, which in turn leads to a distorted perception of sleep. This model also suggests that both daytime and nighttime negative cognitive activities (during the day or in bed) fuel and maintain insomnia. With reference to this model, negative cognitive activities play a central role in aggravating and maintaining insomnia.
Negatively toned activities, also referred as worrying, lead to a direct negative influence on sleep quality (Borkovec, 1979; Borkovec et al., 1998; Takano, Lijima and Tanno, 2012). Worrying causes hyper arousal and has a negative influence on a person’s mood (Kirkegaard, Thomsen, Mehlsen, Christensen & Zachariae, 2003). Moreover, worrying and repetitive thoughts seem characteristic for insomnia (Carney, Harry, Moss & Edinger, 2010). The relationship between insomnia and cognitive activities thus seems evident. However, no study has been done in investigating the difference between the relation with daytime negative cognitive activity and sleep impairment or the relation between nighttime negative cognitive activity and sleep impairment.
Furthermore, there are some caveats to be mentioned about the previous referred studies. Not all measurements were done properly and not always a large effect with worrying and sleep was found. When looking at Takano et al. (2012), at least three important critiques can be made. First of all, the effect of worrying on sleep was non-significant. Secondly, sleep was measured non-daily and thirdly, no sleep diary was used, even though it is the golden standard when measuring sleep (Buysse, Ancoli-Israel, Edinger, Lichtstein & Morin, 2006). Additional, Kirkegaard et. al (2003) measured sleep with a general sleep questionnaire, but did not use a sleep dairy. Carney et al. (2010) did measure sleep daily with a sleep dairy, however the effect size was limited. The influence of worrying on sleep was also researched with a sleep diary in the thesis of van Baar (2013), however no effect was found. According to the author, this was caused by the questionnaire, which was not specific enough for worrying. To complement previous studies, this study will measure worrying twice a day with a specific questionnaire about worrying: The Anxiety and Preoccupation about Sleep Questionnaire (APSQ) by Jansson-Fröjmark, Harvey, Lundh, Norrel-Clarke and Linton (2010). Moreover, sleep is measured by a sleep diary.
The model of Harvey (2002) suggests that both daytime worries and nighttime worries can cause sleep impairment, see Figure 1. However, the studies by Weise, Ong, Tesler, Kim and Roth (2013) and Wiclow & Espie (2000) both show that only nighttime worrying was related with sleep impairment, however general worrying was not made evident. Moreover, the thesis of van Baar (2013) did not found a relation between daytime worrying and sleep impairment. This study aims to investigate if sleep related worries during the day and during the nighttime are related to sleep impairment.
Fig. 1
Application of the cognitive model to the night and the day. Adapted from “A cognitive model of insomnia” van Harvey, A. G., 2002, Behaviour Research and Therapy, 40, p. 874. Copyright 2002 Elsevier Science Limited.
In addition to the above, other factors can be of importance as well, when looking at the model of Harvey (2002). According to this model, arousal and distress also play a central role in developing and maintaining insomnia. Stressful life events seem to have a negative impact on sleep quality, especially for individuals with a tendency to worry (Guastella & Moulds, 2007). It seems that high levels of stress are related with sleep impairment and lower sleep efficiency (Åkerstedt, Kecklund & Axelsson, 2007; Kant et. al., 1995). Zoccola, Dickerson and Lam (2009) tested if worries and stress are related with insomnia. The results showed that people, who tend to worry more often, are more perceptive of stress and therefore have more impaired sleep after a stressful life event. In contrast, some studies found that stress moderated worrying or only found an effect of stress on sleep when mediated by worrying (Brosschot et. al., 2005; Hall et. al., 2000; Morin, Rodrigue & Ivers, 2003;). Moreover, researchers found that stress can incite sleep, which gives the brain an opportunity to recuperate from the stressful life event (Haynes, Adams & Franzen, 1981; Philbert et. al., 2011). These two studies found a restorative effect of stress on sleep. The cognitive model of insomnia suggests that worrying stimulates stress and arousal. However, these last studies suggest the opposite: stress stimulates sleep. Based on these studies, stress seems to be related with insomnia, but it remains unclear what the exact relation between stress, worrying and insomnia is.
Current study
This study investigates the relation between stress, daytime worry and nighttime worry and sleep measured on a daily basis. The DSM-IV states that the sleeping problems cannot occur during the course of another mental disorder, therefore anxiety and depression are controlled for in this study. The expectations are based on the model of Harvey (2002). Based on this model the first hypothesis is that higher levels of reported stress are associated with lower levels of sleep. The second hypothesis is that higher levels of daytime worrying are associated with lower levels of sleep. And the third hypothesis is that nighttime worrying also is related with lower levels of sleep.
Method
Procedure
This study was part of a larger randomized waitlist controlled trail, investigating the efficacy of online CBT-I. CBT-I was free of charge if participants completed the seven day diary. Participants were recruited by a popular-scientific Dutch insomnia website. If participants were interested to join this study, they could provide their email-address on www.insomnie.nl and complete the online Dutch baseline questionnaire, the Sleep-50 (Spoormaker et. al., 2005). When participants completed the informed consent, they were asked to fill in a sleep dairy for seven executive days (sleep dairy was automatically sent at 6.00 AM and evening diary was automatically sent at 7.00 PM). After they completed at least six dairies, participants were randomized in the online CBT-I condition or waitlist condition. Because the treatment was conducted online, participants could follow the treatment at home, trough the website www.slaapgezonder.nl. Participants also gained access to the digital version of the insomnia self-help book, when completing six weeks of CBT-I treatment. After the post-test was filled out, the waitlist condition also received free treatment. All participants were asked to complete a follow-up questionnaire, after three months and after six months.
Participants
Ninety-eight participants were invited to participate in this study. Ninety-three started the online questionnaire via the website www.insomnie.nl from August to October 2013. Inclusion criteria were: fulfilling insomnia disorder based on DSM-V criteria. Furthermore, the participants were required to have an email address and had to be older than 18 years. Exclusion criteria were sleep apnea, suicide ideation, psychoses or schizophrenia, pregnant women or woman that were breastfeeding their child, working nightshifts, already received cognitive behavioral therapy for treating insomnia, started any form of psychotherapy in the last six months, and an alcohol or drug addiction. Twenty-five participants were excluded from the study (see Figure 2 for a flowchart). Moreover participants were required to complete 12 of the total 14 dairies (two a day). Sixteen 6
participants were excluded due to not completing enough diaries (less than 12 diaries). The final sample consisted of 52 participants, varying between 21 and 77, with a mean age of 47,76. The final sample included 42 woman (80.7%) and 10 men (19.3%), for more baseline characteristics see Table 1.
Fig. 2 Flowchart. Table 1
Demographic characteristics.
n total Response n percentage
Job 52 Yes No 39 13 75% 25% Gender 52 Male Female 10 42 19.3% 80.7%
Medication for sleeping 52 Yes
No
26 26
50.0% 50.0%
In psychological treatment 52 Yes
No
1 51
1.9% 98.1%
Insomnia due to a physical condition 52 Yes No
4 46
11.5% 88.5%
Insomnia longer than three months 52 Yes No 51 1 98.1% 1.9% Invited to participate n = 98 Started questionnaire n = 93 Started diary n = 68 Included in analyses n =52 Excluded, because: - Alcohol/drug dependence (n=3) - Suicide ideation (n=2) - Night shift (n=5)
- Already had treatment (n=12) - Apnea (n=3)
Total excluded (n=25)
Not included because: - Did not finish at least
six diaries (n = 16)
Power
The power was calculated to establish if the effect of this study is an actual effect found in the general population. G*power 3.1 software was used to calculate the number of participants to achieve a sufficient power. We aimed to achieve a power of .8. This study needed at least 55 participants to attaint an effect size of f2 = .15 with a power of .8 (p < .05, two-tailed). After
calculation this study achieved a power of .78 (p < .05, two-tailed).
Measurements
Diary measures
The sleep dairy is a standard instrument to asses and measure sleep. In this study the The Core Consensus Sleep Diary (Buysse, Ancoli-Israel, Edinger, Lichstein & Morin, 2006; Carney, Buysse, Ancoli-Israel, Edinger, Krystal, Lichstein, et al., 2012) was used. With the sleep diary the following variables were recorded; bedtime, final arising time, sleep onset latency (SOL), number of nocturnal awakenings (NWAK), wake after sleep onset (WASO), and daily use of sleep medication. From these variables time in bed (TIB = final arising time), total sleep time (TST = TIB – SOL – WASO) and sleep efficiency ((SE = TST/TIB) x100) were calculated, (table 2). Also Total Time Awake (TTA) was calculated (TTA = SOL + NWAK + WASO).
Table 2
Sleep variables, abbreviations and definitions Sleep Measure Abbreviation Definition
Number of Awakenings NWAK Number of awakenings, excluding the final awakening before the final arising Sleep Quality SQ Subjective sleep quality, typically defined by responses on an ordinal or visual
analog scale Sleep Efficiency
(percentage)
SE Percent of time in bed spent asleep. (percentage) When using sleep diaries, this is
calculated from other self-report variables: TST/TIB x 100
Sleep-Onset Latency SOL How many minutes it takes to fall latency asleep, starting from the moment of intention to fall asleep.
Time in Bed TIB Time in bed, starting from the moment of intention to fall asleep and concluding with the final arising
Total Sleep Time TST Actual time slept. When using sleep diaries, this is typically calculated from other self-report variables (TIB–SOL–
WASO–TWAK)
Terminal Wakefulness TWAK Amount of awake time between the final awakening and the time of getting out of bed
Wake After Sleep Onset WASO Total amount of time awake during the night, excluding SOL and TWAK
Dependent sleep variables based on “Standard Research Assessment of Insomnia” by Buysse et al. 2006, Sleep, 29, No. 9, p. 1162.
Stress was measured with a subscale of the Depression Anxiety Stress Scale 21 (DASS-21) (Lovibond & Lovibond, 1995). The stress scale exist of seven items and is reliable with α =.89, inter-item range of r =.53 (.34-.77) and has a test-retest correlation of .85 (de Beurs, van Dyck, Marquenie, Lange, & Blonk, 2001).
For measuring insomnia related worrying the Anxiety and Preoccupation about Sleep Questionnaire (APSQ) by Jansson-Fröjmark, Harvey, Lundh, Norrel-Clarke and Linton (2010) was used. The worry scale contained six items and is very reliable α =.91.
Baseline Questionnaire
Depression was measured by the Center for Epidemiologic Studies Depression Scale (CES-D) questionnaire by Radloff (1977). The questionnaire exists of 20 items focused on the mood of the participant of last week. The CES-D (Dutch version) has a good reliability of α = 0.80-0.90 and the test-retest correlation is .90 (Beekman, van Limbeek, Deeg, Wouters & van Tilburg, 1994; Demirchyan, Petrosyan & Thompson, 2011; Radloff, 1977).
Anxiety was measured by the anxiety subscales of the Hospital Anxiety and Depression Scale (HADS-A) developed by Zigmond & Snaith (1983). The subscale exists of seven items with a sum score of 0-21, with a cutoff score of eight. The reliability of the HASD-A is α = 0.80 to 0.84 (Spinhoven, Ormel, Sloekers, Kempen, Speckens & van Hemert, 1997).
Subjective sleeping problems were measured by the Insomnia Severity Index (ISI) of Morin (1993). This questionnaire exist of seven five-point scale items, with a sum score of 0-28 and a cutoff score of 15. The ISI has a good intern consistency of α = .74 (Bastien, Vallières & Morin, 2001).
To exclude participants suffering from sleeping apnea, the apnea subscale of the SLEEP-50 was used (Spoormaker et. al., 2005). This subscale exists of eight items. The sum score varies from 8-32 with a cutoff score of 15. The subscale has a good reliability of α = 0.85 and inter-item range of r=.65 - .89.
2.3 Statistical Analyses
First Pearson correlations were calculated for CES-D, HADS, SE, SOL, WASO, TAT, TWAK, daytime worries, nighttime worries and stress, see Table 4. For testing the relationship between the dependent and independent variables a multilevel linear regression was used (Hox, 2002). The reason for applying a multilevel analysis is that within this study two levels can be identified. The first level was time and the second level were differences on an individual level. Another advantage of multilevel modelling is that it can easily handle missing data. Therefore, participants who did not finish all their diaries for all seven days could still be included in the analysis.
NWAK, SOL, TWAK, TTA, and WASO were not normally distributed and to correct for this skewed distribution the variables were log transformed. Even after log-transforming, one participant remained an outlier (Z-score of -6.19 on TTA). This score was excluded from the analysis. Furthermore some participants had difficulty using the 24-hour time format, as example they confused 23:00 PM with 11:00 AM. This was corrected for eight cases. The analyses were executed with the normal scores and the Z-scores. This was done in order to get unstandardized (b) and standardized (β) coefficients. Analyses were conducted on 1) all participants and 2) on only the people who finished there dairy on time. In the result section, only the data of the people, who finished the diary on time, are reported. For the results of the full group, please see supplemental table 2-6. Severity index, anxiety and depression are controlled for in this study. Significant levels were set at p<.05 (two-tailed).
Results
Table 3Mean scores for sleep measures, anxiety, depression, stress and insomnia severity index.
Variables n Minimum Maximum Mean (SD) Cutoff score
ISI 52 9.00 26.00 18.08 (.21) >15 CES-D 52 3.00 38.00 17.25 (.46) >15 HADS 52 1.00 19.00 6.87 (.22) >15 TST 363 .00 710.00 331.74 (127.49) SE 363 .00 100.00 66.46 (23.17) <85 SOL 363 .00 148.80 82.20 (33.00) WASO 363 .00 151.20 87.00 (40.80) TAT 363 .00 165.00 126.00 (20.40) TWAK 363 .00 151.20 88.20 (39.00) Nighttime worry 361 10.00 50.00 26.00 (10.18) Daytime worry 358 10.00 50.00 26.74 (9.37) Stress 358 7.00 28.00 12.39 (4.67)
Note. The time variables SOL, WASO, TAT, TWAK, and TST are noted in minutes. TST = total sleep time; SOL = sleep onset latency; SE = Sleep efficiency; WASO = Wake after sleep onset; TAT = Total awake time; TWAK = Terminal wakefulness.
Table 4
Correlations between diary variables.
Variables CESD HADS SOL WASO TAT TWAK
Nighttime worry .241** .345** .262** .074 .187** .043 Daytime worry .189** .258** .119* .045 .034 -.022 Stress .319** .435** .094* -.001 -.104* -.097*
Note. ** = p<.01 (two-tailed), * = p<.05 (two-tailed). TST = total sleep time; SOL = sleep onset latency; SE = Sleep efficiency; WASO = Wake after sleep onset; TAT = Total awake time; TWAK = Terminal wakefulness.
Multilevel Linear Regression analysis
Only nighttime worry was significantly associated with sleep efficiency (β = -.048, b = -1.12, p < .001), see Table 5. No significant relationship was found between stress, daytime worry and sleep efficiency, day, anxiety, depression and insomnia severity index. For this model 9.85% of the variance was explained on the individual level and 19.06% was explained on the day level. Together this model explained 15.87%.
Table 5
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of sleep efficiency.
Sleep Efficiency β (SE) b (SE) 95% CL Constant 94.537(10.880)* [.269 - 2.153] Day -.003(.023) -.067(.543) [-.049 - .044] ISI -.043(.025) -1.000(.577) [-.093 - .007] Anxiety .003(.031) .059(.709) [-.059 - .064] Depression .023(.015) .533(.351) [-.007 - .053] Nighttime worrying -.048(.007)** -1.12(.172)** [-.063 - -.034] Daytime worrying .010(.009) .229(.203) [-.007 - .027] Stress .012(.015) .271(.358) [-.018 - .042] Note. ** = p<.01 (two-tailed), R2= 15.87%
Only nighttime worry was significantly associated with total sleeping time (β = .049, b = -6.298, p<.001), see Table 6. No significant relationship was found between daytime worrying, stress and day, anxiety, depression, insomnia severity index and total sleeping time. For this model 5.73% of the variance was explained on the individual level and 21.43% was explained on the day level. Together this model explained 15.77%.
Table 6
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of total sleeping time.
Total sleeping time
β (SE) b(SE) 95% CL Constant 455.850(62.260)** [-.006 – 1.953] Day .035(.024) 4.470(2.994) [-.012 - .082] ISI -.025(.026) -.3122(3.314) [-.077 - .028] Anxiety -.005(.032) -.636(4.067) [-.069 - .059] Depression .020(.016) 2.519(2.011) [-.012 - .051] Nighttime worrying -.049(.007)** -6.298(.949)** [-.064 - -.035] Daytime worrying .005(.009) .625(1.115) [-.012 - .022] Stress .013(.015) 1.683(1.963) [-.017 - .044] Note. ** = p<.01 (two-tailed), R2 = 15.77%.
Only nighttime worry was significantly associated with sleep onset latency (β = -.035, b = .035, p < .001), see Table 7. No significant relationship was found between stress, daytime worrying and days, anxiety, depression, insomnia severity index and sleep onset latency. For this model 6.44% of the variance was explained on the individual level and 11.99% was explained on the day level. Together this model explained 11.34%.
Table 7
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of sleep onset latency.
Sleep onset latency
β (SE) b(SE) 95% CL Constant -.683(.546) [-1.776 - .410] Day -.020(.020) -.020(.020) [-.061 - .020] ISI .009(.029) .009(.029) [-.050 - .068] Anxiety .030(.036) .029(.036) [-.043 - .102] Depression -.016(.018) -.016(.018) [-.051 - .019] Nighttime worrying .035(.007)** .035(.007)** [.022 - .049] Daytime worrying -.001(.008) -.001(.008) [-.017 - .015] Stress -.017(.014) -.017(.014) [-.045 - .011] Note. ** = p<.01 (two-tailed), R2 =11.34%. 12
Nighttime worry (β = .018, b = .018, p = .02) and insomnia severity index (β = .062, b = .062, p = .005) was significantly associated with wake after sleep onset, see Table 8. No significant relationship was found between daytime worrying, stress and wake after sleep onset, day, anxiety and depression. For this model 42.91% of the variance was explained on the individual level and 2.38% was explained on the day level. Together this model explained 14.87%.
Table 8
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of wake after sleep onset.
Wake after sleep onset
β (SE) b(SE) 95% CL Constant -.956(.399) [-1.755 - -.156] Day -.013(.025) -.013(.025) [.062 - .036] ISI .062(.021)** .062(.021) [.020 - .103] Anxiety -.010(.026) -.010(.026) [-.062 - .041] Depression -.024(.013) -.024(.013) [-.050 - .002] Nighttime worrying .018(.008)* .018(.008) [.003 - .036] Daytime worrying .001(.009) .001(.009) [-.017 - .019] Stress -.005(.016) -.005(.016) [-.037 - .026]
Note. * = p<.05 (two-tailed), ** = p<.01 (two-tailed), R2 = 14.87%.
No significant relationship was found between stress, daytime, nighttime worrying and terminal wakefulness, day, anxiety, insomnia severity index and depression, see Table 9. For this model 13.75% of the variance was explained on the individual level and 1.74% was explained on the day level. Together this model explained 4.08%.
Table 9
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of terminal wakefulness.
Terminal Wakefulness β (SE) b(SE) 95% CL Constant -.462(.397) [-1.257 - .333] Day .037(.026) .037(.026) [-.015 - .088] ISI .024(.021) .024(.021) [-.018 - .065] Anxiety -.004(.025) -.004(.025) [-.054 - .047] Depression .003(.013) .003(.013) [-.022 - .028] Nighttime worrying .005(.008) .005(.008) [-.011 - .022] Daytime worrying -.009(.010) -.009(.010) [-.028 - .010] Stress -.005(.017) -.005(.017) [-.038 - .028] R2 = 4.08%.
Nighttime worrying (β = .039, b = 5.111, p < .001) and insomnia severity index (β = .049, b = 2.300, p = .026) was only significantly associated with total awake time, see Table 20. No significant relationship was found between daytime worrying, stress and total awake time and day, anxiety, depression. For this model 31.82% of the variance was explained on the individual level and 12.92% was explained on the day level. Together this model explained 18.22%.
Table 10
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of total awake time.
Total awake time
β (SE) b(SE) 95% CL Constant -.991(.410)* [-1.813 - -.168] Day -.002(.024) -.002(.024) [-.051 - .046] ISI .049(.021)* .049(.022)* [.006 - .092] Anxiety .007(.026) .007(.026) [-.046 - .060] Depression -.023(.013) -.023(.013) [-.050 - .003] Nighttime worrying .039(.008)** .039(.008)** [.024 - .054] Daytime worrying -.010(.009) -.010(.009) [-.027 - .008] Stress -.020(.016) -.020(.016) [-.051 - .010]
Note. * = p<.05 (two-tailed), ** = p<.01 (two-tailed), R2 = 18.22%.
Discussion
This study investigated the relationship between stress, daytime worry and nighttime worry and sleep. In line with our expectations, nighttime worry was related with sleep impairment. However, stress and daytime worry were not related with any sleep measurement, which was unexpected. Nighttime worry was related with lower sleep efficiency, lower sleeping time, longer sleep onset, longer wake after sleep onset and longer terminal wakefulness; however this was not the case for daytime worry and stress. Based on the model of Harvey (2002) the first hypothesis stated that higher levels of reported stress are associated with lower levels of sleep. The second hypothesis stated that a higher level of daytime worrying was associated with lower levels of sleep. The third and final hypothesis stated that nighttime worrying also is related with lower levels of sleep. Only the third hypothesis can be supported by the results of this study.
With reference and stipulated in the introduction, to date studies have generally failed to measure worry on a day to day basis. Based on this study, it seems that only a relation between insomnia and nighttime negative cognitive activities was found, when measuring worry on a day to day basis. Based on our results nighttime negative cognitive activities seem of more importance then daytime negative cognitive activities. This contradicts previous research. The reviews of Borkovec (1979) and Borkovec et al. (1998) state that worrying has a negative influence on sleep. Two other studies (Kirkegaard Thomsen, Mehlsen, Christensen & Zachariae, 2003; Takano, Lijima & Tanno, 2012) found a negative relation between sleep and worrying. Takano et. al. (2012) found that worrying had a direct negative influence on sleep quality. A relation between insomnia and cognitive activities seems evident, but was not found in this study. Perhaps, when making a distinction between daytime worries and nighttime worries, it becomes clear that only nighttime worries are of relevance when investigating and treating insomnia.
The results of this study suggest that daytime worries are of lesser importance than the model of Harvey (2002) stated. This leads to disqualifying the second hypothesis that was based on the cognitive model of Harvey (2002). This model suggests that daytime worrying is equally important as nighttime worries in causing and maintaining sleep impairment. The results of this study do not support this theory. Therefore our first hypothesis is supported, but our second hypothesis was not made evident en needs to be disqualified. Opposed to our second hypothesis, daytime worrying may not be directly related to sleeping impairment based on the results of this study. The results of the thesis of van Baar (2013) and the results of the study of Weise, Ong, Telser, Kim and Roth (2013) also support this conclusion.
Another unexpected result is the lack of a relation between stress and sleep impairment. Our third hypothesis implied that higher levels of reported stress are associated with lower levels of sleep. This hypothesis is not supported by our results. The relation between stress and sleep impairment remained unclear, as mentioned in the introduction. Some previous research showed that stress had a negative relation with sleep (Guestella et. al., 2007; Hall et. al., 2000; Morin et. al., 2003; Brosschot et. al., 2005). Hence two of these studies only found an effect of stress on sleep, when mediated by negative toned cognitions (Guestella et. al., 2007; Morin et. al., 2003; Brosschot et. al., 2005). However, two studies suggested that stress incited sleep (Haynes, Adams & Franzen, 1981; Philbert et. al., 2011). This contradicted the cognitive model of Harvey (2002) which suggested that distress is preceded by negative toned cognitions and both are related to sleeping problems. In this study stress did not seem to explain the relation with sleep impairment in any model. This is a remarkable observation. These results suggest that stress has no relationship with sleep impairment. Clearly, more research is needed to investigate the relation between stress and sleep impairment.
Before the results are interpreted, there are some limitations that need to be addressed. One limitation of this study is the measurement of stress. The cognitive model of Harvey (2002) states that sleep is negatively influenced by distress and arousal. We measured stress by a questionnaire, however it did not show a significant relation with sleep impairment in our models, as previously discussed. Some people are not able to report stress very objectively or are used to a certain level of stress that they are unable to detect and report an increase of stress, also called reporting bias. A valid recommendation for future research is using an objective way to measure stress, perhaps a cortisol measurement (Bonnet, Léger, Baubet, Debilly & Cespuglio, 1997) or measuring the HPA axis by blood analysis (Kudielka, Buske-Kirschbaum, Hellhammer & Kirschbaum, 2004). Another recommendation is using a questionnaire that is more focused on distress or perceived stress, by using scales that measure perceived stress such as The Perceived Stress Scale (PSS) (Morin, Rodrigue & Ivers, 2003), distress such as the Psychiatric Epidemiology Research Interview (PERI) or daily stress such as The Daily Stress Inventory (DSI) (Morin, Rodrigue, & Ivers, 2003; Dohrenwend, Shrout, Egri & Mendelsohn, 1980).This argument also applies to the measurement of sleep. A polysomnography or EEG can be used to measure sleep objectively, to reduce reporting bias and recall bias.
We measured daytime and nighttime worrying with a questionnaire. But as discussed above, some people or not able to report worries very objectively or they are unaware of their worries, also called reporting bias or recall bias. The nighttime worrying questionnaire was automatically sent at 6.00 AM, nonetheless some people may be unable to report or reproduce their nighttime worries the next morning. Recommended for future research, is a more objective way of measuring worries, like an actigraphic (voice-activated) measurement as used in the study of Wiclow & Espie, 2000.
A third limitation is that we did not measure affect or arousal in this study. The cognitive model of Harvey (2002) states that sleep is negatively influenced by distress and arousal. Maybe this is not general stress, as measured in this study, but more focused on mood and bodily (hyper) arousal. According to one study, pre-sleep cognitive arousal directly led to sleep impairment: an average of eight minutes longer sleep onset (Wuyts et. al., 2012). Moreover, affect (or mood) is independently related to sleep impairment when measured separately next to negative toned cognitions (Kirkegaard et. al., 2003). This has not been researched for stress, therefore stress maybe not independently related to sleep impairment, but entwined with mood or affect. It could be a possibility that stress is a mediator of arousal or affect as one study suggests (Furunato & Harsh, 2006). A recommendation for further research is measuring arousal, when investigating the relation between stress, cognition and sleep. More important physiological arousal must be measured specific for example: by The Pre-Sleep Arousal Scale (PSAS) (Morin, Rodrigue, & Ivers, 2003). Affect can be measured by the Positive and Negative Affect Schedule (PANAS) (Watson, Clark & Tellegen, 1988).
Some technical limitations are also in place. The total explained variance of none of the models exceeded 19%. This leaves room for unexplained variance to remain large. Another limitation is that we did not have enough participants to achieve the desired power; this may have played a role in the lack of an effect of daytime worry and stress.
A final limitation is that we used a clinical sample with high scores on reported insomnia and depression. This sample does not represent the general population, but the general poor sleeper, which makes sense, because CBT-I is intended for helping poor sleepers. Hence selecting people who meet the DSM-V diagnosis of insomnia also seem to be more preoccupied with sleep compared to healthy sleepers. This was evident from our high ISI-score and may have enhanced the results of this study. This sample also tends to worry more and be more depressed then the general population. This was also evident from our higher CESD and worry scores. Furthermore, a final limitation is that a systematic bias is still a possibility because not everyone is able to connect to the internet on a daily basis or owns a computer.
In the analysis of this study only people who filled in their dairy in time were selected. Based on the results of this study there seems to be a difference when analyzing sleep diaries that were completed on time (Tables 5-10) and dairies that were not (supplemental Tables 1-5) . Maybe this has something to do with recall bias. Perhaps people forget the intensity of their worries overnight. Therefore, another recommendation is separating the individuals who did not complete their dairy on time, when analyzing the data. Another interesting issue that can be extracted from our data is
that there seems to be a significant difference in relationship between daytime worries, nighttime worries and sleep. Therefore, measuring these worrying at both daytime and nighttime is essential for further study.
Another interesting issue is that the largest part of the variance was on the day level (and not on the individual level, except for sleeping efficiency and total awake time). As consequence, nighttime worrying can be better explained by differences between days, then differences between individuals. Perhaps, coping skills between individuals or not as influential as different factors during the day. Therefore, people may experience more worries during the week, then during the weekend. Possible factors that can explain differences between days may be arousal or affect (mood) which were not measured in this study (as mention before).
Taken all these limitations into account it still seems evident that nighttime worry is most important when looking at the relation between worry, stress and sleep impairment. Interpreting the results; if these results are solid, which they seem to be, then research needs to focus more on nighttime worrying. Based on our results, perhaps nighttime factors are more important and have more influence on sleep, than daytime factors. Maybe people can forget their daytime worries, but lie awake at night because of their nighttime worries. When we link these findings to the cognitive model of Harvey (2002) it partially disconfirms the model. As discussed in the introduction, this model suggests that daytime and nighttime worrying are the main factors in causing and maintaining insomnia. These factors fuel arousal, distress, attention bias and perception bias to daytime functioning and sleep impairment. Notwithstanding, our results suggest that daytime worrying is not as important as nighttime worrying. This conclusion also applies for stress. The model of Harvey suggests that arousal and distress lead to an attention bias and perception bias. Nonetheless, the results of this study suggest that stress plays no role in aggravating and maintaining sleep impairment. Therefore, the results of this study partially disconfirm the model of Harvey (2002).
If only nighttime worrying is related with sleep impairment, then this would mean that CBT-I just needs to focus on nighttime worries and that daytime worries are not as relevant. Moreover, it is interesting that stress did not have a relation with sleep impairment in this study. Perhaps daytime stress can be coped with or people use sleep to recuperate from a stressful day as two studies suggest (Haynes, Adams & Franzen, 1981; Philbert et. al., 2011). This would mean that worrying at day and stress don’t have to be addressed in CBT-I. Perhaps the focus of CBT-I should be on nighttime worries, when people go to sleep. Moreover, more variance was explained on a day level and not on an individual level. Perhaps, the tendency to worry is not as influential as day-to-day factors, like the difference between a workday and weekend.
Concluding, this study consistently shows that with most of the sleeping measurements, only nighttime worry is associated with sleeping impairment. This is one of the first studies to investigate the difference between daytime and nighttime worry and their relation with sleep impairment. It is becoming evident that daytime worrying may not be directly related to sleeping impairment based on our results, the results of the thesis of van Baar (2013), the study of Wiclow & Epsie (2000) and the study of Weise, Ong, Telser, Kim and Roth (2013). Also the relation between stress and sleep still remains unclear. Measuring worries at day and at night clearly shows that only nighttime worrying is related with sleeping impairment. So it is important for future research and CBT-I to focus on nighttime worrying. These findings can help make CBT-I more effective when treating insomnia. Perhaps if CBT-I focusses more on nighttime worries, it may improve the effectiveness of CBT-I.
References
Åkerstedt, T., Kecklund, G., & Axelsson, J. (2007). Impaired sleep after bedtime stress and worries. Biological Psychology, 76(3), 170-173.
American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Arlington, VA: American Psychiatric Association.
Bastien, C. H., Vallières, A., & Morin, C. M. (2001). Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Medicine, 2, 297-307.
Beekman, A. T., van Limbeek, J., Deeg, D. J., Wouters, L., & van Tilburg, W. (1994). A screening tool for depression in the elderly in the general population: The usefulness of Center for Epidemiological Studies Depression Scale (CES-D). Tijdschrift voor Gerontologie en Geriatrie, 25, 95-103.
de Beurs, E., van Dyck, R., Marquenie, L. A., Lange, A. & Blonk, R. W. B. (2001). De DASS: een vragenlijst voor het meten van depressie, angst en stress. Gedragstherapie, 34, 35-53.
Bonnet, C., Léger, L., Baubet, V., Debilly, G. & Cespuglio, R. (1997). Influence of a 1 h immobilization stress on sleep states and corticotropin-like intermediate lobe peptide CLIP or ACTH, Ph-ACTH. Brain 18 – 39 18 – 39 contents in the rat. Brain Research, 751, 54-63. Borkovec, T. B. (1979). Pseudo (experiential)-insomnia and idiopathic (objective) insomnia:
Theoretical and therapeutic Issues. Advances in Behaviour Research and Therapy, 2, 27-55.
Borkovec, T. D., Ray, W. J., & Stöber, J. (1998). Worry: A cognitive phenomenon intimately linked to affective, physiological, and interpersonal behavioral processes. Cognitive Therapy and Research.
Brosschot, J. F., Gerin, W. & Thayer, J. F. (2005). The perseverative cognition hypothesis: A review of worry, prolonged stress-related physiological activation, and health. Journal of Psychosomatic Research, 60, 113-124.
Brosschot, J. F., Pieper, S. & Thayer, J. F. (2005). Expanding stress theory: Prolonged activation and perseverative cognition. Psychoneuroendocrinology, 30, 1043–1049.
Buysse, D. J., Ancoli-Israel, S., Edinger, J. D., Lichstein, K. L., & Morin, C. M. (2006). Recommendations for a standard research assessment of insomnia. Sleep, 9, 1155-1173.
Carney, C. E., Buysse, D. J., Ancoli-Israel, S., Edinger, J. D., Krystal, A. D., Lichstein, K. L., et al.
(2012). The consensus sleep diary: Standardizing prospective sleep self-monitoring. Sleep, 35, 287-302.
Carney, C. E., Harris, A. L., Moss, T. G., & Edinger, J. D. (2010). Distinguishing rumination from worry in clinical insomnia. Behaviour Research and Therapy, 48(6), 540-546.
Daley, M., Morin, C. M., LeBlanc, M., Grégoire, J. P., Savard, J., & Baillargeon, L. (2009). Insomnia and its relationship to health-care utilization, work absenteeism, productivity and accidents. Sleep Medicine, 10, 427–438.
Demirchyan, A., Petrosyan, V., & Thompson, M. E. (2011). Psychometric value of the Center for Epidemiologic Studies Depression (CES-D) Scale for screening of depressive symptoms in Armenian population. Journal of Affective Disorders, 133, 489-498.
Dohrenwend, B. P., Shrout, P. E., Egri, G., & Mendelsohn, F. S. (1980). Nonspecific psychological distress and other dimensions of psychopathology: Measures for use in the general population. Archives of General Psychiatry, 37(11), 1229-1236.
Fortunato, V. J., & Harsh, J. (2006). Stress and sleep quality: The moderating role of negative affectivity. Personality and Individual Differences, 41, 825-836.
Guestella, A. J., & Moulds, M. L. (2007). The impact of rumination on sleep quality following a stressful life event. Personality and Individual Differences, 42, 1151-1162.
Hall, M., Vasko, R., Buysse, D., Omboa, H., Chen, Q., Cashmere, J. D., et al. (2004). Acute stress affects heart rate variability during sleep. Psychosomatic Medicine, 66, 56–62.
Harvey, A. G. (2002). A cognitive model of insomnia. Behaviour Research and Therapy, 40, 869-893. Haynes, S. N., Adams, A. & Franzen, M. (1981). The effects of pre-sleep stress on sleep-onset
insomnia. Journal of Abnormal Psychology, 90, 601-606.
Hox, J. J. (2002).Multilevel Analysis: Techniques and Applications. Mahwah: Lawrence Erlbaum Associates.
Jacobs, G. D., Pace-Schott, E. F., Stickgold, R., & Otto, M. W. (2004). Cognitive behavior therapy and pharmacotherapy for insomnia: A randomized controlled trial and direct
comparison. Archives of Internal Medicine, 164, 1888-1896.
Janssen, R., de Vries, S. J., Verstappen, C. C. P., & de Leijer, C. (2011). De epidemiologie van slaapstoornissen. Bijblijven, 27, 7-16.
Jansson-Fröjmark, M., Harvey, G. A., Lundh, L., Norell-Clarke, A. & Linton, J. S. (2011). Psychometric Properties of an insomnia-specific measure of worry: The anxiety and preoccupation about sleep questionnaire. Cognitive Behaviour Therapy, 40, 65-76.
Kant, G. J., Pastel, R. H., Bauman, R. A., Meiniger, G. R., Maughan, K. R., Robinson, T. N., et al. (1995). Effects if chronic stress on sleep in rats. Physiology & Behavior, 57, 359-365.
Kirkegaard Thomsen, D., Mehlsen, M. Y., Christensen, S., & Zachariae, R. (2003). Rumination- relationship with negative mood and sleep quality. Personality and Individual Differences, 34, 1293-1301.
Kudielka, B. M., Buske-Kirschbaum, A., Hellhammer, D. H., & Kirschbaum, C. (2004). HPA axis responses to laboratory psychosocial stress in healthy elderly adults, younger adults, and children: impact of age and gender. Psychoneuroendocrinology,29(1), 83-98.
Lovibond, P. F., & Lovibond, S. H. (1995). The structure of negative emotional states: Comparison of the depression anxiety stress scales (DASS) with the Beck Depression and Anxiety Inventories. Behaviour Research and Therapy, 33, 335-343.
Milner, C. E., & Belicki, K. (2010). Assessment and treatment of insomnia in adults: A guide for clinicians. Journal of Counseling and Development, 88, 236-244.
Morin, C. M. (1993). Insomnia: Psychological assessment and management. New York: Guilford Press. Morin, C. M. & Benca, R. (2012). Chronic insomnia. Lancet, 379 (9821), 1129-1141.
Morin, C. M., Rodrigue, S. & Ivers, M. (2003). Role of stress, arousal, coping skills in primary insomnia. Psychosomatic Medicine, 65, 259–267.
Morin, C. M., LeBlanc, M., Daley, M., Gregoire, J. P., Mérette, C. (2006). Epidemiology of insomnia: Prevalence, self-help treatments, consultations, and determinants of help-seeking behaviors. Sleep Medicine, 7, 123-130.
Pace-Schott, E. F., Kaji, J., Stickgold, R., & Hobson, J. A. (1994). Nightcap measurement of sleep quality in self-described good and poor sleepers. Sleep, 17, 688–692.
Philbert, J., Pichat, P., Beeské, S., Decobert, M., Belzung, C. & Griebel, G. (2011). Acute inescapable stress exposure induces long-term sleep disturbances and avoidance behavior: A mouse model of post-traumatic stress disorder (PTSD). Behavioural Brain Research, 221, 149-154. Ohayon, M. M. (2002). Epidemiology of insomnia: What we know and what we still need to learn.
Sleep medicine Reviews, 6(2), 97-111.
Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurements, 1, 385-401.
Riemann, D., & Perlis, M. L. (2009). The treatments of chronic insomnia: a review of benzodiazepine receptor agonists and psychological and behavioral therapies. Sleep Medicine Reviews, 13(3), 205-214.
Spinhoven, Ph., Ormel, J., Sloekers, P. P. A., Kempen, G. I. J. M., Speckens, A. E. M., & van Hemert, A. M. (1997). A validation study of the Hospital Anxiety and Depression Scale (HADS) in different groups of Dutch subjects. Psychological Medicine, 27, 363-370.
Spoormaker, V. I., Verbeek, I., van den Bout, J., & Klip, E. C. (2005). Initial Validation of the SLEEP-50 Questionnaire. Behavioral Sleep Medicine, 3, 227-246.
Takano, K., Lijima, Y., & Tanno, Y. (2012). Repetitive thought and self-reported sleep disturbance. Behavior Therapy, 43, 779-789.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales.Journal of personality and social psychology,54(6), 1063.
Weise, S., Ong, J., Tesler, N. A., Kim, S., & Roth, W. T. (2013). Worried sleep: 24-h monitoring in high and low worriers. Biological Psychology, 94(1), 61-70.
Wicklow, A., & Espie, C. A. (2000). Intrusive thoughts and their relationship to actigraphic measurement of sleep: Towards a cognitive model of insomnia. Behaviour Research and Therapy, 38(7), 679-693.
Wuyts, J., De Valck, E., Vandekerckhove, M., Pattyn, N., Bulckaert, A., Berckmans, D., et al. (2012). The influence of pre-sleep cognitive arousal on sleep onset processes. International Journal of Psychophysiology, 83, 8-15.
Zigmond, A. S., & Snaith, R. P. (1983). The Hospital Anxiety and Depression Scale. Acta Psychiatrica Scandinavica, 67(6), 361-370.
Zoccola, P. M., Dickerson, S. S., & Lam, S. (2009). Rumination predicts longer sleep onset latency after an acute psychosocial stressor. Psychosomatic Medicine, 71(7), 771-775.
Supplemental Table 1
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of sleep efficiency
Sleep Efficiency β (SE) b (SE) Constant 95.783(10.300)* Dag .000(.022) -.002(.509) ISI -.049(.024)* -1.130(.549)* Anxiety .000(.029) -.009(.669) Depression .027(.014) .632(.334) Nighttime worrying -.045(.007)** -1.050(.158)** Daytime worrying .009(.008) .210(.189) Stress .006(.015) .145(.341) * = p<.05 (two-tailed), ** = p<.01 (two-tailed) Supplemental Table 2
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of total sleeping time
Total sleeping time
β (SE) b(SE) Constant 456.274(58.474)* Day .043(.022) 5.420(2.838) ISI -.028(.024) -3.603(3.121) Anxiety -.006(.030) -.794(3.811) Depression .022(.015) 2.812(1.896) Nighttime worrying -.047(.007)** -5.976(.859)** Daytime worrying .005(.008) .664(1.035) Stress .008(.015) 1.066(1.866) ** = p<.01 (two-tailed) 24
Supplemental Table 3
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of sleep onset latency
Sleep onset latency
β (SE) b(SE) Constant -.820(.512) Day -.011(.018) -.011(.018) ISI .015(.028) .015(.028) Anxiety .035(.034) .035(.034) Depression -.019(.017) -.019(.017) Nighttime worrying .037(.006)** .037(.006)** Daytime worrying -.005(.007) -.005(.007) Stress -.014(.013) -.014(.013) ** = p<.01 (two-tailed) Supplemental Table 4
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of wake after sleep onset
Wake after sleep onset
β (SE) b(SE) Constant -1.026(.402)* Day -.002(.023) -.002(.023) ISI .062(.021) .062(.021)** Anxiety .003(.026) .003(.026) Depression -.027(.013) -.027(.013)* Nighttime worrying .014(.007) .014(.007)* Daytime worrying -.002(.009) -.002(.009) Stress .004(.015) .004(.015) * = p<.05 (two-tailed), ** = p<.01 (two-tailed) 25
Supplemental Table 5
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of terminal wakefulness
Terminal Wakefulness β (SE) b(SE) Constant -.420(.409) Day .023(.024) .023(.024) ISI .022(.022) .022(.022) Anxiety .026(.026) -.004(.026) Depression .013(.013) -.002(.013) Nighttime worrying .007(.007) .003(.008) Daytime worrying .009(.009) .003(.009) Stress .016(.016) .016(.016) Supplemental Table 6
Day, depression, anxiety, nighttime worrying, daytime worrying, stress and insomnia severity index as predictors of total awake time
Total awake time
β (SE) b(SE) Constant -1.178(.433)* Day -.002(.024) .003(.022) ISI .049(.021)* .048(.023)* Anxiety .007(.026) .015(.028) Depression -.023(.013) -.033(.014)* Nighttime worrying .039(.008)** .040(.007)** Daytime worrying -.010(.009) -.004(.008) Stress -.020(.016) -.015(.015) * = p<.05 (two-tailed), ** = p<.01 (two-tailed) 26
