Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Feb 1.
Published in final edited form as: J Sleep Res. 2023 Aug 14;33(1):e13992. doi: 10.1111/jsr.13992

Effect of cognitive behavioural therapy and yoga for generalised anxiety disorder on sleep quality in a randomised controlled trial: The role of worry, mindfulness, and perceived stress as mediators

Ryan J Jacoby 1,*, Mackenzie L Brown 2, Sarah T Wieman 3, David Rosenfield 4, Susanne S Hoeppner 5, Eric Bui 6, Elizabeth A Hoge 7, Sat Bir S Khalsa 8,**, Stefan G Hofmann 9,**, Naomi M Simon 10,**
PMCID: PMC10840983  NIHMSID: NIHMS1915438  PMID: 37577773

Abstract

Sleep disturbances are present in approximately 65% of individuals with generalized anxiety disorder (GAD). Although both Kundalini Yoga (KY) and cognitive behavioral therapy (CBT) are effective treatment options for GAD, little is known about how these treatments compare in improving sleep for GAD and what drives these changes. Accordingly, we examined the effects of CBT, KY, and stress education (SE; an attention control condition) on subjective sleep quality (as measured by the Pittsburgh Sleep Quality Index [PSQI] and Insomnia Severity Index [ISI]) in a randomized controlled trial of 226 adults with GAD (M age = 33.37; 70% female; 79% White). We hypothesized that both CBT and KY would outperform SE in improving sleep disturbances. Three potential mediators of sleep improvement (worry, mindfulness, perceived stress) were also examined. In line with hypotheses, PSQI and ISI scores significantly improved from pre- to post-treatment for all three treatment groups (p’s < .001, d’s > .97). However, contrary to predictions, sleep changes were not significantly greater for CBT or KY compared to SE. In mediation analyses, within-person deviations in worry, mindfulness, and stress each significantly mediated the effect of time on sleep outcomes. Degree of change in sleep attributable to worry (CBT>KY>SE) and perceived stress (CBT, KY>SE) was moderated by treatment group. Personalized medicine as well as combined treatment approaches should be studied to help reduce sleep difficulties for patients with GAD who do not respond.

Keywords: Anxiety, Sleep, Mechanisms, Cognitive Behavioral Therapy, Insomnia, Stress, Yoga

Generalized Anxiety Disorder (GAD) is a common anxiety disorder characterized by excessive anxiety and worry about varied topics affecting 5.7% of Americans in their lifetime (Ruscio et al., 2017). GAD is also associated with physical symptoms and is the only anxiety disorder that includes sleep disturbance in its diagnostic criteria (e.g., difficulty falling or staying asleep; American Psychiatric Association, 2013). Fifty to seventy percent of patients with GAD endorse insomnia symptoms (i.e., persistent trouble falling asleep, staying asleep, and/or poor quality sleep associated with significant distress or impairment; American Psychiatric Association, 2013) that are as severe as their anxiety symptoms (e.g., Boland & Ross, 2015). Furthermore, sleep disturbances in GAD are associated with a wide range of impairments (Bélanger et al., 2004; Brenes et al., 2009; Glidewell et al., 2015), including impaired concentration and memory, decreased occupational functioning, increased absenteeism, and increased use of medical services (Benz et al., 2020; Harvey et al., 2009). Chronic insomnia may also be a risk factor for treatment relapse (Ohayon & Roth, 2003). Therefore, the additive clinical effects of anxiety and sleep problems warrant further exploration of treatment approaches that may address sleep in the context of GAD.

Cognitive-behavioral therapy (CBT) is the gold standard psychological treatment for anxiety (e.g., J. K. Carpenter et al., 2018). Clinical trials of CBT for GAD and elevated worry have shown positive effects on sleep (Belleville et al., 2010; Brenes et al., 2020; Bush et al., 2012; Ramsawh et al., 2016). However, many trials have found insomnia to be a residual symptom following anxiety disorder treatment (Glidewell et al., 2015), and the mechanisms by which CBT for GAD improves sleep are not well understood. There is evidence that the core symptom of GAD, pathological worry, may be a key mechanism in the development of sleep problems (Åkerstedt et al., 2007; Harvey et al., 2009). Elevated worry before bed is consistently associated with sleep interference (e.g., lower sleep efficiency) in healthy controls (e.g., Åkerstedt et al., 2007; Brosschot et al., 2007). Moreover, pathological worries about sleep (e.g., ruminative thoughts about the consequences of lost sleep, unrealistic expectations about sleep) may contribute to the maintenance of insomnia (Cousineau et al., 2016). Accordingly, the positive effects of CBT for GAD on sleep may be in part due to targeting pathological worries (Belleville et al., 2010; Sánchez-Ortuño & Edinger, 2012), although this has not yet been directly tested.

Complementary and integrative mind-body practices such as yoga are increasingly popular interventions for GAD and anxiety (Belleville et al., 2010; Jeter et al., 2015; Sánchez-Ortuño & Edinger, 2012). Yoga includes physical exercises, breath regulation, relaxation, and meditation, which are believed to facilitate reductions in psychophysiological arousal and increases in mindfulness (Hofmann et al., 2015; Hoge et al., 2013). There is also evidence for improvements in sleep/insomnia with yoga and mindfulness practices (S. B. S. Khalsa & Goldstein, 2021; Rusch et al., 2019; F. Wang et al., 2016; X. Wang et al., 2019). However, there is limited research investigating whether mind body interventions improve sleep among patients with GAD specifically (Hoge et al., 2013; M. K. Khalsa et al., 2015; Vøllestad et al., 2011). More research is also needed to clarify the mechanisms, such as mindfulness, by which these interventions have their effect on sleep in GAD. Research suggests that mindfulness likely improves insomnia symptoms due to reducing hyperarousal as well as improving self-regulation of thought processes (Ong & Kalmbach, 2023) suggesting that mindfulness is a promising mechanism for further study.

In summary, although evidence supports the efficacy of CBT and mind-body interventions for improving sleep among patients with GAD, there is a dearth of research directly comparing these treatments and testing potential mechanisms. Accordingly, in this secondary data analysis, we examined the efficacy of both group CBT and Kundalini Yoga (KY), compared with a stress education (SE) attention control treatment (which involves psychoeducation about the effects of stress on the body and addressing factors that can alleviate stress), on subjective sleep quality among patients with GAD enrolled in a randomized controlled trial (RCT; Hofmann et al., 2015; Simon et al., 2021). We hypothesized that scores on the Pittsburgh Sleep Quality Index (PSQI) and Insomnia Severity Index (ISI) would significantly decrease from baseline to post-treatment (regardless of intervention; Hypothesis 1A). We also examined whether these changes were moderated by treatment group (CBT, KY, SE) and hypothesized that CBT and KY would both outperform SE in improving sleep quality (CBT and KY > SE; Hypothesis 1B).

We additionally examined three potential mediators of sleep quality improvements (Aim 2) including worry and mindfulness, the previously mentioned targets of CBT and KY respectively. We also examined general perceived stress as a non-specific mediator because although the frequency of daily stressors did not differ between participants with insomnia and those who reported satisfaction with their sleep, individuals with insomnia perceive daily events to be more stressful (Morin et al., 2003), suggesting that sensitivity to stress may contribute to sleep difficulties. We hypothesized that all three mediators (worry, mindfulness, perceived stress) would be related to sleep quality outcomes (i.e., ISI and PSQI scores as a multivariate outcome) over and above the effects of the other mediators (Hypothesis 2). In moderated mediation analyses (Aim 3), we also examined whether mediated pathways were different across the three treatment groups, because some treatments might be more effective at changing certain mediators. Specifically, we predicted that CBT would change worry more than the other treatments (Hypothesis 3a) and that KY would change mindfulness more than the other treatments (Hypothesis 3b).

Method

Participants

Participants (N=226; M age = 33.37; 70% female; 79% White; 12.4% Hispanic) with a primary diagnosis of GAD as determined by the Structured Diagnostic Interview for DSM-IV (SCID-IV; First et al., 2002) or the Anxiety Disorders Interview Schedule-5 (ADIS-5; Brown & Barlow, 2014) were enrolled in a two-site RCT (Hofmann et al., 2015; Simon et al., 2021). Participants were randomly assigned to 12 weeks of treatment in one of three groups (2:2:1 ratio): CBT: n = 90, KY: n = 93; SE: n = 43. Demographics of the sample and primary study outcomes can be found in Simon et al. (2021). Institutional review board approvals and written informed consent were obtained.

Inclusion criteria included adults with a primary diagnosis of GAD and a Clinical Global Impression Scale-Severity (CGI-S; Guy, 1976) score of at least “Moderately ill” for GAD symptom severity who were free of or on stable psychotropic medication. Exclusion criteria included: (a) currently practicing mind-body techniques (e.g., yoga, meditation, Tai-Chi) or enrolled in psychotherapy for GAD, (b) more than 5 yoga classes or CBT sessions in the last 3 years, (c) lifetime history of a psychotic, bipolar, or developmental disorder, (d) diagnosis of posttraumatic stress disorder, substance use disorder, eating disorder, or organic mental disorder within the past 6 months, (e) significant suicidal ideation or suicidal behaviors within the past 6 months, (f) women who were pregnant, planning to become pregnant, or on an unreliable form of birth control, (g) cognitive impairment, or (h) serious medical illness or instability or history of head trauma. See Hofmann et al. (2015) for more details.

Measures

The PSQI (Buysse et al., 1989) is a 19-item self-report measure that evaluates past month sleep quality. The PSQI has seven sub-scores with items rated on a 4-point scale: (a) subjective sleep quality (i.e., “For the past month, how would you rate your sleep quality overall”), (b) sleep latency (i.e., how long it takes to fall asleep), (c) sleep duration (i.e., hours of actual sleep), (d) habitual sleep efficiency (i.e., ratio of number of hours slept divided by number of hours spent in bed), (e) sleep disturbance (i.e., reasons for sleep troubles; e.g., pain, trouble breathing, bad dreams), (f) use of sleep medications, and (g) daytime dysfunction (i.e., trouble staying awake and maintaining enthusiasm for daily activities). A global score ranging from 0 to 21 is calculated with higher scores indicating greater sleep disturbance and poorer sleep quality. A cutoff of > 8 was established in clinical populations for poor sleep quality (J. S. Carpenter & Andrykowski, 1998; Ibáñez del Prado & Cruzado, 2020). The PSQI has been shown to have good test-retest reliability and validity (Backhaus et al., 2002); α = .66 - .67 in the current study.

The ISI (Bastien et al., 2001) is a 7-item self-report measure of DSM-IV insomnia in the past 2 weeks, which measures severity (e.g., difficulty falling and staying asleep), distress (e.g., worry about sleep problems), and interference of sleep problems (e.g., self-perception of interference with daily functioning). The ISI uses a 5-point Likert scale, total scores range from 0–28, and higher scores represent more severe insomnia (cutoff ≥ 10 indicate cases of insomnia in community samples; Morin et al., 2011). The ISI has demonstrated good reliability and validity (Bastien et al., 2001); α =.86-.88 in the current study.

The Penn State Worry Questionnaire-Past Week (PSWQ-PW; Stöber & Bittencourt, 1998) is a self-report measure of pathological worry severity over the past week. The PSWQ-PW includes 15-items rated on a 7-point Likert scale from 0 (Never) to 6 (Almost always; e.g., “My worries overwhelmed me”). Items worded in the positive direction are reverse scored, with total scores from 0 to 90 (higher scores indicate greater worry severity). The PSWQ-PW has good test-retest reliability and high internal consistency (Stöber & Bittencourt, 1998); α = .89 - .96 in the current study.

The Five-Facet Mindfulness Questionnaire (FFMQ; Baer et al., 2006) is a 39-item self-report measure of five key components of mindfulness: (a) observing (e.g., “I pay attention to sensation, such as the wind in my hair or sun on my face”), (b) describing (e.g., “I’m good at finding words to describe my feelings”), (c) acting with awareness (e.g., “I find it difficult to stay focused on what’s happening in the present”), (d) non-judging of inner experience (e.g., “I make judgments about whether my thoughts are good or bad”), and (e) non-reactivity to inner experience (e.g., “When I have distressing thoughts or images I am able just to notice them without reacting”). Items are rated on a 5-point Likert scale from 1 (Never or rarely true) to 5 (Very often or always true); items worded in the negative direction are reverse scored. The overall score ranges from 39 to 195 (higher scores indicate more mindfulness). The five-factor structure has been validated in meditating and non-meditating samples (Baer et al., 2008); α’s = .80-.92 in the current study.

The Perceived Stress Scale (PSS; Cohen et al., 1983) is a 10-item self-report measure of the degree to which situations in one’s life are appraised as stressful over the last month (e.g., “In the last month, how often have you found that you could not cope with all the things that you had to do”). Items are rated on a 5-point Likert scale from 0 (Never) to 4 (Very often). Items worded in the positive direction are reverse scored; total scores range from 0–40, with higher scores indicating more perceived stress. The PSS has been shown to have adequate internal and test-retest reliability (Cohen et al., 1983); α = .80-.89 in the current study.

Procedure

Participants were randomized to group CBT, KY, or SE for 12 consecutive weekly 60-minute sessions and daily 20-minute homework practices between sessions. Each group type was led by two trained instructors with 3–6 patients per group. The assessments in this secondary analysis occurred at baseline and after week 12 of treatment.

Treatments.

Group CBT consisted of psychoeducation, cognitive restructuring, progressive muscle relaxation, worry exposures, and in vivo exposure exercises (mindfulness was not addressed). Group KY (as taught by Yogi Bhajan) included breathing techniques, meditation, yoga theory, philosophy and psychology, physical postures/exercises, and deep relaxation. Group SE received psychoeducation about the stress response, the negative stress cycle, stress and health/illness, and the contribution of lifestyle behaviors such as caffeine and alcohol intake. This included providing general psychoeducation about sleep physiology (e.g., the stages of sleep), misconceptions about sleep, and sleep hygiene, but this was neither prescriptive nor personalized. For additional details on the treatment manual, see Hofmann et al. (2015).

Statistical Analysis

Our primary analyses included the full sample of study participants (regardless of whether they scored above the clinical cutoff for insomnia). Sleep quality outcomes for Aim 1 (PSQI and ISI) were analyzed using multi-level modeling (MLM). The predictors were Time (weeks from baseline), Treatment Group (dummy coded to represent comparisons between CBT and SE and between KY and SE), and Time x Treatment Group. MLM includes all participants who complete at least one assessment (all 226 participants in the current study), regardless of other missing data. The MLMs had three levels: the two repeated measurements of sleep quality were nested within participants, who were nested within their treatment cohort. We followed with exploratory analyses to investigate whether specific aspects of sleep disturbance were impacted differently by the treatments by examining the seven PSQI subscales separately (correcting for multiple tests using the Benjamini-Hochberg correction).

For Aims 2–3, we conducted a longitudinal moderated mediation analysis using MLM. Since we had two separate, strongly correlated measures of sleep quality, we performed a single multivariate mediation analysis (i.e., PSQI and ISI comprised the multivariate outcome) using multivariate MLM (MMLM; Hox et al., 2018), rather than perform a separate mediation analysis for each outcome. Instead of calculating the regression coefficients to maximize the fit of the predictors to a single outcome (as in MLM), MMLM maximizes the fit to both outcomes simultaneously, calculating the regression coefficients to maximize the overall probability across both outcomes. So, the regression coefficients in MMLM will be similar to (but not exactly equal) the average of the regression coefficients if the outcomes had been analyzed separately.

For the mediation analysis, we examined whether PSWQ, FFMQ, and/or PSS mediated the effect of Time on the outcomes (See Figure 1). Measures were assessed at baseline and week 12. We examined the within-person, longitudinal relations between the variables in our model because between-person relations can be confounded by between-subjects 3rd variable confounds (L. Wang & Maxwell, 2015). The effect of Time on the mediators and on the outcome inherently calculate the within-subjects changes in those variables from baseline to post-treatment. Specifically, we followed L. Wang and Maxwell (2015) and disaggregated each mediator into its between-subjects component (the average level of the mediator over baseline and post), and its within-subjects component (the within-person deviation for each mediator at each time point from its own mean; L. Wang & Maxwell, 2015). Both the within-person deviations of each mediator at each timepoint, and the between-subjects mean on the mediator, were included in the model (see Figure 1). All 3 mediators were included in a single mediation model (see Supplement for more details).

Figure 1.

Figure 1.

Open in a new tab

PSWQ, FFMQ, and PSS as Mediators of the Change Over Time in the Multivariate Sleep Outcome.

Note. CBT = Cognitive Behavioral Therapy; KY = Kundalini Yoga; SE = Stress Education; FFMQ = Five-Facet Mindfulness Questionnaire; PSWQ-PW = Penn State Worry Questionnaire-Past Week; PSS = Perceived Stress Scale; PSQI = Pittsburgh Sleep Quality Index; ISI = Insomnia Severity Index. The mediator boxes in the model represent the two separate components of each mediator: the mean level of the mediator across assessments, and the deviations at each timepoint between the value of the mediator at that timepoint and the mean level of the mediator for that participant. The path coefficients reported in the figure are for the deviation component of that mediator, since the mean component of the mediator may simply represent third variable confounds between participants. Paths in bold are significant. The non-bolded path from Time to outcome was not significant. *** p < .001, ** p < .01, * p <.05, + p <.1. The subscript “t” indicates that the variable was measured repeatedly over time. Treatment Group was tested as a moderator of the change in PSWQ, FFMQ, and PSS over time. Path coefficients that appear above the path line represent the average change over time (per week) in the mediator across the three treatment groups (indicated by the “AVG” superscript). Treatment Group was a significant moderator of the effect of time on PSWQ and PSS, so the path coefficient for each treatment group is presented below the path line, with superscripts for each group. Treatment group was not a significant moderator of the change over time in FFMQ, so separate path coefficients for each treatment group are not presented. Paths from the mean of each mediator to the outcome are not shown because they cannot mediate change over time given that they don’t change over time.

The sleep quality outcomes were PSQI and ISI (as the multivariate outcome). The deviations in each mediator at each timepoint was a time-varying predictor of outcome at each timepoint and reflected the within-person relation between each mediator and the sleep outcomes over time. Since the change over time in the different mediators may vary between treatment groups (e.g., FFMQ may change more in KY than in SE), we allowed the effect of Time on the mediators to be moderated by Treatment Group (dummy coded as described in Aim 1). See Figure 1. The “a” paths in the mediation analysis were the effect of Time (i.e., the change over time) on each mediator using MLM. The “b” paths estimated the within-person relationships between the repeated measurements of the mediators (as time-varying predictors) with the repeated measurements of the outcomes (L. Wang & Maxwell, 2015) using MMLM (since we had a multivariate outcome). Mediated pathways were tested for significance using the program RMediation (Tofighi & MacKinnon, 2011). As an effect size measure of the mediated pathways, we report PM, which is the proportion of the effect of Time on sleep quality that was mediated by each particular mediator (PM=a*b/c). In additional exploratory analyses (See Supplement) we also examined two specific subscales of the FFMQ (acting with awareness and non-judging of inner experience) as mediators because a previous meta-analysis demonstrated that these two FFMQ subscales had the strongest relationships with affective symptoms (J. K. Carpenter et al., 2019). Finally, to determine if our findings applied to the subset of participants reporting substantial insomnia, we recalculated our analyses using the subgroup with an ISI score ≥ 10 at baseline (See Supplement). Additional details for the statistical analytic plan can be found in the Supplement.

Power Analyses.

Power analyses using the power analysis program RMASS2 indicted that, given the number of participants in each treatment at each time point, we have >.90 power to detect a medium effect size (Cohen’s d = .50) difference between treatments on sleep quality outcomes for this secondary analysis. For the mediation analysis, Fritz and Mackinnon (2007) showed that sample sizes > 75 are sufficient to have power > .80 to detect mediation when the mediated pathways were of “half-medium” effect size (β = .26).

Results

Preliminary analyses

Descriptive statistics for outcome and mediator measures by intervention group and time point can be found in Table 1. Posttreatment assessments were completed by 155 participants (68.6%). As reported in the primary outcomes paper (Simon et al., 2021), dropouts did not differ from completers on baseline characteristics.

Table 1.

Raw means and standard deviations for outcome and mediator measures by intervention group and time point

Kundalini Yoga CBT Stress Education
PRE (n = 93)a POST (n = 59)b PRE (n = 90)c POST (n = 67) PRE (n = 43)d POST (n = 29)e
PSWQ-PW 64.22 (11.73) 45.80 (19.00) 64.25 (12.70) 40.96 (17.13) 64.21 (10.38) 52.04 (17.52)
FFMQ Total 114.30 (15.75) 128.93 (18.64) 113.61 (18.94) 122.14 (20.49) 112.62 (15.42) 122.24 (17.45)
 Observe 24.91 (6.28) 26.34 (7.61) 23.53 (6.00) 23.81 (6.18) 24.86 (5.47) 25.07 (5.60)
 Describe 26.40 (6.43) 27.99 (6.53) 25.18 (7.38) 25.85 (6.95) 24.78 (6.35) 26.34 (6.44)
 Acting with Awareness 22.54 (6.22) 26.97 (5.98) 24.30 (5.86) 25.81 (6.31) 21.92 (5.68) 24.59 (5.09)
 Non-judgment 23.32 (7.44) 28.74 (6.96) 23.87 (7.89) 27.77 (7.73) 24.02 (7.24) 26.52 (8.67)
 Non-reactivity 17.13 (4.34) 18.90 (4.59) 16.73 (4.53) 18.90 (4.78) 17.04 (3.64) 19.72 (3.63)
PSS 23.67 (5.20) 18.17 (6.57) 23.84 (5.87) 17.86 (6.37) 23.83 (4.57) 20.45 (7.40)
PSQI 8.71 (3.19) 6.12 (3.41) 8.08 (3.93) 6.46 (3.25) 8.33 (3.98) 7.27 (3.40)
ISI 13.08 (6.65) 8.56 (5.67) 12.53 (6.71) 8.60 (5.33) 12.26 (6.15) 10.01 (5.84)
Open in a new tab

Note. CBT = Cognitive Behavioral Therapy; PSWQ-PW = Penn State Worry Questionnaire-Past Week; FFMQ = Five-Facet Mindfulness Questionnaire; PSS = Perceived Stress Scale; PSQI = Pittsburgh Sleep Quality Index; ISI = Insomnia Severity Index.

a

Kundalini Yoga Group - PRE: n = 92 for FFMQ, PSS, and ISI; n = 85 for PSQI

b

Kundalini Yoga Group - POST: n = 57 for PSQI

c

CBT Group - PRE: n = 85 for ISI

d

Stress Education Group - PRE: n = 42 for PSWQ-PW; n = 41 for FFMQ, PSS, and ISI; n = 40 for PSQI

e

Stress Education Group - POST: n = 26 for PSQI

Baseline Characteristics.

In the full sample, 47.6% (n = 100) were above the clinical cutoff of the PSQI at baseline (J. S. Carpenter & Andrykowski, 1998) and 63.7% (n = 142) were above the clinical cutoff on the ISI for insomnia (Morin et al., 2011). There were no site differences in baseline PSQI levels (t(212) = −0.97, p = .33, d = 0.13); one site had worse baseline insomnia as measured by the ISI (M = 13.60, SD = 7.08 vs. M = 11.84, SD = 5.93), t(208) = −2.00, p = .046, d = 0.28). Additional baseline characteristics and baseline correlations between mediators and outcomes can be found in the Supplement.

Aim 1: Impact of Treatment for GAD on Sleep Quality.

At post-treatment, 32.0% (n = 48) were above the clinical cutoff of the PSQI and 41.3% (n = 64) were above the cutoff on the ISI for insomnia. PSQI scores decreased (improved) across all three treatment groups from pre-treatment (estimated mean [SE]: 8.38 [0.12]) to post-treatment (6.66 [0.15]) with a large effect size (b=−.14; 95% CI=−.17 to −.11; t=−9.03, df=350, p<.001, d=0.97; Figure 2). The Time by Treatment Group interactions, testing change over time in CBT vs. SE (p=.73) and change over time in KY vs. SE (p = .16), were not significant, suggesting that the three treatment groups did not differentially improve overall sleep quality as measured by the PSQI.

Figure 2.

Figure 2.

Open in a new tab

Changes in PSQI Scores over Time by Treatment Group

Note. Figure depicts estimated marginal means (EMM) and standard errors (SEs). PSQI = Pittsburgh Sleep Quality Index; CBT = Cognitive Behavioral Therapy; KY = Kundalini Yoga; SE = Stress Education. We scaled the graph to clearly show the interaction (full PSQI range 0–21; cut-off score >8 indicates poor sleep quality).

MLMs examining changes in the seven PSQI subscales showed that 4 subscales improved significantly over time after correction for multiple tests using the Benjamini-Hochberg adjustment: (a) subjective sleep quality, t(184)=5.03, p<.001, d=0.74, (b) sleep latency, t(168)=5.05, p<.001, d=0.78, (c) sleep duration, t(167)=2.29, p=.02, d=0.35, and (d) daytime dysfunction, t(182)=4.84, p<.001, d=0.72. The other three subscales demonstrated false discovery rates greater than .05: habitual sleep efficiency (p=.04, d=0.31), sleep disturbances (p=.05, d=0.30), and use of sleep medications (p=.10, d=0.26). There were no significant Treatment Group x Time interactions (ps>0.05, ds<0.28).

ISI scores also decreased (i.e., improved) across the three treatment groups from pre (12.71 [0.19]) to post (9.26 [0.23]) with a large effect size (b=−.29; 95% CI=−.34 to −.24; t=−11.40, df=376, p<.001, d=1.18; Figure 3). The Time by Group interactions testing change in CBT vs. SE (p=.34) and change in KY vs. SE (p=.09) were not significant, suggesting that the three treatment groups did not differentially improve insomnia symptoms.

Figure 3.

Figure 3.

Open in a new tab

Changes in ISI Scores over Time by Treatment Group

Note. Figure depicts estimated marginal means (EMM) and standard errors (SEs). ISI = Insomnia Severity Index; CBT = Cognitive Behavioral Therapy; KY = Kundalini Yoga; SE = Stress Education. We scaled the graph to clearly show the interaction (full ISI range 0–28; cut-off score = 15 indicates moderate insomnia).

Aim 2: Mediation of Treatment Effects

MMLM analyses showed that, on average across treatment groups, the “a” paths from Time to deviations in PSWQ, FFMQ, and PSS were all significant, indicating that all three mediators improved significantly over time (Figure 1). The “b” paths from deviations in PSWQ and PSS to sleep quality outcomes were also significant, indicating that lower PSWQ and PSS deviations (lower within-person worry and perceived stress) were related to lower (better) scores on sleep quality outcomes within-subjects over time (Figure 1). Higher FFMQ deviations (greater mindfulness) were marginally related to lower (better) sleep quality (Figure 1). As a result, PSWQ, FFMQ, and PSS deviations were all significant mediators of the effect of Time on sleep quality outcomes (PSWQ: a*b=.012, 95% CI: .0002-.023, p<.05, percent mediated [PM]=32.4%; FFMQ: a*b=.007, 95% CI: .001-.013, p<.05, PM =18.9%; PSS: a*b = .010, 95% CI: .002-.019, p<.05, PM =27.0%). Thus, together the 3 mediators accounted for over 78% of the total effect of time on sleep quality.

Aim 3: Moderated Mediation.

The changes in PSWQ and PSS (but not FFMQ) were moderated by Treatment Group. PSWQ decreased significantly from pre- to post-treatment for all three treatments (Figure 1), but the decrease was: (a) greater in CBT than SE (difference=−.06; 95% CI=−.09 to −.03, t=−3.76, df=310, p<.001, d=0.43), (b) greater in KY than SE (difference=−.04; 95% CI=−.07 to −.003, t=−2.11, df=310, p=.035, d=0.24), and (c) greater in CBT than KY (difference=−.03; 95% CI=−.05 to −.0003, t=−1.99, df=310, p=.047, d=0.23). Hence, PSWQ was a significant mediator of change in sleep quality in all three treatment groups, but the change attributable to PSWQ was strongest in CBT (a*b=.015, 95% CI=−.029 to −.001, p<.05, PM =40.5%), followed by KY (a*b=.012, 95% CI=−.024 to −.01, p<.05, PM =32.3%), and then SE (a*b=.008, 95% CI=−.017 to −.0001, p<.05, PM =21.6%).

PSS also decreased significantly from pre to post for all three treatments (Figure 1), but the decrease was: (a) greater in CBT than SE (difference=−.06; 95% CI=−.10 to −.02, t=−3.00, df=308, p<.001, d=0.34), and (b) greater in KY than SE (difference=−.05; 95% CI=−.09 to −.01, t=−2.65, df=308, p=.009, d=0.30), but (c) not different in CBT vs. KY. Hence, PSS was a significant mediator of change in sleep quality in all three treatment groups, but the change attributable to PSS was strongest in CBT (a*b=.011, 95% CI=−.016 to −.007, p<.05, PM =29.7%) and KY (a*b=.011, 95% CI=−.015 to −.008, p<.05, PM =29.6%) than in SE (a*b=.006, 95% CI=−.013 to −.001, p<.05, PM = 16.2%). See the Supplement for exploratory mediation analyses of two FFMQ subscales (non-judging and the acting with awareness) as alternatives to the FFMQ total score as mediators of sleep quality outcomes.

Sensitivity Analyses

Sensitivity analyses using the subgroup of participants with an ISI score ≥ 10 at baseline (n = 142, 62.8% of the full sample; Morin et al., 2011) were very similar to the results for the full sample (See Supplement).

Discussion

The current secondary data analysis investigation directly compared the efficacy of two group interventions for GAD (CBT and KY) versus an attention control group (SE) in improving sleep quality. The sample reported high rates of insomnia (64%, as measured by the ISI), consistent with existing literature (Bélanger et al., 2004). In line with hypotheses, all three interventions, including the SE control, led to significant improvements in sleep quality with large effect sizes. These findings are consistent with existing literature demonstrating the efficacy of CBT (Belleville et al., 2010; Bush et al., 2012; Ramsawh et al., 2016), mind-body interventions (e.g., Hoge et al., 2013; Vøllestad et al., 2011), and SE (Hoge et al., 2013), for addressing sleep disturbance in GAD, and suggest that each effectively improves co-occurring sleep difficulties in GAD. Nonetheless, there remained a substantial percentage of patients (32–41%) who remained above the clinical cutoffs of our sleep quality measures, indicating that more work is needed.

The SE control condition may have showed efficacy in improving sleep quality because psychoeducation about sleep and sleep hygiene was taught in the curriculum. Sleep hygiene includes behavioral and environmental recommendations intended to promote sleep and has been associated with improvements in sleep quality (S. B. S. Khalsa & Goldstein, 2021; O’Donnell & Driller, 2017). This information was provided in an educational format and was not personalized to participants, nor were participants instructed to change their sleep habits. And yet, while CBT and KY outperformed SE in decreasing symptoms of GAD (Simon et al., 2021), the inclusion of sleep hygiene and other non-specific factors in the SE curriculum may have contributed to similar improvements in sleep quality across the three interventions.

When examining specific components of sleep quality (as measured by the PSQI), subjective sleep quality, sleep latency, sleep duration, and daytime dysfunction due to sleep disturbances were the components that specifically improved during treatment. Two past trials examining CBT for GAD also showed improvements in subjective sleep quality and sleep latency (Bush et al., 2012; Ramsawh et al., 2016), but these studies failed to show improvements in sleep duration or daytime dysfunction. Sleep efficiency, sleep disturbances (e.g., due to things like pain or trouble breathing), and frequency of use of sleep medications did not improve significantly in the current study. However, at baseline, use of sleep medications (24.8%) and certain sleep disturbances (e.g., being unable to breathe comfortably, 25.1%) were not frequently reported, perhaps limiting the ability of these aspects to improve.

In line with our mediation hypotheses, reductions in all three constructs, worry (PM=32.4%), mindfulness (PM=18.9%), and perceived stress (PM=27.0%), mediated the effects of change over time in sleep quality outcomes (Total PM=78.3%). In line with hypotheses, the amount of change in sleep quality outcomes attributable to worry was greater in CBT than KY, which was greater than SE. This is in line with findings suggesting that: (a) worry is a key mechanism in the formation of sleep problems and pathological worries about sleep maintain insomnia (Åkerstedt et al., 2007; Cousineau et al., 2016; Harvey et al., 2009), and (b) mindfulness meditation reduces emotional reactivity and ruminative thoughts which contribute to poor sleep quality (Rusch et al., 2019). Contrary to hypotheses, changes in FFMQ were not moderated by treatment group. We also found that the amount of change attributable to perceived stress was greater in CBT and KY than in SE. Of note, the mechanisms tested are all constructs related to negative affect, and the interventions have overlapping components (e.g., relaxation in CBT, breathing and meditation in yoga) which complicate the interpretation of our moderated mediation analyses.

For patients whose sleep quality did not improve, an important next step would be to continue to test integrative interventions such as Mindfulness Based Therapy for Insomnia (Perini et al., 2021) that includes psychoeducation about sleep physiology and sleep hygiene, targets worry beliefs about sleep via cognitive/mindfulness exercises, and addresses perceived stress through meditation and yoga-based practices. Similarly, research has also suggested that treatment of GAD with CBT enhanced with KY may improve sleep (M. K. Khalsa et al., 2015). Thus, combined approaches are promising future directions for addressing GAD and sleep comorbidities. Alternatively, meta-analytic findings show that CBT for insomnia demonstrates a moderate effect on comorbid anxiety symptoms (ES=0.41; Belleville et al., 2011); however, none of the studies included enrolled individuals with GAD specifically which is a direction for future research. Finally, personalized medicine approaches could determine which treatment approaches (e.g., CBT vs. KY) are best for certain patients.

Although a strength of the current study is the longitudinal design, our mediational analyses are correlational (e.g., we cannot infer that reductions in worry cause improvements in sleep quality because we were not able to perform a meaningful time lagged analysis with measures that were only assessed at baseline and post-treatment). Moreover, while we were powered to detect medium effects in this secondary data analysis, smaller effect sizes may have been missed. Another limitation was that sleep quality measures were not administered mid-treatment, which limited our ability to look at more complex mediation models. The relationship between anxiety and sleep disturbance is likely bidirectional (Glidewell et al., 2015; Harvey et al., 2009), such that anxiety/worry may interfere with sleep but also lack of sleep may lead to impaired mood regulation (Babson et al., 2009; Jansson-Fröjmark & Lindblom, 2008). Thus, future studies should include multiple mid-treatment measures of sleep quality in addition to the mediators in order to explore these relationships.

Additionally, the primary outcome measures used (ISI and PSQI) are both retrospective self-report measures. Although these measures are common, daily diary or sleep logs are the gold-standard self-report method for sleep quality outcome data (e.g., Halpern et al., 2014), and a combination of these methods provides an excellent assessment of one’s sleep quality (Harvey et al., 2009). Thus, future studies might utilize a daily diary method to obtain information on sleep onset, number of awakenings, total time spent asleep, and subjective sleep quality as it is likely that results from the current trial would have differed across measurement types. Moreover, ecological momentary assessment (EMA) could assess how the relationships between anxiety and sleep quality change pre/post treatment. Furthermore, future studies should include objective measures of sleep disturbance such as actigraphy or polysomnography to objectively capture sleep disturbance improvements via behavioral interventions for GAD.

In conclusion, the present study sought to examine the differential efficacy of group CBT and KY compared with a SE control for improving sleep quality in GAD and whether worry, mindfulness, and/or perceived stress mediated the effects of treatment on sleep difficulties. Findings suggest that each of these interventions are efficacious for improving sleep quality in individuals with GAD, and that worry, mindfulness, and stress each significantly mediate these effects. The amount of change in sleep outcomes attributable to worry (CBT>KY>SE) and perceived stress (CBT, KY>SE) was moderated by treatment group. Future treatment studies should include intensive time series methods (e.g., daily sleep diaries, EMA) and objective measures (e.g., actigraphy, polysomnography).

Supplementary Material

Supinfo
Fig S1

Disclosure Statement:

Research reported in this publication was supported by the National Center for Complementary and Integrative Health of the National Institutes of Health awarded to principal investigators Drs. Hofmann and Simon under award numbers R01AT007257 & R01AT007258. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Dr. Jacoby has received salary and research support from the National Institute of Mental Health (NMH; K23MH120351). She is also paid as a faculty member of the Massachusetts General Hospital Psychiatry Academy. She receives book royalties from Hogrefe Publishing.

Dr. Rosenfield received grants and/or consulting fees from the NCCIH, NIH, the Department of Defense, and Rosenfield Analytics during the conduct of the study.

Dr. Bui received grants from the NIH during the conduct of the study and received royalties from Springer and from Wolters Kluwer, personal fees from Cerevel Therapeutics LLC, and grants from the US Department of Defense, Osher Center for Integrative Medicine, Patient-Centered Outcomes Research Institute, and Elizabeth Dole Foundation outside the submitted work.

Dr. Khalsa received grants from the NCCIH and NIH during the conduct of the study; received grants and personal fees from Kundalini Research Institute and grants from Kripalu Center for Yoga & Health outside the submitted work; and is a practitioner and certified instructor in Kundalini yoga as taught by Yogi Bhajan. He receives book royalties from Handspring Publishing.

Dr. Hofmann receives financial support by the Alexander von Humboldt Foundation (as part of the Alexander von Humboldt Professur), the Hessische Ministerium für Wissenschaft und Kunst (as part of the LOEWE Spitzenprofessur), NIH/NIMH R01MH128377, NIH/NIMHU01MH108168, Broderick Foundation/MIT, and the James S. McDonnell Foundation 21st Century Science Initiative in Understanding Human Cognition – Special Initiative. He receives compensation for his work as editor from SpringerNature. He also receives royalties and payments for his work from various publishers.

Dr. Simon in the past 2 years received grants from the National Institutes of Health (NIH) during the conduct of the study; received grant funding from the American Foundation for Suicide Prevention, Patient-Centered Outcomes Research Institute, Ananda Scientific, and support from the Cohen Veteran’s Network; received personal fees from Cerevel, Vanda Pharmaceuticals, Engrail, Praxis Therapeutics, Bionomics Limited, Genomind, and Wiley; performed grant reviews for American Foundation for Suicide Prevention; receiving royalties for UpToDate contributions from Wolters Kluwer and APA Publishing (Textbook of Anxiety, Trauma and OCD Related Disorders 2020); and has spousal stock from G1 Therapeutics and Zentalis outside the submitted work.

Footnotes

All other authors report no conflicts of interest.

Clinical Trials Registration: GATE: Generalized Anxiety – A Treatment Evaluation, ClinicalTrials.gov Identifier: NCT01912287 (https://clinicaltrials.gov/ct2/show/NCT01912287)

Contributor Information

Ryan J. Jacoby, Massachusetts General Hospital / Harvard Medical School, Department of Psychiatry, 185 Cambridge St, Suite 2000, Boston, MA 02114, USA.

Mackenzie L. Brown, University of Louisville, Department of Psychology, 2301 S 3rd St, Louisville, KY 40292, USA.

Sarah T. Wieman, Suffolk University, Department of Psychology, 73 Tremont St., Boston, MA 02108, USA.

David Rosenfield, Southern Methodist University, Department of Psychology, P.O. Box 750442, Dallas, TX 75275, USA.

Susanne S. Hoeppner, Massachusetts General Hospital / Harvard Medical School, Department of Psychiatry, 185 Cambridge St, Suite 2000, Boston, MA 02114, USA.

Eric Bui, Normandie Univ, UNICAEN, INSERM, U1237, PhIND “Physiopathology and Imaging of Neurological Disorders”, NEUROPRESAGE Team, (Institut Blood and Brain @ Caen-Normandie), GIP Cyceron, Caen, France & Centre Hospitalier Universitaire Caen Normandie, Caen, France.

Elizabeth A. Hoge, Georgetown University Medical Center, Department of Psychiatry, 2115 Wisconsin Ave, NW, Suite 200, Washington, DC, 20007, USA.

Sat Bir S. Khalsa, Brigham and Women’s Hospital / Harvard Medical School, Division of Sleep and Circadian Disorders, Department of Medicine, 221 Longwood Avenue, Boston, MA, 02115, USA.

Stefan G. Hofmann, Philipps-University Marburg, Department of Clinical Psychology, Schulstrasse 12, 35037 Marburg/Lahn, Germany.

Naomi M. Simon, New York University Grossman School of Medicine, Department of Psychiatry, One Park Avenue, 8th Floor, New York NY 10016, USA.

Data Availability Statement:

The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. At the time of data collection, the IRB consent form did not allow for release of individual patient data.

References

  1. Åkerstedt T, Kecklund G, & Axelsson J (2007). Impaired sleep after bedtime stress and worries. Biological Psychology, 76(3), 170–173. 10.1016/j.biopsycho.2007.07.010 [DOI] [PubMed] [Google Scholar]
  2. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Author. [Google Scholar]
  3. Babson KA, Feldner MT, Trainor CD, & Smith RC (2009). An experimental investigation of the effects of acute sleep deprivation on panic-relevant biological challenge responding. Behavior Therapy, 40(3), 239–250. 10.1016/j.beth.2008.06.001 [DOI] [PubMed] [Google Scholar]
  4. Backhaus J, Junghanns K, Broocks A, Riemann D, & Hohagen F (2002). Test–retest reliability and validity of the Pittsburgh Sleep Quality Index in primary insomnia. Journal of Psychosomatic Research, 4. [DOI] [PubMed] [Google Scholar]
  5. Baer RA, Smith GT, Hopkins J, Krietemeyer J, & Toney L (2006). Using self-report assessment methods to explore facets of mindfulness. Assessment, 13(1), 27–45. 10.1177/1073191105283504 [DOI] [PubMed] [Google Scholar]
  6. Baer RA, Smith GT, Lykins E, Button D, Krietemeyer J, Sauer S, Walsh E, Duggan D, & Williams JMG (2008). Construct validity of the Five Facet Mindfulness Questionnaire in meditating and nonmeditating samples. Assessment, 15(3), 329–342. 10.1177/1073191107313003 [DOI] [PubMed] [Google Scholar]
  7. Bastien CH, Vallières A, & Morin CM (2001). Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Medicine, 2(4), 297–307. 10.1016/s1389-9457(00)00065-4 [DOI] [PubMed] [Google Scholar]
  8. Bélanger L, Morin CM, Langlois F, & Ladouceur R (2004). Insomnia and generalized anxiety disorder: Effects of cognitive behavior therapy for GAD on insomnia symptoms. Journal of Anxiety Disorders, 18(4), 561–571. 10.1016/S0887-6185(03)00031-8 [DOI] [PubMed] [Google Scholar]
  9. Belleville G, Cousineau H, Levrier K, & St-Pierre-Delorme M-È (2011). Meta-analytic review of the impact of cognitive-behavior therapy for insomnia on concomitant anxiety. Clinical Psychology Review, 31(4), 638–652. 10.1016/j.cpr.2011.02.004 [DOI] [PubMed] [Google Scholar]
  10. Belleville G, Cousineau H, Levrier K, St-Pierre-Delorme M-E, & Marchand A (2010). The impact of cognitive-behavior therapy for anxiety disorders on concomitant sleep disturbances: A meta-analysis. Journal of Anxiety Disorders, 24(4), 379–386. 10.1016/j.janxdis.2010.02.010 [DOI] [PubMed] [Google Scholar]
  11. Benz F, Knoop T, Ballesio A, Bacaro V, Johann AF, Rücker G, Feige B, Riemann D, & Baglioni C (2020). The efficacy of cognitive and behavior therapies for insomnia on daytime symptoms: A systematic review and network meta-analysis. Clinical Psychology Review, 80, 101873. 10.1016/j.cpr.2020.101873 [DOI] [PubMed] [Google Scholar]
  12. Boland EM, & Ross RJ (2015). Recent advances in the study of sleep in the anxiety disorders, obsessive-compulsive sisorder, and posttraumatic stress disorder. The Psychiatric Clinics of North America, 38(4), 761–776. 10.1016/j.psc.2015.07.005 [DOI] [PubMed] [Google Scholar]
  13. Brenes GA, Divers J, Miller ME, Anderson A, Hargis G, & Danhauer SC (2020). Comparison of cognitive-behavioral therapy and yoga for the treatment of late-life worry: A randomized preference trial. Depression and Anxiety, 37(12), 1194–1207. 10.1002/da.23107 [DOI] [PubMed] [Google Scholar]
  14. Brenes GA, Miller ME, Stanley MA, Williamson JD, Knudson M, & McCall WV (2009). Insomnia in older adults with generalized anxiety disorder. The American Journal of Geriatric Psychiatry, 17(6), 465–472. 10.1097/JGP.0b013e3181987747 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Brosschot JF, Van Dijk E, & Thayer JF (2007). Daily worry is related to low heart rate variability during waking and the subsequent nocturnal sleep period. International Journal of Psychophysiology, 63(1), 39–47. APA PsycInfo. 10.1016/j.ijpsycho.2006.07.016 [DOI] [PubMed] [Google Scholar]
  16. Brown TA, & Barlow DH (2014). Anxiety and Related Disorders Interview Schedule for DSM-5 (ADIS-5) - Lifetime Version: Client Interview Schedule 5-Copy Set. OUP USA Treatments That Work. [Google Scholar]
  17. Bush AL, Armento MEA, Weiss BJ, Rhoades HM, Novy DM, Wilson NL, Kunik ME, & Stanley MA (2012). The Pittsburgh Sleep Quality Index in older primary care patients with generalized anxiety disorder: Psychometrics and outcomes following cognitive behavioral therapy. Psychiatry Research, 199(1), 24–30. 10.1016/j.psychres.2012.03.045 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Buysse DJ, Reynolds CF III., Monk TH., Berman SR., & Kupfer DJ. (1989). The Pittsburgh Sleep Quality Index: A new instrument for psychiatric practice and research. Psychiatry Research, 28(2), 193–213. [DOI] [PubMed] [Google Scholar]
  19. Carpenter JK, Andrews LA, Witcraft SM, Powers MB, Smits JAJ, & Hofmann SG (2018). Cognitive behavioral therapy for anxiety and related disorders: A meta-analysis of randomized placebo-controlled trials. Depression and Anxiety, 35(6), 502–514. 10.1002/da.22728 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Carpenter JK, Conroy K, Gomez AF, Curren LC, & Hofmann SG (2019). The relationship between trait mindfulness and affective symptoms: A metaanalysis of the Five Facet Mindfulness Questionnaire (FFMQ). Clinical Psychology Review, 74. 10.1016/j.cpr.2019.101785 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Carpenter JS, & Andrykowski MA (1998). Psychometric evaluation of the Pittsburgh Sleep Quality Index. Journal of Psychosomatic Research, 45(1), 5–13. 10.1016/S0022-3999(97)00298-5 [DOI] [PubMed] [Google Scholar]
  22. Cohen S, Kamarck T, & Mermelstein R (1983). A global measure of perceived stress. Journal of Health and Social Behavior, 24(4), 385–396. 10.2307/2136404 [DOI] [PubMed] [Google Scholar]
  23. Cousineau H, Marchand A, Bouchard S, Bélanger C, Gosselin P, Langlois F, Labrecque J, Dugas MJ, & Belleville G (2016). Insomnia symptoms following treatment for comorbid panic disorder with agoraphobia and generalized anxiety disorder. Journal of Nervous and Mental Disease, 204(4), 267–273. APA PsycInfo. 10.1097/NMD.0000000000000466 [DOI] [PubMed] [Google Scholar]
  24. First M, Spitzer R, Gibbon M, & Williams J (2002). Structured Clinical Interview for DSM-IV-TR Axis I Disorders, Research Version, Patient Edition (SCID-I/P). Biometrics Research, New York State Psychiatric Institute. [Google Scholar]
  25. Fritz MS, & MacKinnon DP (2007). Required sample size to detect the mediated effect. Psychological Science, 18(3), 233–239. 10.1111/j.1467-9280.2007.01882.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Glidewell RN, McPherson Botts E, & Orr WC (2015). Insomnia and anxiety: Diagnostic and management implications of complex interactions. Sleep Medicine Clinics, 10(1), 93–99. 10.1016/j.jsmc.2014.11.008 [DOI] [PubMed] [Google Scholar]
  27. Guy W (1976). Clinical Global Impressions. In ECDEU Assessment Manual for Psychopharmacology—Revised (pp. 218–222). National Institute of Mental Health. DHEW Pub. No. ADM 76–338. [Google Scholar]
  28. Halpern J, Cohen M, Kennedy G, Reece J, Cahan C, & Baharav A (2014). Yoga for improving sleep quality and quality of life for older adults. Alternative Therapies in Health and Medicine, 20(3), 37–46. [PubMed] [Google Scholar]
  29. Harvey AG, Hairston IS, Gruber J, & Gershon A (2009). Anxiety and sleep. In Antony MM& Stein MB (Eds.), Oxford handbook of anxiety and related disorders. (2008–13193-047; pp. 611–621). Oxford University Press; APA PsycInfo. [Google Scholar]
  30. Hofmann SG, Curtiss J, Khalsa SBS, Hoge E, Rosenfield D, Bui E, Keshaviah A, & Simon N (2015). Yoga for generalized anxiety disorder: Design of a randomized controlled clinical trial. Contemporary Clinical Trials, 44, 70–76. 10.1016/j.cct.2015.08.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hoge EA, Bui E, Marques L, Metcalf CA, Morris LK, Robinaugh DJ, Worthington JJ, Pollack MH, & Simon NM (2013). Randomized controlled trial of mindfulness meditation for generalized anxiety disorder: Effects on anxiety and stress reactivity. The Journal of Clinical Psychiatry, 74(8), 786–792. 10.4088/JCP.12m08083 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Hox JJ, Moerbeek M, & Schoot R van de. . (2018). Multilevel Analysis: Techniques and Applications, Third Edition (3rd edition). Routledge. [Google Scholar]
  33. Ibáñez del Prado C, & Cruzado JA (2020). A screening method for sleep disturbances at the end-of-life. Palliative and Supportive Care, 18(4), 468–472. 10.1017/S1478951520000024 [DOI] [PubMed] [Google Scholar]
  34. Jansson-Fröjmark M, & Lindblom K (2008). A bidirectional relationship between anxiety and depression, and insomnia? A prospective study in the general population. Journal of Psychosomatic Research, 64(4), 443–449. 10.1016/j.jpsychores.2007.10.016 [DOI] [PubMed] [Google Scholar]
  35. Jeter PE, Slutsky J, Singh N, & Khalsa SBS (2015). Yoga as a therapeutic intervention: A bibliometric analysis of published research studies from 1967 to 2013. The Journal of Alternative and Complementary Medicine, 21(10), 586–592. APA PsycInfo. 10.1089/acm.2015.0057 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Khalsa MK, Greiner-Ferris JM, Hofmann SG, & Khalsa SBS (2015). Yoga-enhanced cognitive behavioural therapy (Y-CBT) for anxiety management: A pilot study. Clinical Psychology & Psychotherapy, 22(4), 364–371. 10.1002/cpp.1902 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Khalsa SBS, & Goldstein MR (2021). Treatment of chronic primary sleep onset insomnia with Kundalini Yoga: A randomized controlled trial with active sleep hygiene comparison. Journal of Clinical Sleep Medicine, 17(9), 1841–1852. 10.5664/jcsm.9320 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Morin CM, Belleville G, Bélanger L, & Ivers H (2011). The Insomnia Severity Index: Psychometric indicators to detect insomnia cases and evaluate treatment response. Sleep, 34(5), 601–608. 10.1093/sleep/34.5.601 [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Morin CM, Rodrigue S, & Ivers H (2003). Role of stress, arousal, and coping skills in primary insomnia. Psychosomatic Medicine, 65(2), 259–267. [DOI] [PubMed] [Google Scholar]
  40. O’Donnell S, & Driller MW (2017). Sleep-hygiene education improves sleep indices in elite female athletes. International Journal of Exercise Science, 10(4), 522–530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Ohayon MM, & Roth T (2003). Place of chronic insomnia in the course of depressive and anxiety disorders. Journal of Psychiatric Research, 37(1), 9–15. 10.1016/S0022-3956(02)00052-3 [DOI] [PubMed] [Google Scholar]
  42. Ong J, & Kalmbach D (2023). Mindfulness as an adjunct or alternative to CBT-I. Sleep Medicine Clinics, 18(1). 10.1016/j.jsmc.2022.09.002 [DOI] [PubMed] [Google Scholar]
  43. Perini F, Wong KF, Lin J, Hassirim Z, Ong JL, Lo J, Ong JC, Doshi K, & Lim J (2021). Mindfulness-based therapy for insomnia for older adults with sleep difficulties: A randomized clinical trial. Psychological Medicine, 1–11. 10.1017/S0033291721002476 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Ramsawh HJ, Bomyea J, Stein MB, Cissell SH, & Lang AJ (2016). Sleep quality improvement during cognitive behavioral therapy for anxiety disorders. Behavioral Sleep Medicine, 14(3), 267–278. 10.1080/15402002.2014.981819 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rusch HL, Rosario M, Levison LM, Olivera A, Livingston WS, Wu T, & Gill JM (2019). The effect of mindfulness meditation on sleep quality: A systematic review and meta-analysis of randomized controlled trials. Annals of the New York Academy of Sciences, 1445(1), 5–16. 10.1111/nyas.13996 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Ruscio AM, Hallion LS, Lim CCW, Aguilar-Gaxiola S, Al-Hamzawi A, Alonso J, Andrade LH, Borges G, Bromet EJ, Bunting B, Caldas de Almeida JM, Demyttenaere K, Florescu S, de Girolamo G, Gureje O, Haro JM, He Y, Hinkov H, Hu C, … Scott KM (2017). Cross-sectional comparison of the epidemiology of DSM-5 generalized anxiety disorder across the globe. JAMA Psychiatry, 74(5), 465–475. 10.1001/jamapsychiatry.2017.0056 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Sánchez-Ortuño MM, & Edinger JD (2012). Internight sleep variability: Its clinical significance and responsiveness to treatment in primary and comorbid insomnia. Journal of Sleep Research, 21(5), 527–534. APA PsycInfo. 10.1111/j.1365-2869.2012.01010.x [DOI] [PubMed] [Google Scholar]
  48. Simon NM, Hofmann SG, Rosenfield D, Hoeppner SS, Hoge EA, Bui E, & Khalsa SBS (2021). Efficacy of yoga vs cognitive behavioral therapy vs stress education for the treatment of generalized anxiety disorder: A randomized clinical trial. JAMA Psychiatry, 78(1), 13. 10.1001/jamapsychiatry.2020.2496 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Stöber J, & Bittencourt J (1998). Weekly assessment of worry: An adaptation of the Penn State Worry Questionnaire for monitoring changes during treatment. Behaviour Research and Therapy, 36(6), 645–656. 10.1016/s0005-7967(98)00031-x [DOI] [PubMed] [Google Scholar]
  50. Tofighi D, & MacKinnon DP (2011). RMediation: An R package for mediation analysis confidence intervals. Behavior Research Methods, 43(3), 692–700. 10.3758/s13428-011-0076-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Vøllestad J, Sivertsen B, & Nielsen GH (2011). Mindfulness-based stress reduction for patients with anxiety disorders: Evaluation in a randomized controlled trial. Behaviour Research and Therapy, 49(4), 281–288. 10.1016/j.brat.2011.01.007 [DOI] [PubMed] [Google Scholar]
  52. Wang F, Eun-Kyoung Lee O, Feng F, Vitiello MV, Wang W, Benson H, Fricchione GL, & Denninger JW (2016). The effect of meditative movement on sleep quality: A systematic review. Sleep Medicine Reviews, 30, 43–52. 10.1016/j.smrv.2015.12.001 [DOI] [PubMed] [Google Scholar]
  53. Wang L, & Maxwell SE (2015). On disaggregating between-person and within-person effects with longitudinal data using multilevel models. Psychological Methods, 20(1), 63–83. 10.1037/met0000030 [DOI] [PubMed] [Google Scholar]
  54. Wang X, Li P, Pan C, Dai L, Wu Y, & Deng Y (2019). The Effect of Mind-Body Therapies on Insomnia: A Systematic Review and Meta-Analysis. Evidence-Based Complementary and Alternative Medicine: ECAM, 2019, 9359807. 10.1155/2019/9359807 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supinfo
NIHMS1915438-supplement-Supinfo.docx (28.9KB, docx)
Fig S1
NIHMS1915438-supplement-Fig_S1.pdf (62KB, pdf)

Data Availability Statement

The data underlying this article cannot be shared publicly due to the privacy of individuals that participated in the study. At the time of data collection, the IRB consent form did not allow for release of individual patient data.

RESOURCES