• DOI:10.7326/M14-2841
• Corpus ID: 21617330

Cognitive Behavioral Therapy for Chronic Insomnia

@article{Trauer2015CognitiveBT,
  title={Cognitive Behavioral Therapy for Chronic Insomnia},
  author={James M. Trauer and Mary Y Y Qian and Joseph S. Doyle and Shantha M. W. Rajaratnam and David Cunnington},
  journal={Annals of Internal Medicine},
  year={2015},
  volume={163},
  pages={191-204}
}

• J. Trauer, M. Qian, +2 authors D. Cunnington
• Published 4 August 2015
• Psychology, Medicine
• Annals of Internal Medicine

Insomnia is a prevalent condition, with 5% to 15% of adults meeting formal diagnostic criteria for chronic insomnia (15) (now termed insomnia disorder [6]) and one third reporting dissatisfaction with sleep. Insomnia is associated with both medical and psychiatric comorbidity, being linked to anxiety; depression (7); chronic health problems, such as hypertension (8, 9) and type 2 diabetes (10); health care use; nonmotor vehicle accidents; pain (11); and use of medication and alcohol (1215). Symptoms of insomnia have functional consequences even in the absence of a formal diagnosis (16), with the high economic burden of the condition largely mediated through the productivity cost of work absenteeism (17). Hypnotics, such as benzodiazepines and related drugs, are the most commonly used treatment for insomnia, with around 6% to 10% of U.S. adults using hypnotics in 2010 (18) and 27 daily doses of such drugs being taken per 1000 U.S. persons (19). In Australia, around 90% of primary care encounters for insomnia result in hypnotic prescription (20). Furthermore, despite a lack of evidence, use of second-generation antipsychotics (such as quetiapine) is also increasing, possibly due to patient and physician dissatisfaction with available treatments and a perceived lack of alternatives (21, 22). In this context, considering nonpharmacologic treatment options for insomnia disorder is important. Cognitive behavioral therapy for insomnia (CBT-i) is an effective nonpharmacologic treatment that improves sleep outcomes with minimal adverse effects (23) and is preferred by patients to drug therapy (24). The approach to CBT-i has been refined in recent years, and it is now most commonly studied as a combined cognitive and behavioral treatment incorporating some or all of 5 components. The components are described in Table 1, and although the precise efficacy of each has not been determined, the package of care is more effective than separate delivery of the cognitive or behavioral components (25). Although previous meta-analyses have been performed (2629), no recent meta-analysis has assessed the efficacy of this now-established package of care. We present a meta-analysis of the efficacy of CBT-i on sleep diary outcomes, compared with control, for the treatment of adults with chronic insomnia. Table 1. Components of CBT-i Methods We performed a systematic review and meta-analysis in accordance with the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines, using methods from the Cochrane Handbook for Systematic Reviews of Interventions. The predetermined methods were registered online with PROSPERO (CRD42012002863) (30), with full methods presented in section 1 of the Supplement. Supplement. Supplementary Material Data Sources and Searches We searched MEDLINE, EMBASE, PsycINFO, CINAHL, the Cochrane Library, and PubMed Clinical Queries from inception to 31 March 2015 with the terms sleeplessness, chronic insomnia, insomniac, insomnia, insomni*, sleep initiation and maintenance disorders, cognitive behavioural therapy, cognitive behavioral therapy, cognitive behavioural therapies, cognitive behavioral therapies, sleep hygiene, stimulus control, relaxation, relaxation techniques, behavior modification, behavior therapy, cognitive therapy, imagery, and psychotherapy in any language. We also reviewed the reference lists of 4 review articles on the topic (2628, 31) and briefly screened references by using the same search strategy without limitation to randomized, controlled trials. Study Selection Eligible studies were randomized, controlled trials involving CBT-i in adults (aged 18 years) with chronic insomnia. We defined CBT-i as multimodal therapy delivered in person on at least 2 occasions and incorporating at least 2 of the 5 most widely accepted components of CBT-i: cognitive therapy, stimulus control, sleep restriction, sleep hygiene, and relaxation therapy. For the primary analysis, acceptable control groups included sham therapy, waiting list, no treatment, sleep hygiene, or information provision. Studies were excluded if medical, sleep, or psychiatric comorbid conditions were listed as inclusion criteria, but they were not excluded on the basis of the frequency of comorbid conditions in included patients. We adopted this approach because excluding all studies that allowed patients with comorbid conditions would have markedly depleted the number of included studies, and because patients with chronic insomnia seen in clinical practice are likely to have a range of noninteracting comorbid conditions. Moreover, because only a subgroup of included studies reported the proportion of patients with comorbid conditions, we wished to avoid penalizing studies that reported in greater detail. Data Extraction and Quality Assessment Two authors independently confirmed the eligibility of studies, with all discrepancies resolved by consensus. One of these 2 authors extracted data, which were verified by a third author. We contacted the corresponding author of all included studies published after 1 January 2000 to request clarification of data and methods. Study quality assessments were performed independently by the 2 authors who extracted and verified data using the Cochrane Collaboration tool for assessing risk of bias (32). Data Synthesis and Analysis Our main outcome measures of interest were sleep diary measures of sleep onset latency (SOL), wake after sleep onset (WASO), total sleep time (TST), and sleep efficiency (SE%) (Table 2). These end points were assessed at 3 time points that we defined for the purpose of this analysis: immediately after treatment, early follow-up (4 weeks to <6 months after completion of the intervention), and late follow-up (6 to 12 months after completion of the intervention). Because studies most frequently reported results as the mean and SD at a point in time rather than the SD of the mean change over time, the SD of the change over time was imputed in most cases. All analyses used random-effects models, with heterogeneity assessed using the I 2 statistic and publication bias assessed with funnel plots and the Egger test (33, 34). Table 2. Glossary Six sensitivity analyses were performed that were limited to studies with particular intervention characteristics. First, we limited the analysis to studies incorporating sleep restriction because this may be among the most effective components of CBT-i (35). Second, because the optimal dosage of CBT-i is unknown but may be 4 sessions (36), we limited the analysis to studies involving at least 4 in-person contacts. Third, to consider the incremental effect of CBT-i incorporating a greater number of components, we limited the analysis to studies involving at least 4 components. Fourth, we limited the analysis to studies using a comparator group other than sleep hygiene because this may or may not be an effective stand-alone treatment (31). Fifth, we restricted the analysis to studies delivering treatment on an individual basis only, rather than in a group setting. Finally, to determine whether a tendency existed for studies with significant results to follow patients longer, we restricted the posttreatment analysis to studies with follow-up time points. In addition, we performed 3 sensitivity analyses in which we varied the correlation coefficients used to impute SDs. All statistical tests were 2-tailed, with P values less than 0.05 considered statistically significant. Statistical analyses were performed using Stata, version 13.0 (StataCorp), and R, version 3.1.3 (R Foundation for Statistical Computing). Role of the Funding Source This study received no funding. Results Our formal search strategy identified 292 references for review of the title and abstract. Of these, the full text was obtained and reviewed for 91 articles that were considered potentially appropriate for inclusion, and 20 studies met all inclusion criteria (although only 19 contributed data to the pooled estimates presented). The study flow diagram with reasons for exclusion is presented in Figure 1. Figure 1. Summary of evidence search and selection. CBT-i = cognitive behavioral therapy for insomnia; RCT = randomized, controlled trial. * Restricted to references not returned on MEDLINE search. Study Characteristics Table 3 shows descriptive data for the 20 included studies, which involved a total of 1162 patients (range, 20 to 201 patients), with values presented for only the groups that contributed data to this meta-analysis when possible. Most study populations were of late or middle age (mean age, 55.6 years), and 9 studies incorporated age restrictions as exclusion criteria. Sex was predominantly female (64.3%), and most studies were performed in developed countries (n= 19). Of the studies excluded on the basis of the population studied, 5 enrolled hypnotic-dependent patients, 1 enrolled obese persons, 3 enrolled patients with any comorbid condition, 1 enrolled patients with moderate to severe hot flashes, and 3 did not require a formal diagnosis of insomnia. Most studies referenced accepted definitions for insomnia (n= 17), most often an edition of the Diagnostic and Statistical Manual of Mental Disorders (n= 13), the International Classification of Sleep Disorders (n= 7), or both (Table 3). Nineteen studies excluded patients on the basis of presence of comorbid conditions, with all of these excluding patients with psychiatric or sleep-related comorbid conditions, 17 excluding those with medical comorbid conditions, 17 excluding those with medication or drug use, 4 excluding pregnant women, and 3 excluding shift workers. However, approaches differed in whether any comorbid condition was sufficient for exclusion or whether only patients with severe, unstable, or treatment-requiring conditions were excluded. Fourteen studies screened all patients with polysomnography at baseline to detect unrecognized sleep disorders. Table 3. Characteristics of Rando