Effectiveness of Cognitive Behavioural Therapy to Improve Sleep Outcomes in Patients with Schizophrenia: A Systematic Review of Randomised Control Trials ()
1. Introduction
Schizophrenia, a type of psychosis, affects more than 20 million people globally (World Health Organization, 2022), and 1% of people experience psychotic symptoms at least once in their lives in the UK (NICE, 2014). As per Diagnostic Criteria (DSM-V), schizophrenia traits can be expressed through positive symptoms (hallucinations and delusions), negative symptoms (affective flattening, avolition, and alogia), disorganized speech, and catatonic behaviour (American Psychiatric Association, 2013; Royal College of Psychiatrists, 2017). Individuals may develop psychosis due to heredity, environmental stress, or adverse childhood experiences (Batinic, 2019). Sleep disruptions are strongly associated with psychotic experiences (Laskemoen et al., 2020) and can predict psychosis among at-risk populations (Reeve et al., 2018).
Insomnia is defined by difficulty commencing, maintaining, and restoring sleep, resulting in daytime dysfunction, mood variants, poor concentration, and fatigue (NICE, 2020; Sedky et al., 2020). The sleep’s quality is poor enough to carry on day to day functioning properly, continue with educational goals successfully or maintain social relationships normally. The client never feels refreshed no matter how much sleep he has been gone through. He feels that he is usually awake while lying on the bed for long hours in order to get a sound sleep. It can be a disorder or coexist with other health conditions, increasing treatment burden and relapse of the disorder (Khurshid, 2018). There are two kinds of insomnia: acute and chronic. Acute insomnia, usually associated with environmental stress, lasts for less than three months, whereas chronic one sustains for a long time and is a common symptom in psychiatric/medical conditions, especially among psychotic patients (Afonso et al., 2011). It is estimated that about 80% of people with psychosis also experience sleep disruptions (Klingaman et al., 2015).
People with schizophrenia experience insomnia, mainly due to the coexisting factors such as sleep-related fixed beliefs, trauma, lifestyle, lack of motivation, and medication side-effects which affect their functioning, leading to disability (Chiu et al., 2016). Moreover, conditions such as hallucinations, delusions, irregular circadian rhythm, daytime inactivity, and bedtime fear also contribute to sleep disturbances (Waite et al., 2016). Conversely, sleep disturbances lead to cognitive deficits (Davies et al., 2017), weak processing speed and inhibition (Laskemoen et al., 2020), which put people at increased risk of suicidality (Miller et al., 2019; Malik et al., 2014), obesity (Palmese et al., 2011), stroke, diabetes (David et al., 2016) and depression and anxiety (Blanchard et al., 2020). Moreover, the causal link of insomnia with psychotic events (Reeve et al., 2017; Barrett et al., 2020) indicates a potential to treat it in the first place.
To treat comorbidity between insomnia and psychosis, antipsychotics are often referred (NICE, 2014); however, drugs may prove unhelpful or cause severe side effects to some patients. Therefore, to prevent the patients from delayed rehabilitation due to sleep deprivation, health care professionals recommend psychosocial interventions as the preferred treatment (Davidson et al., 2019; Brabban et al., 2009). These interventions include Open Dialogue, Psychosocial Inpatient/outpatient Treatment, Major Role Therapy (Cooper et al., 2020), and Music Relaxation Therapy (Bloch et al., 2010). Among these, Cognitive Behavioural therapy is considered as the most structured and beneficial technique to deal with insomnia comorbidity with psychosis (Joober et al., 2017) delivered face to face (Chiu et al., 2018; Hwang et al., 2019) or online (Freeman et al., 2017; Denis et al., 2020).
1.1. Cognitive Behavioural Therapy
CBT is basically problem-solving focused therapy. It aims to alter thinking patterns within patients to activate positive change in behaviour. Cognitive psychologists study the cognitive areas (like thought processing, problem solving, perception, attention, memory and language) scientifically and have established framework in areas: schema theory, information processing model, and neuroscience model (Dobson & Dozois, 2019). CBT also deals with cognitive areas (negative perceptions of the world, negative beliefs, negative automatic thoughts, and negative concepts) termed as automatic thoughts (Kosugi et al., 2019). CBT not only attempts to change the present situation; rather it focuses on the change in the behaviour. At first it brings a change in perception or cognitive schemata, that sets a path for a change in behaviour which can also be observed. Thus, it makes the patients capable of changing their thought patterns and makes them able to take the steps to bring a change in their lives (Sheldon, 2011).
CBT based on educational model is based on the assumption that all behavioural reactions are learned therefore it aims to unlearn the behaviour that client wants to get rid of. Second important trait of CBT is that its techniques rely on the inductive methods which is based on the facts, not assumptions. The client is ready to test his self-created hypothesis and after testing its validity in the real world; he begins to think about changing his thinking according to the new information and understanding of the situation. Third important trait of CBT is that it uses the Socratic Method in which therapist tries to gain a deep understanding of clients’ concerns (Clark & Egan, 2015). This is the best way for a therapist to find out the real causal factors to diagnose or treat clients’ problems accordingly.
Evidence suggests that CBT, as an effective technique to deal with positive and negative symptoms while linking the contexts one perceives and behaves (Batinic, 2019; Millan et al., 2012), improves sleep quality (Ye et al., 2016) through controlling stimulus, restructuring thoughts, and relaxation training (Robertson et al., 2019). A systematic review conducted by Heller and colleagues (2021) through searching literature between 2002-2019 showed the effects of Cognitive Behavioural Therapy on pain perception when it is associated with self-knowledge and cognitive restructuring. It can be administered individually, at a group level, face to face, or online. These days digital CBT has gained much popularity and success (Freeman et al., 2017) with cost-effectiveness and feasibility (Luik et al., 2017). However, patients in the UK are still not getting the best treatment from clinicians (Barrett et al., 2020) due to either the complexity of CBT or the challenges involved in the implementation processes in terms of time, cost, and patient lifestyle (Rehman et al., 2017). Therefore, to deal with such barriers and improve the intervention’s adaptability among psychotic patients, a systematic review on CBT’s effectiveness to improve sleep quality will be carried out.
1.2. Objective
This study explores the mechanism through which CBT works effectively in the context of experiencing sleep disruptions under the paradigm of psychosis. It will explore the dimensions and the contexts under which CBT help improve the patients’ sleep quality (primary outcome) and psychological/emotional well-being (secondary outcome). This information will help policymakers and health professionals to consider it before implementation.
2. Method
2.1. Inclusion Criteria
2.1.1. Type of Studies
Randomised Control Trials, peer-reviewed, with full-text availability, in English language, published in the last ten years (2010-2020), will be included. CBT is a widely discussed intervention in a range of mental disorders (Hofmann et al., 2012); however not so frequently discussed for insomnia in psychosis. Other systematic reviews discuss the intervention either within the context of multiple disorders (Thoma et al., 2015; Pigeon et al., 2017) or early psychosis (Davies et al., 2017; Hertenstein et al., 2019). Therefore, the search will be restricted to only RCTs to get evidence-based knowledge and insight for future recommendations.
2.1.2. Type of Participants
All adults (≥18), regardless of gender identity and ethnicity, diagnosed with schizophrenia or schizoaffective disorder or with psychotic symptoms (diagnosed by DSM-V or ICD-10), with short- or long-term conditions associated with sleep disturbances will be considered. However, participants diagnosed with sleep Apnoea and Parasomnia disorder will be excluded. Besides studies with mixed age groups, children, adolescents, or subgroups like army or prisoners will not be included. The sample of less than 10 participants, being underpowered, will also be excluded (Biau et al., 2008).
2.1.3. Type of Interventions
This review will include CBTI, the most preferred intervention to assess the sleep outcomes (Khurshid, 2018), including all modalities. Its efficacy has been established for both psychiatric disorders (e.g., schizophrenia) as well as non-psychiatric disorders (e.g., insomnia) (Chand et al., 2020). However, due to the disorder’s complexity and its adjunctive nature with medication (Wong et al., 2019), antipsychotics cannot be excluded from the intervention.
2.1.4. Types of Comparators
The alternative therapy will be either a placebo or a standard treatment as usual or a control group.
2.1.5. Types of Outcome Measures
In this review, both standardized self-reported and objective (wrist actigraphy) measures to assess sleep outcomes will be included. The outcomes reporting sleep quality (improvement or reduction), alone or with other psychotic experiences, will be considered.
2.2. Search Strategy
An online bibliographic search will be performed through databases: PubMed, Psych Info, and Cochrane Library using terms: “schizophrenia” OR “psychosis,” OR “schizoaffective disorder” OR “psychotic symptoms” AND “sleep disturbance” OR “insomnia” OR “sleep disruption” following the similar procedure used by Robertson et al. (2019). Through this reiterative process, all retrieved peer-reviewed research published within the last ten years (2010-2020), meeting initial criteria, will be exported to RefWorks Citation Manager. These studies will be further screened for duplication, predetermined criteria, and objectives to have a final list of studies to be included in this review. The PICO framework can be viewed in TableA3 (see Appendix B).
2.3. Data Extraction and Quality Assessment
The selected papers will be assessed, and the following information will be extracted: study aims, study design, year of publication, sample size, sampling method, sample characteristics, sleep assessment tools, dimensions of CBT, characteristics of the control condition, sleep outcomes, time of measurement, follow-up, and setting.
The articles’ quality will be assessed through the checklist provided by the Critical Appraisal Skills Programme, UK (CASP, 2018).
3. Results
3.1. Selection of Studies
In this process, at first databases PubMed, Cochrane and PsychInfo were explored for the relevant papers while applying the search criteria (see Appendix B) and 597 papers were selected for further screening. These papers were further examined for full text RCTs and 82 papers were extracted. Finally, while applying the inclusion and exclusion criteria, five most relevant RCTs were included in this review (see Figure 1).
The RCTs (1: Chiu et al., 2018; 2: Denis et al., 2020; 3: Freeman et al., 2015; 4: Freeman et al., 2017; & 5: Sheaves et al., 2018a) selected for this review have also been assessed for their quality (CASP, 2018) while assigning a number 1 for each “yes” out of total score “9” indicating low (0 - 3), moderate (4 - 6) and high quality (7 - 9). The three of these (1, 2, 5) qualified for the moderate, and two (3, 4) for high quality (TableA2: Appendix A).
Figure 1. Prisma flow diagram for the systematic review.
3.2. Trials Characteristics
The summary of all the characteristics of data can be viewed in TableA1 (Appendix A). Among these, most studies were conducted in the UK (2, 3, 4, 5) and one (1) in Australia. The three studies were delivered face-to-face (1, 3, 5), and the other two (2, 4) were administered online.
Three were pilot trials (2, 3, 5) mainly planned to evaluate the acceptability, adherence, cost effectiveness and efficacy of the therapy in target population. Participants were mostly British white, adult inpatients (5), outpatients (1, 3), and university students (2, 4). The mean age of the participants was 40 (median) years ranged from 20 to 41.4, and the mean proportion of male to female was 50%, including studies with male-only (5) and female-only (2). The sample size ranged from 40 to 3755.
Most participants had symptoms of insomnia (diagnosed or self-reported), and in most of the studies (1, 3, 5), were diagnosed with non-affective disorder. Various tools have been adopted in these studies to measure insomnia symptoms. Pittsburgh Sleep Quality Index (PSQI) in study 1, Sleep Condition Indicator (SCI) in 2, Insomnia Severity Index (ISI) in 3, ISI and SCI-8 in 4, and ISI was used in study 5 to assess sleep disruptions. All studies used self-reported measures to assess sleep outcomes except study 3, in which actigraphy was used to measure sleep-wake cycles.
CBT was delivered in six (3, 5) to eight (1, 2, 4) flexible sessions, following distinct assessment timing as well as separate follow-ups for each study starting from 6 weeks (1), through 12 weeks (5), 22 weeks (4), 24weeks (3) until six months (2). CBTI in study 1 completed in 4 - 6 weeks with four weekly sessions. The “Sleepio” intervention in studies (2, 4) was delivered in six weekly online sessions. The most extended face-to-face schedule was 8 sessions over 12 weeks in study 3 (minimum dose 4). The most intensive STAC intervention in study 5 was delivered in 2 weeks with eight sessions (minimum dose 5).
The highest retention rate was in study 5 (100%), 96% provided follow-up in study 3, 80% completed more than two sessions in study 1, 69% reached mid intervention and 47% completed follow-up in study 2; and in study 4, 69% attended one, 50% two and 18% attended six sessions.
In all studies, CBTI has shown a significant improvement in insomnia symptoms ranging from small to medium effect in study 2 (d = 0.42) to large in 1, 3, 4 and 5 (d = 0.66; d = 1.9, d = 1.1, d = 0.9) respectively.
Regarding secondary outcomes, most studies (3, 4, 5) have reported significant improvement in psychological wellbeing with small to medium effect size (d = 0.3, d = 0.29, d = 0.3) respectively, as well as significant reductions in distress (1: d = 0.4) and stress (2). Some secondary outcomes related to psychotic symptoms being out of scope will not be addressed in this review.
3.3. Strengths
All the studies were Randomised Control Trials, the most rigorous method to ensure the reliability and validity of the results regarding research biases (Ernest et al., 2015). While avoiding selection bias, participants in all studies were randomly allocated to groups (treatment vs. control) except a study (1) in which it was not explicitly described. Similarly, other than one open-label study (1), in all trials, accessors were blind to the group allocations, minimising the effect of interpretation bias. To avoid the confounding variable effects, baseline similarities were also assessed in all studies.
All studies followed a systematic statistical procedure and provided evidence (95% CI) for the reliability and validity of results. Study 2 measured adherence to the therapy and study 3 reported fidelity of the therapy as well.
3.4. Weaknesses
Blinding the assessor, participants, and investigators from the allocation groups is always preferable to produce vigorous results in RCTs (Renjith, 2017) which is often not feasible in behavioural sciences. In this review, it was also not realistic to keep the participants/therapists blind from their group allocation due to the intervention’s specific nature. Thereby the intervention group may be influenced by the whole procedure but not specifically by the intervention. To resolve this issue, it is advised (Kendall, 2003) to keep the two groups similar. In this connection, only one study (2) has assigned an additional task, “puzzles,” to the control group to compensate the intervention group’s cognitive burden.
Moreover, the imbalance between males and females in all studies except one (1), warrants ascertainment bias, and the small sample size in most in-person trials could not provide promising results. However, variations in sample size, CBTI components, therapy sessions, and adherence are likely to be found equally in both delivery modes (Matthews et al., 2013; Ye et al., 2016) and pose severe validity threats.
3.5. CBT Evaluation
All studies utilized standard CBT elements: psychoeducation, goal setting, stimulus control, sleep restrictions, belief restructuring, circadian rhythm, articulatory suppression, paradoxical intention, daytime activity, mindfulness, relaxation techniques, and sleep hygiene, with few adaptations. For instance, in addition to exploring the effect of CBT on insomnia, study 1 also aimed to explore the sleep subtypes among psychotic patients and assessed the effectiveness of CBTI in clusters. Accordingly, clusters representing severe insomnia, regular sleep, and hypersomnia would need different strategies to manage sleep disruptions. Study 5 incorporated CBT elements in Sleep Treatment at Acute Crisis (STAC) intervention and adopted light/dark therapy (Wirz-Justice et al., 2013) to adjust biological clock in patients. Study 3 adapted CBT for the delusions and hallucinations.
Thus, in this systematic review, after careful evaluation of 597 studies, only five studies were included as per set inclusion/exclusion criteria. Among these Randomised Control Trials, Cognitive Behavioural Therapy was adapted for psychotic patients and insomnia symptoms improved alone or in combination with other psychological and physical symptoms.
4. Discussion
CBT in this review has shown a significant improvement in insomnia symptoms. This result is consistent with other studies (Hwang et al., 2019; Robertson et al., 2019). However, the primary aim to investigate the mechanisms underpinning CBT to reduce insomnia was not the objective of the studies included in this review. In this context, study 1 has offered a unique solution to enhance CBT’s effectiveness on sleep outcomes by creating sleep profiles (1: classic severe insomnia; 2: insomnia with normal sleep; 3: insomnia with hypersomnia), revealing the significant effect of CBT on insomnia reduction. In addition to sleep profile, CBT has also shown improvement in daytime efficiency in people with psychosis as per their response type (strong responder, non-responder, selective responder) (Waters et al., 2020).
A similar perspective was discussed in study 5, in which a consistent sleep window was introduced as per patient type (morning type and evening type). Thus, given the patient’s preference for chronotype, the implementation of therapy adapted. Consisting with this, Bradley et al. (2018) have utilized circadian rhythm regulation techniques to improve young people’s sleep quality.
Similarly, studies (1, 3) have shown a change in total sleep time and sleep efficiency by adopting strategies such as stimulus control, psychoeducation, beliefs restructuring, and sleep restrictions. Extending to this, a transdiagnostic treatment approach (Sheaves et al., 2018b), while utilizing some creative adaptations to CBT, has shown significant benefits of using objective measures for insomnia patients.
Online studies offer certain advantages over face-to-face regarding privacy, flexibility, and cost-effectiveness (Luik et al., 2017). Participants had flexibility to adopt the therapy at their own pace as per their needs (study 4). These designs are usually incorporated with an online library, community support, and message reminders through emails (study 2). Other web-based approaches based on fully automated systems provide evidence for the effectiveness of CBT’s digital application (Espie et al., 2012). However online studies (2, 4) have shown a higher dropout rate (50%), which is a common factor in online studies (Scott et al., 2020).
All face-to-face trials in this review have shown a significant improvement in sleep outcomes. However, these trials have been conducted by trained clinical psychologists in highly controlled settings with strict schedules. It is not always feasible in real life, providing interventions with great intensity and validity (Addington & Lecomte, 2012). For instance, in study 3, 41% attending CBT, and 4% from the control group had no insomnia at week 12. This remarkable result could be due to extensive care (e.g., calls and text messages between sessions) and robust methodology to implement the intervention.
Moreover, these trials recruited in-active control groups which estimate an absolute effect size rather than a relative effect size (Karlsson & Bergmark, 2015). Trials with active control groups (Espie et al., 2012) can further explore the therapy effectiveness from various perspectives (Karlsson & Bergmark, 2015).
5. Limitations
The pilot trials in this review were not primarily conducted to assess which CBT elements bring a change in insomnia. Therefore, the focused information was not explicitly available. The second limitation was that the relevant studies included either comorbidities (excluding criteria) or psychosis as excluding criteria in their research. Moreover, the variant samples, like patients diagnosed with non-affective psychosis (1, 5, 3) with severe (1, 3, 4, 5) and mild (2) insomnia symptoms and university students (2, 4) with mild psychotic symptoms, restrict the generalizability of the results to a broader population.
6. Future Research
While enhancing the generalizability of the results, new research should be focused on more homogeneous samples, sharing more similarities in terms of psychological and physical health conditions. Similarly, it will be helpful to consider cultural diversities and different belief systems to implement the intervention accurately (David et al., 2016). Insomnia is a contributing factor in psychosis (Reeve et al., 2018; Robertson et al., 2019). Therefore, assessing sleep outcomes associated with other psychiatric comorbidities and medication adherence would also be valuable.
Additionally, web-based interventions, being widely available in sleep disorders should also extend to mental health disorders (Pot-Kolder et al., 2020) and need to be integrated with traditional in-person approaches (Luik et al., 2017). Nevertheless, in addition to subjective reports, objective measures (watches) may also be incorporated in therapy (Sheaves et al., 2018b) to enhance credibility and patient motivation (Chiu et al., 2016).
7. Conclusion
CBTI has been successfully adapted across large populations, from patients diagnosed with a non-affective disorder to the general population with mild psychotic symptoms through an automatised online system or intensive face-to-face sessions. Still, to understand underlying mechanisms and to increase the availability and effectiveness of CBT on a large scale, more robust clinical trials would be required.
Appendix A
Table A1. The summary of data characteristics extracted from the five selected articles.
Table A2. The results extracted from the CASP checklist for the selected five papers in this review.
Interpretation of Scores: Low quality 0 - 3, Moderate quality 4 - 6, High quality 7 - 9 (CASP, 2018).
Appendix B
The PICO framework can be viewed from TableA3.
Table A3. The PICO framework for the review.
1. Chiu, V. W., Ree, M., Janca, A., Iyyalol, R., Dragovic, M., & Waters, F. (2018). Sleep Profiles and CBT-I Response in Schizophrenia and Related Psychoses. Psychiatry Research, 268, 279-287. https://doi.org/10.1016/j.psychres.2018.07.027
2. Denis, D., Eley, T. C., Rijsdijk, F., Zavos, H. M. S., Keers, R., Espie, C. A., Luik, A. I., Badini, I., Derveeuw, S., Hodsoll, J., & Gregory, A. M. (2020). Is Digital Cognitive Behavioural Therapy for Insomnia Effective in Treating Sub-Threshold Insomnia: A Pilot RCT. Sleep Medicine, 66, 174-183. https://doi.org/10.1016/j.sleep.2019.10.007
3. Freeman, D., Waite, F., Startup, H., Myers, E., Lister, R., McInerney, J., Harvey, A. G., Geddes, J., Zaiwalla, Z., Luengo-Fernandez, R., Foster, R., Clifton, L., & Yu, L.-M. (2015). Efficacy of Cognitive Behavioural Therapy for Sleep Improvement in Patients with Persistent Delusions and Hallucinations (BEST): A Prospective, Assessor-Blind, Randomised Controlled Pilot Trial. The Lancet Psychiatry, 2, 975-983. https://doi.org/10.1016/S2215-0366(17)30328-0
4. Freeman, D., Sheaves, B., Goodwin, G. M., Yu, L.-M., Nickless, A., Harrison, P. J., Emsley, R., Luik, A. I., Foster, R. G., Wadekar, V., Hinds, C., Gumley, A., Jones, R., Lightman, S., Jones, S., Bentall, R., Kinderman, P., Rowse, G., Brugha, T. et al. (2017). The Effects of Improving Sleep on Mental Health (OASIS): A Randomised Controlled Trial with Mediation Analysis. The Lancet Psychiatry, 4, 749-758. https://doi.org/10.1016/S2215-0366(17)30328-0
5. Sheaves, B., Freeman, D., Isham, L., McInerney, J., Nickless, A., Yu, L.-M., Rek, S., Bradley, J., Reeve, S., Attard, C., Espie, C. A., Foster, R., Wirz-Justice, A., Chadwick, E., & Barrera, A. (2018a). Stabilising Sleep for Patients Admitted at Acute Crisis to a Psychiatric Hospital (OWLS): An Assessor-Blind Pilot Randomised Controlled Trial. Psychological Medicine, 48, 1694-1704. https://doi.org/10.1017/S0033291717003191