Hypnosis Intervention Effects on Sleep Outcomes: A Systematic Review

Literature Search

Records of Ovid MEDLINE, PsychINFO, EMB Reviews, and Cochrane Central Register of Controlled Trials and Cochrane Database of Systematic Reviews were searched to include the work published before January 12, 2017. The search strategy is described in Figure 1.

Study Selection and Classification

Inclusion criteria were: (1) primary research published in English in a peer-reviewed scientific journal; (2) RCTs or prospective studies with at least 10 subjects; (3) included adults (18 years or older), regardless of health conditions; (4) outcomes of interest included measurable objective or subjective data reflecting sleep quality; and (5) hypnosis or hypnotherapy studied alone or as a part of a multicomponent intervention. Only interventions defined as “hypnosis,” “hypnotherapy” or “hypnotic” were considered. Approaches such as guided imagery or relaxation methods often used along with hypnosis were not included in this review on their own unless described as “hypnotic” to minimize heterogeneity in intervention approaches and to ensure that the interventions were likely to result in development of a hypnotic state with an increased capacity for response to suggestion inherent in hypnosis definition.¹² The full text of studies meeting the inclusion/exclusion criteria, the relevant review articles, and the articles with insufficient information to determine eligibility from the abstract were retrieved and reviewed by two authors for determination for inclusion in this review (Figure 1).

Methodological Quality Assessment

The methodological quality of the included publications was evaluated using the modified Scale for Assessing Scientific Quality of Investigation-Complementary and Alternative Medicine (SASQI-CAM) version.¹⁸ The scale was derived from previous work¹⁹ with several items added for improving the quality of research on CAM and integrative medicine approaches based on expert suggestions.^20,21 This modified 19-item scale has scores ranging from zero to 20, and it is specifically adapted to take into account special features of integrative medicine research (eg, challenges in blinding participants to the study interventions). A SASQI-CAM score higher than 9 is considered to be of sufficient scientific quality.²² Two reviewers independently assessed the risk of bias of each study in accordance with the guidelines provided for complementary and integrative approaches.¹⁸ Any discrepancies were resolved by consensus.

Data Extraction and Ratings

Both quantitative and qualitative data related to the study characteristics and outcome measures of interest were extracted for this review. For each study, data were extracted by one of the trained staff and then reviewed by a second member. Again, any discrepancies were resolved by consensus. The following data were collected: type of study, country where study was conducted, medical condition, number of participants, mean age, sex, whether sleep was primary or secondary outcome, hypnosis provider qualifications, hypnosis intervention type, intervention procedures, intervention length, length of trial, control groups, follow-up periods, rate of intervention completion, outcome measures, and results at postintervention and follow-up.

The outcomes were classified as positive when ≥ 50% of assessed sleep items had values of P < .05 or were described as more effective compared to a control group, mixed when fewer than 50% of measured items were positive; and negative when no group improved or both intervention and control groups improved but were not significantly different from each other.

Overview of the Included Studies

We identified 139 studies via searches, of which 24 met our inclusion criteria.^23–46 The data from these publications were synthesized for qualitative analyses. Quantitative analyses and meta-analytical techniques were not applied due to heterogeneity related to the diversity of assessed populations and health conditions, as well as wide variability in hypnosis interventions and outcome measures.

Table 1 provides more detailed information about the included studies. The included studies were published between 1973 and 2016. More than half of the included studies (54.2%) were conducted in the United States, with the rest conducted in European countries (37.5%) including Denmark, England, France, Spain, and Switzerland; or countries in other regions (8.3%) including Australia, Canada, and Israel. The included studies collectively assessed 1,330 participants. The average study sample size was 55.3 (standard deviation = 43.7) with the range of 14 to 187. The participants' ages ranged from 18 to 83 years (weighted average for the 19 studies where participants' ages were reported was 41.2 years), and typically most of the participants for whom the sex was reported were women (mean = 66.1%, range 21% to 100%). The studies assessed participants with diverse health conditions including insomnia or other sleep complaints (eg, sleep walking or sleep terror), depressive symptoms, PTSD, cancer, chronic pain conditions including low back, orofacial, temporomandibular pain, irritable bowel syndrome, and fibromyalgia. Studies also enrolled participants suffering from hot flashes, as well as healthy participants, including college and medical students. Most of the included studies (66.7%) did not require sleep complaints for study inclusion.

Table 1 Description of the included studies.

Among the controlled studies (75.0% of all included studies), 72.2% used active control groups (eg, relaxation, sedative-hypnotic medication, therapy, simple monitoring, structured attention, education, stimulus response), 16.7% employed non-active control groups (eg, standard treatment, waitlist), and 11.1% employed both active and nonactive control groups.

Sleep-related outcomes were primary in 50% of the included studies. The outcomes ranged from objective sleep assessment using polysomnography to custom sleep diaries. Most commonly used sleep assessments were the Pittsburgh Sleep Quality Index (PSQI)⁴⁷ and the Medical Outcomes Survey (MOS) sleep problem index.⁴⁸ Both instruments assess several dimensions of sleep quality (eg, sleep initiation, maintenance, etc.) over the previous month; both PSQI and MOS have been shown to have adequate reliability and validity in various settings and with diverse populations^47,49–53 and are intended to be standardized sleep assessments for clinicians and researchers. The custom instruments used in 29.2% of the included studies were primarily diary-like assessments; however, it is difficult to ascertain the exact nature of the assessed outcomes when such custom instruments have been used.

Forty-five percent of the included studies assessed hypnosis effects not only immediately following the intervention but also included additional follow-up assessments. Follow-up periods ranged from 1 month to 24 months postintervention.

Methodological Quality Assessment

Of the included studies, 58.3% were RCTs. This review included other controlled trials such as within-subject crossover studies^30,31 (16.7%), pre-post prospective observational studies^{32,37,41,42,46} (20.8%), and one RCT in which both arms underwent hypnosis following the completion of other treatments conditions.²⁶ This RTC was considered pre/post in regard to the hypnosis portion of the intervention.²⁶

The mean overall SASQI-CAM score¹⁸ of the included studies was 10 out of 20 possible (standard deviation = 4.3) and the range was 3 to 17. The most commonly missed quality items were SASQI-CAM item 6, related to assessing subject's expectations of benefits from the intervention, with only one publication³³ explicitly describing this, and item 10, which focused on monitoring and reporting side effects and adverse events; this information was included in only 2 publications.^33,39 More recently published work received higher methodological quality scores compared to earlier research (Table 2).

Table 2 Methodological quality of the included studies.

Comparison of Positive, Mixed, and Negative Studies

Of 24 of the included studies, more than half (58.3%) reported a positive effect of the hypnosis intervention on the assessed sleep outcome over time. Mixed results indicating improvements in some but not all sleep outcomes were presented in 12.5% of the included studies. The rest of the publications (29.2%) demonstrated negative results, with no effects of hypnosis or comparable effects of hypnosis and control conditions on sleep outcomes.

Table 3 presents comparisons of the positive, mixed, and negative studies on several variables. Positive and negative studies had similar average methodological quality scores (presented as mean ± standard deviation: 9.4 ± 4.7 and 9.9 ± 4.6, respectively), average sample size (53.6 ± 43.2 and 50.7 ± 48.4, respectively), and average intervention duration (3.8 ± 2.0 and 3.6 ± 2.0 hours, respectively). Among studies with negative hypnosis effects on sleep outcomes, 85.7% were RCTs and 85.7% used an active control group. Among the studies reporting positive effects, only 50% were RCTs, and an active control group was used in 50% of the positive studies, with 14.3% using a nonactive control group and 35.7% employing no control group. Samples with sleep complaints were recruited in 35.7% of the studies reporting a positive effect of hypnosis on sleep and in 28.6% of the studies reporting negative hypnosis effects on sleep. Next, sleep was the primary outcome in 57.1% of the studies showing benefits of hypnosis for sleep and in 42.9% of the studies showing no hypnosis effect on sleep.

Table 3 Comparisons between positive, mixed, and negative studies.

Description of the Subset of Studies Targeting Sleep

Only one-third of the reviewed studies had inclusion criteria requiring participants to have sleep disturbances. Considering only the studies recruiting samples with sleep problem, most of them (62.5%) reported positive findings, 12.5% reported mixed results, and 25.0% reported negative results. All except one of these studies were controlled, and two studies focused on mechanistic aspects of hypnosis for sleep outcomes rather than efficacy. Table 4 shows comparisons of the positive, mixed, and negative studies that enrolled samples with sleep disturbances. Average sample sizes were 36.4 ± 14.7, 108.0 ± 0, and 29.5 ± 7.5 for positive, mixed, and negative studies, respectively. Average probability of bias based on SASQI-CAM scoring appeared similar and high for studies reporting benefit of hypnosis on sleep and studies reporting no benefits of hypnosis on sleep outcomes (6 ± 3.2, and 6 ± 3.0, respectively). Treatment duration averaged 2.5 ± 2.5 for positive and 2.5 ± 0.5 for negative studies. Most of the studies enrolling samples with sleep problems assessed sleep with subjective assessments, except one study reporting positive outcomes of hypnosis on sleep.

Table 4 Information about the subset of studies enrolling samples with sleep complaints.

Eighty percent of the positive studies from this subset recruited samples with psychiatric comorbidities such as PTSD, anxiety, or depression; none of the studies reporting negative results indicated significant comorbidities in their participants. Sleep was the primary outcome in 62.5% of the studies showing benefits of hypnosis for sleep and in both of the studies showing no hypnosis effect on sleep.

Description of the Subset of Studies with SASQI-CAM Scores Indicating Lower Risk of Bias

Twelve of the included studies received a score higher than 9 on SASQI-CAM, indicating adequate study quality and lower risk of bias.²² The patterns observed for this subsample of the reviewed publications paralleled those observed for the full sample. Seven studies (58.3%) demonstrated an overall positive effect of the hypnosis intervention on the assessed sleep outcome. Two studies (16.7%) showed mixed results, indicating improvements in some but not all sleep outcomes. The remaining three studies (25%) demonstrated negative results, with no beneficial effects of hypnosis or comparable effects of hypnosis and control conditions on sleep outcomes.

Table 5 presents comparisons of the positive, mixed, and negative studies scoring higher than 9 on SASQI-CAM on several variables. Positive and negative studies had similar average sample size (71.0 ± 48.9 and 79.7 ± 56.1, respectively), but average intervention duration was 2.9 ± 2.2 hours for positive studies and 4.7 ± 0.5 hours for negative studies. Except for two crossover placebo-controlled studies reporting positive results, the rest of the high-quality studies were RCTs. All high-quality studies used a control group, with most including active controls. Samples with sleep complaints were recruited in one of the studies reporting a positive effect of hypnosis on sleep and in none of the studies reporting negative hypnosis effects on sleep. Next, sleep was the primary outcome in 57.1% of the studies showing benefits of hypnosis for sleep and in none of the studies showing no hypnosis effect on sleep.

Table 5 Information about the subset of studies with lower risk of bias (SASQI-CAM score higher than 9).

Hypnosis Intervention Procedures

Hypnosis is an approach that can be easily customized and adapted to a client with the choice of specific hypnosis techniques and individualized suggestions, and it also can be added to another therapy. This flexibility is one of the attractive features of this therapeutic approach. In most included studies (79.2%), hypnosis was used as a stand-alone intervention and the remaining studies combined hypnosis with another therapeutic intervention (eg, cognitive behavioral therapy, cognitive processing therapy). Few of the included studies (8.3%)^25,43 reported using sleep hygiene instructions along with hypnosis training. The studies varied greatly in the lengths of hypnosis intervention: from 13-minute audio recordings with hypnosis for deep sleep to 8 sessions of hypnosis, each lasting 1 hour. Most of the studies (83.3%) used individual hypnosis sessions. Two of the included studies (8.3%) used group hypnosis sessions^45,46 and the remaining 8.3% of the studies used an audio recording only.^30,31 The number of sessions averaged 4 ± 1.2 (range, 1 to 8), with the total time spent in intervention averaging 3.6 ± 1.9 hours (0.23- to 8-hour range).

Home practice was encouraged in 54.2% of the included studies, with most of these studies offering audio recordings for such a practice. Adherence to the home practice instructions was not typically assessed.

One included study⁴⁴ compared different hypnosis treatment packages and their effects on pain and sleep quality outcomes. The findings indicated no significant effect of hypnosis on sleep outcomes but showed that after hypnosis participants had a significantly greater reduction in the study primary outcome (pain intensity) than after control biofeedback treatment. Interestingly, there was no significant difference among the three hypnosis conditions (8 self-hypnosis group training sessions with audio recordings for home practice, 8 self-hypnosis group training sessions without recordings, or 2 self-hypnosis group training sessions with audio recordings for home practice and weekly reminder telephone calls). The authors concluded that two sessions of self-hypnosis with audio recordings for home practice may be as effective as 8 sessions of hypnosis treatment for pain management. It is of interest to assess the optimal number of sessions of sleep-focused hypnotherapy for individuals with sleep problems.

The hypnosis interventions described in the review targeted sleep disturbance (25%), pain symptoms (29.2%), and immune system (4.2%), focused on increase of relaxation and reduction stress (12.5%), symptoms associated with hot flashes (12.5%), and the remaining studies provided no specific intervention focus.

For studies specifically targeting insomnia symptoms, hypnosis approaches involved using a metaphor of a fish going deeper into the water for deeper sleep, using an age-regression technique prompting the participant to focus on his or her earlier memories of periods when restorative sleep could be easily achieved, providing directions for relaxation with visualizations including relaxing scenes, and delivering ego-strengthening suggestions.

In the studies targeting pain symptoms, suggestions included instructions for dissociating from pain, controlling pain perception, imagined analgesia, decreased pain unpleasantness, and managing breakthrough pain. Hypnoanalgesic techniques taught in the studies included age regression, dissociations, fantasy, changing pain image, glove anesthesia, and hypnoplasty. Hypnosis for reducing hot flashes symptoms incorporated suggestions for relaxation and coolness. Also commonly included were suggestions for relaxation, safe place imagery, coping with stress, and suggestions for effective hypnosis practice. Several studies (20.8%) specified that the sessions were individually tailored with customized suggestions.

Hypnotic susceptibility was assessed and reported in 37.5% of the included studies. In these studies, the assessment was completed with the Harvard Group Scale of Hypnotic Susceptibility⁵⁴ (55.6%), Stanford Hypnotic Susceptibility Scale^55,56 (33.3%), or the Hypnotic Induction Profile⁵⁷ (11.1%). Some studies found that only highly susceptible participants benefitted from hypnotic suggestions for sleep improvement and those with low hypnotic susceptibility did not.^30,31 The authors of one study indicated that participating in hypnotic susceptibility assessment prior to intervention might have influenced the participants' expectations for success in hypnosis.³⁶ The other studies reported no effect of hypnotic susceptibility on sleep outcomes. Most of the studies assessed in this review did not include the assessment of hypnotic susceptibility or tested its relation to the study outcomes.

Adverse Event Reporting

One of the goals of this review was to assess adverse events in the published studies utilizing hypnosis. Hypnosis is typically described as the form of treatment with low or no associated side effects⁵⁸; however, the practitioners of hypnosis are often cautioned that hypnosis may cause unpleasant side effects that may include dizziness, confusion, unpleasant somatic symptoms, false memories, or panic attacks in some people.⁵⁹ The risk of moderate to severe after-effects of hypnosis was previously estimated to be 7% in research and clinical samples but it is believed that adverse events due to hypnosis may be underreported in clinical settings because these complications are typically short-lived.⁶⁰ Therefore, the true incidence of such adverse events is not clear. The evidence from this review suggests that few studies utilizing hypnosis interventions report monitoring or assessing adverse experiences. Only two reviewed studies^39,61 totaling 231 participants mentioned some form of monitoring of adverse events. All reported no adverse experiences associated with hypnosis. One study,³⁶ though not specifically assessing adverse experiences, reported that some people in the hypnosis group showed general dissatisfaction with the procedure.

Overview of the Qualitative Analyses

Overall, more than half of the included studies indicated some positive effects of hypnosis on sleep outcomes. However, the body of evidence, in general, was of variable methodological quality. The studies reporting positive findings had a similar average methodological quality and sample sizes compared to those reporting negative hypnosis effects. However, on the one hand, more than 80% of the negative studies were controlled trials and a similar percentage had an active control group compared to only half of the positive studies having an RCT design and using an active control group. On the other hand, the studies reporting positive effects of hypnosis on sleep outcomes were more likely to have a sample with sleep problems and focus on a sleep outcome as a primary outcome measure. The results regarding outcomes were similar when only a subgroup of studies with samples experiencing sleep problems was considered. Furthermore, when only a subset of studies with higher SASQI-CAM scores indicating sufficient quality of the study was considered, the patterns were similar to those observed in the whole sample.

The hypnosis interventions were primarily delivered as separate treatment programs in several individual sessions with increasing interest in using tailored suggestions. Using sleep hygiene instructions along with hypnosis was reported in few hypnosis interventions for sleep improvement. For the studies targeting sleep and conditions affected by sleep (pain, stress, hot flashes) home practice was encouraged in most of the included studies, though adherence to such practice and its influence on the outcomes was rarely assessed. In studies that reported hypnosis provider, the interventions were delivered by a trained clinical professional with training and expertise in clinical hypnosis. The adverse event rate was very low given that more than 200 participants from two RCTs reported no adverse events; however, the information about adverse events was not reported in most studies.

Implications for Clinical Practice

A recent review assessing efficacy, safety, and applications of medical hypnosis did not list sleep problems among the suggested potential applications of medical hypnosis in daily clinical practice.⁵⁸ The reason for this may be that, to date, rigorous research evaluating hypnosis interventions for sleep disturbances is lacking. Our review indicates that hypnosis may be a promising approach for sleep improvement that merits further investigation.

Several attractive features of hypnosis supported by the presented evidence should be highlighted. First, successful hypnosis interventions for sleep may be relatively short. Our review indicates that on average, 3 to 4 sessions were sufficient to obtain benefits for sleep outcomes in the studies reporting positive effects of hypnosis on sleep. The minimum number of sessions needed for an effective treatment remains open for debate and may depend on individual tailoring, with evidence leaning towards multiple (3 to 5) sessions with an individually tailored treatment plan. Though most of the studies reporting benefits of hypnosis on sleep were using individual sessions, benefits of hypnosis were also observed in studies utilizing group sessions and audio recording only, highlighting that sleep improvements can be achieved with more efficient interventions. The research also suggests that home practice supported by audio recordings of sessions may boost the efficacy of treatment and reduce the number of treatment sessions.

Second, hypnosis is a flexible approach with a big toolbox of techniques that can be tailored to the specific individual and their symptoms. Our review indicates that beneficial hypnosis for sleep improvement included sleep-specific suggestions, suggestions for relaxation and stress reduction, and age regressions. Another promising strategy for hypnosis treatment for sleep improvement may be an inclusion of ego-strengthening suggestions in the treatment sessions. These suggestions that are not focusing on specific sleep symptoms may be beneficial for improving therapeutic alliance, achieving heightened insight and better self-esteem in the patients and thus optimizing the treatment effects from hypnosis.⁶² Sleep hygiene instructions may also be included in a hypnosis treatment program for sleep improvement. Though inclusion of sleep hygiene instructions was not commonly reported in the studies included in the review, sleep hygiene may boost treatments targeting sleep.⁶³

Next, though hypnotic susceptibility is believed to be important in how people respond to hypnosis, few included studies assessed it. The data are not conclusive about the role of hypnotic susceptibility in the success of the hypnotic interventions. Only some studies showed the importance of hypnotic susceptibility for the success of the intervention^30,31 whereas others failed to show it or did not assess it. It was also suggested that patients in the midrange of hypnotic susceptibility may be the most appropriate clients for hypnosis treatments for some health conditions,²⁵ but this aspect needs further investigation.

Further, hypnosis appears to be an approach with a low incidence of adverse effects. Though only a few studies reported adverse event monitoring in this review, those that included this information reported no adverse events. Additionally, a high rate of completion of the hypnosis interventions observed in the included studies (Table 1) suggests general acceptance of this modality by patients and study participants. Sleep problems are typically treated with sleep medications that have known unpleasant side effects and have an unknown efficacy for long-term use.^2,7 These may be some of the reasons why those affected by sleep problems are likely to seek alternative approaches for management of insomnia.¹¹ Empirical evidence from published research indicates either a beneficial effect of hypnosis interventions on sleep or hypnosis effects comparable to those of control interventions with a low rate of adverse experiences from hypnosis. Therefore, this approach may be important to explore further for its utility for treating sleep problems especially in populations in which using pharmacological treatments may be undesirable (eg, children, elderly, patients with multiple comorbidities, etc.)

Currently, a nonpharmacological approach of choice for treating insomnia is CBT-I.^8,64 A recent systematic review and meta-analysis evaluating cognitive behavioral therapies for treatment of insomnia indicated that CBT-I is efficacious for some but not all dimensions of sleep problems.⁶⁴ Though CBT-I has been considered a reliable treatment for chronic insomnia for the past 25 years, some limitations noted in the recent work include the need for 5 or more sessions of therapy to observe most optimal response and a shortage of trained therapists to provide the CBT-I.^8,64 Hypnosis can be provided by any clinician (physician, nurse, psychologist) trained in hypnotherapy, and diverse patient populations can be trained to induce beneficial changes in health using self-hypnosis without the presence of a clinician as demonstrated in previous research assessing hypnosis for other health issues.^33,65–69 Further, hypnosis is extremely flexible and can be performed in different settings (not just as individual, group, or on-line sessions similar to CBT-I but also as audio recorded sessions); the studies included in this review indicate that a beneficial effect can be obtained in four or fewer sessions with hypnosis; however, the optimal number of sessions for best results is yet to be determined. To date, the rigorous research directly comparing efficacy and cost-effectiveness of CBT-I and hypno-therapy for sleep is scarce. Of interest, few studies evaluating interventions combining cognitive therapy approaches with hypnosis indicate beneficial synergistic effects,^29,39,70 suggesting that hypnosis might be a useful adjunct therapy for other psychotherapy approaches.

Our review indicates that research on using hypnosis for sleep problems is still in the early stages. Because of several attractive features of hypnosis previously described and promising research to date showing potential benefits of hypnosis in the absence of serious adverse events, hypnosis for sleep disorders deserves further evaluation.

Limitations

The review is limited in several respects. The included studies assessed relatively small samples of predominantly middle-aged female participants who may or may not have sleep complaints. Therefore, it is not clear how these results can be applied to the general population of individuals experiencing sleep problems.

Our selection procedures included filtering results based on language (English only). Furthermore, we allowed hypnosis interventions to focus on conditions other than sleep as long as sleep outcomes were included in the data. All of these issues may have affected the generalizability of these results to the general population with sleep problems.

Another limitation is the methodological diversity of the studies and the extent of bias of the included studies. Our review search strategy cast a broad net to include studies that reported at least one sleep-related outcome. Overall the quality of the included research was modest, with only half of the included studies receiving a score higher than 9, indicating lower risk of bias on SASQI-CAM. Only one-third of the studies included in this review recruited samples with sleep problems. Further, hypnosis intervention focused on sleep improvement in only a portion of the included studies; and several studies used interventions focused on pain, immune system, or hot flashes symptoms. We also included studies of different designs in addition to RCTs. Therefore, there was considerable heterogeneity in the included studies related to study samples, outcome measures, and clinical interventions, which complicates the qualitative analyses and interpretation of the results.

Additionally, though SASQI has been specifically designed for evaluating integrative medicine approaches and has been increasingly used in systematic reviews assessing efficacy of integrative approaches,^18,22,71 its reliability and validity have yet to be formally assessed, thus potentially limiting objectivity of the study quality assessment.

Some limitations are specific to integrative medicine approaches such as hypnosis; for example, formal blinding of participants and clinicians to treatment arms is not always possible in trials assessing mind-body approaches. Thus, the extent to which the awareness of the group assignment affected the study outcomes is difficult to predict.

Recommendations for Future Research

Important future directions related to hypnosis include determining minimal effective dose (ie, the length of treatment) and delivery method (individual versus group versus recording), as well as the optimal combination of suggestions and techniques for sleep improvement hypnosis. Another important topic to explore is determining specific populations who would most likely benefit from hypnosis and specific populations for whom hypnosis is not an optimal intervention. Future studies must consider symptomatology, hypnotic susceptibility, expectancy, adherence to treatment, and other patient characteristics. Methodological quality of future research could be enhanced by careful selection of control conditions; assessment of participant expectancy of improvement; use of reliable sleep assessments; reporting adverse events; monitoring provider adherence to intervention protocols; reporting participant adherence to the intervention; providing detail on clinician qualifications in hypnosis; and use of longer follow-up periods to assess outcomes.