Psychological Interventions for Irritable Bowel Syndrome
Irritable bowel syndrome (IBS) patients often present psychoform symptoms or psychiatric disorders. Among the psychological factors studied in IBS patients, two seem to influence mostly its severity: catastrophizing and somatization. Somatization is an independent risk factor for IBS. In addition, somatization more than the severity of IBS influences the way the patients perceive their illness, the outcome and the efficacy of treatment. Irritable bowel syndrome patients demonstrate greater catastrophizing scores than controls, and pain catastrophizing is a significant predictor of gastrointestinal symptoms related to pain. In this context we analysed the data regarding the efficacy of two psychological treatments in IBS: cognitive behavioral therapy and hypnosis. Cognitive behavioral therapy is focused on replacing maladaptive coping strategies with more positive cognitions and behaviors. Several studies showed that cognitive behavioral therapy is effective in reducing bowel symptoms in IBS, both post-treatment and short-term follow-up. Gut-directed hypnotherapy has beneficial short-term effects in improving gastrointestinal symptoms of patients with IBS, and the results are maintained after one year in half of the patients. Psychological treatments are a suitable option for selected IBS patients.
The advent of the biopsychosocial model of functional gastrointestinal disorders (FGIDs) represented a milestone in the understanding, diagnosis, and management of these disorders. Since the development of the biopsychosocial model, about 30 years ago, there is increasing evidence that psychological factors play an important role in the pathogenesis and outcome of FGIDs, especially in irritable bowel syndrome (IBS).
Very often there is an association between IBS and
psychiatric disorders. Even if most IBS patients do not have psychiatric illnesses, they do manifest psychoform symptoms and somatoform complaints . There are some unanswered questions regarding the association observed between psychological factors and IBS. Is this high occurrence of psychiatric disorders or maladaptive coping in IBS a characteristic of IBS patients, or do they represent a risk factor? Do these factors influence the self-reported severity of IBS? Do they modify the self-reported outcomes of IBS, i.e. general health status, quality of life (QoL)? Which of these factors can we influence through therapy to obtain symptom alleviation?
Interestingly, van Tilburg et al.  developed a model of psychological influences on IBS. The authors collected data regarding several psychological factors such as neuroticism, abuse history, life events, anxiety, somatization and catastrophizing from 286 IBS patients. In their model, the two most important variables associated with IBS severity were catastrophizing and somatization. Anxiety had an important but indirect effect because it was associated with both increases in catastrophizing and somatization. Other psychological variables such as neuroticism and stressful life events also play a role in IBS patients by increasing anxiety. The authors concluded that “the most fruitful approach to curb the effects of psychological factors on IBS is to reduce catastrophizing and somatization” . This model did not include depression, but there are data showing that depression favors catastrophic thinking and through this thinking style patients experience more intense pain . In a previous paper , all the psychosocial factors studied throughout the time in relation with IBS were presented and analysed. This review will focus on the latest data since 2010 on catastrophizing, somatization and psychological treatment in IBS.
SOMATIZATION AND IBS
Somatization disorder (SD) is the tendency to report multiple unexplained physical symptoms (e.g. headache, back pain, fatigue), without the evidence of a medical condition that could explain them . Up to one quarter of IBS patients referred to tertiary care centers have SD, but the prevalence of SD in population-based samples vary from less than 1% to 10% [6, 7]. Somatization explains the frequent extra gastrointestinal (GI) symptoms reported by IBS patients: musculoskeletal complaints, urinary and sexual symptoms, headaches and fatigue . Several papers have shown that psychosocial factors indicative of somatization (such as illness behaviour scores, anxiety, sleep problems and somatic symptoms) and somatization are independent risk factors for IBS [7, 9, 10]. In a large community based study from England, subjects free of IBS that reported all markers of somatization at baseline (i.e. high illness behavior scores, anxiety, sleep problems and somatic symptoms) were six times more likely to develop IBS five months later when compared to those who were exposed to none or only one marker. Among these markers, the illness behavior scale was the stronger predictor of outcome (OR = 5.2, 95%CI 2.5-11.0) . Similar results were reported by a recent study that included 2769 subjects from a screening program, among which 258 had IBS. Irritable bowel syndrome was associated with somatization, with an adjusted OR = 2.88, and 95%CI 1.55-5.36 .
One important observation that can be drawn looking at the research in the last years is that somatization more than the severity of IBS influences the way the patients perceive their illness, the outcome, the efficacy of treatment or interventions and the impact on their lives [11-13]. It is generally accepted that patients with IBS have a poor quality of life . Vu et al.  looked for the impact of psychiatric (anxiety, depression, somatization) and extra-intestinal functional disorders on the health-related quality of life (HRQOL) in patients with FGIDs (functional dyspepsia and IBS). In a cohort of 606 FGIDs outpatients, more than 70% had at least one of the above mentioned comorbidities. The presence of these comorbidities influenced the QoL independent of GI symptoms. In addition, comorbidities were stronger predictors of HRQOL than GI symptoms in FGIDs patients . In another study, the presence of somatic and psychological complains (somatization, anxiety, stress, depression) and medical comorbidities, and not the severity of IBS symptoms, correlated with the scores of self-rating of health. The authors concluded that the severity of IBS symptoms has a modest role in how IBS patients describe their health in general .
Patient-reported outcomes (PROs) are instruments developed by FDA and are used to capture clinically important information regarding the therapeutic benefit of treatment in FGIDs. Lackner et al.  studied the influence of cognitive processes on PROs in IBS. The authors examined how factors related with the respondent, such as pain catastrophizing, somatization and anxiety sensitivity are associated with PRO endpoints of severity and concluded that these dysfunctional cognitions observed in IBS influenced patients’ judgments of pain, bloating or bowel habit, in the end affecting the PROs. It is interesting to mention the role of somatization in predicting each GI symptom. Every one unit increase of somatization predicted an increase in IBS severity scores by >0.33, or just over 1/3 of a scale unit. In the end, higher scores of somatization were associated with higher ratings of IBS symptoms . Similar results upon the interplay between somatization and IBS severity are reported by other studies published in the last years .
CATASTROPHIZING AND IBS
Catastrophizing is a maladaptive coping strategy defined as “a negative cognitive process of exaggerated negative rumination and worry” . Pain catastrophizing is the tendency to magnify the seriousness of pain, or feel helpless about it either in direct response to pain or in anticipation of painful stimuli . Irritable bowel syndrome patients demonstrate greater catastrophizing scores than controls . Pain catastrophizing is a significant predictor of GI symptoms related to pain . Catastrophizing, more than pain severity, influences the variance of QOL in IBS  and mediates the relationship between depression and pain severity. Patients with IBS who experience higher levels of depression engage in more catastrophic thinking specific to pain, and partly through this thinking style experience more intense pain and greater activity limitations due to pain . These results of Lackner et al. from 2004 are supported by the work of van Tilburg et al. who showed that catastrophizing and somatization play an important and direct role in IBS symptom severity . A recent paper showed different catastrophizing levels in different countries: the highest scores were observed in China and Romania, compared to other countries . A hypothetical explanation is the influence of living in a specific political regime.
A brief questionnaire to assess catastrophic cognitions in IBS has been developed and validated by Hunt et al. . This is called GI-Cognitions Questionnaire and is able to discriminate between IBS and inflammatory bowel disease patients.
PSYCHOLOGICAL TREATMENT IN IBS
Catastrophizing, somatization and anxiety sensitivity influence patients’ judgments of pain, bloating and bowel habit, thus affecting the PROs . This observation highlights the value of cognitive-behavioral interventions that address the way that patients with IBS experience, interpret, and respond to their symptoms. The model proposed by van Tilburg et al. pointed out two modifiable psychological factors, which have a major influence on IBS: catastrophizing and somatization . To reduce catastrophizing one can try to „reduce stress”, or better, to improve coping with stress. Cognitive behavioral therapy (CBT) is focused mainly on replacing maladaptive coping strategies with more positive cognitions and behaviors. Among psychological therapies, CBT and hypnosis have been used in research studies and showed the highest rates of positive results, and both will be discussed in this article.
HOW DO CBT AND GUT-DIRECTED HYPNOSIS WORK?
There is no clear consensus in the literature with regard to the mechanism of CBT on reducing IBS symptoms. The biopsychosocial model explains how an increased IBS susceptibility is influenced by the interactions between early life factors (genetic predisposition and environmental factors), psychosocial risk factors (e.g. depression and anxiety, coping styles) and physiological dysfunctions (e.g., motility, visceral sensitivity) via the brain-gut axis [25, 26]. At the same time patient responses to IBS symptoms influence the evolution of the disease. Everitt et al.  explained how a patient’s emotions, cognitions and behaviours regarding the symptoms, can increase anxiety and maintain symptoms through the link between the heightened autonomic nervous system and the enteric nervous system. Cognitive behavioral therapy supposedly reduces the impact of CNS activity on gut function  and intervenes on IBS clinical manifestations through education, behavioural and cognitive techniques .
With regard to the CBT mechanism in IBS, several studies suggest that the change may occur through: (1) improvement in psychological distress which leads to improvement in GI symptoms ; (2) direct effect of CBT on GI symptoms that in turn may lead to reductions in psychological distress [28, 29]; (3) changes first of the behaviour, then of the cognition, before impacting on treatment outcome [28, 30]; decrease of visceral sensitivity and of catastrophic cognitions . An interesting aspect mentioned by the authors of a 2015 metaanalysis  is that psychological treatments for IBS, which have the same treatment labels (e.g. cognitive therapy), may not be comparable. The improvements of IBS symptoms may be due not to just one reason, but to several. For example, one reason might be the different hypothesis tested (e.g. improving depressive symptoms versus reductions in visceral hypersensitivity), another reason might be in regard to different techniques applied for the same intermediate theoretical predictors, or a third reason could be that two treatments with different labels may actually target the same intermediate theoretical constructs and utilize the same intervention techniques .
Identifying and understanding the active ingredients of psychological therapies on IBS is crucial , but the differences in the design studies may transform this into an objective hard to complete. These differences start from the diagnostic procedures: according to the NICE (National Institute for Health and Care Excellence) guidelines  the diagnosis is recommended to be made in a positive manner, considering the symptoms which point towards the diagnosis, but in practice exclusion diagnostic procedure is still in use . Studies do not differentiate between the IBS subtypes (IBS- constipation, IBS-diarrhoea, IBS-mixed and IBS-unclassified), which recently have been identified as having different impacts on the QoL . Also some design studies include different diagnostic criteria for IBS (Rome I, II or III) , while others have focused on different mediators. For example, one study showed that change in illness perceptions, catastrophizing, damage beliefs and fear avoidance predict improved work and social adjustment in IBS patients . Another mediation study, that used a protocol focused on changing distress, disagreed with previous findings and argued that the pathway to improvement in IBS was through distress [26, 28]. While some say that symptom-focused approaches  may be ineffective, others state that focusing solely on psychological variables could potentially be ineffective. For example, the IBS patients with purely somatic illness attributions have difficulty in acknowledging the role psychological variables play in maintaining IBS symptoms [32, 36, 37].
Thus, future investigations on the processes and mechanisms through which CBT operates should: (1) have better research designs, with appropriate control groups , (2) investigate which CBT technical components are the keys in making changes (e.g. on bowel symptoms) [26, 29], (3) analyse if the specific skill used in treatment improves after treatment (e.g. mindfulness) , (4) make attempts to measure intermediate constructs at baseline and post-treatment, and (5) try to examine the mediation of symptom improvements through these intermediate constructs quantitatively .
The mechanism through which gut-directed hypnotherapy (GDH) works on IBS symptoms is also not clear. As stated above, hypnotherapy is a form of intervention where suggestibility is used in order to “disable” the more active and analytical aspects of the patient’s mind, whereas aspects of the normally subconscious mind is activated .
The central pain amplification (part of IBS symptom generation), includes increased emotional arousal, cognitive abnormalities (hypervigilance, catastrophizing) and related alterations in the balance of descending inhibitory and facilitatory systems [40, 41]. Functional MRI studies have demonstrated that hypnotherapy appears to lead to normalization of abnormal central pain processing [33, 40], and that amitriptyline reduces brain activation during painful rectal distension [33, 42]. Meanwhile, a report where recording of cerebral-evoked potentials in normal individuals undergoing rectal electrical stimulation was made before and after hypnotherapy, suggested that hypnotherapy reduced cerebral responsiveness [43, 44]. Gerson et al.  believe that the visualization which accompanies gut-directed hypnotherapy reconfigures associations between pain, urgency, and selfcontrol, relieving a level of anxiety that may perpetuate a negative brain-gut interaction.
Cognitive behavioral therapy is a form of psychotherapy where the therapist helps patients to actively dispute their irrational beliefs and to assimilate more efficient rational beliefs, with a positive impact on their emotional, cognitive, and behavioral response . There are different forms of CBT; it is usually administered over weekly sessions (may vary from 8 to 20 sessions), and can be combined with behavioral interventions such as encouraging patients to engage in activities that counter the disability associated with the bowel disorder . Almost all studies published so far showed that CBT is efficient: it reduces the bowel symptom score, improves the QoL [29, 46, 47] and also help patients to better cope with their chronic illness . In general, most studies were small- to-medium-sized trials and most small studies may not have been adequately powered . Only a few studies failed to show the superiority of CBT versus standard care on bowel symptom severity or HRQoL .
Table I shows the studies performed in the last six years, which compared different forms of CBT (internet delivered, based self-management, using personal digital assistant or sessions with a trained psychotherapist) either with usual medical care, or with other interventions (i.e. relaxation therapy). The majority of these studies showed the superiority of CBT in improving bowel symptom severity and some changes in cognition when compared with usual medical care or patients on waiting lists, both in short and long term follow-up. A meta-analysis from 2014 showed that CBT is more effective in reducing IBS bowel symptoms than waiting list controls both at post-treatment and short-term follow-up, but it is not superior to other psychological treatments. The study included 18 randomized controlled trials (RCTs) .
An interesting paper comes from Craske et al. who compared different protocols of CBT . Twenty five patients underwent CBT-interoceptive exposure that targeted erroneous beliefs about IBS symptoms, hypervigilance to IBS symptoms, hypersensitivity to visceral sensations, fear of IBS symptoms, and maladaptive behavioural responses to IBS symptoms. Bowel symptom severity index, visceral sensitivity index, pain vigilance and awareness, IBS-QOL were determined pre-, mid-, post- CBT and at follow-up. The intervention was superior to control intervention addressed to stress management or attention control .
Table I. Studies published in the last 5 years which evaluated the efficacy of cognitive behavioral therapy in patients with irritable bowel syndrome
|Study||Number of patients||Type of therapy||Main results / conclusions|
|Jang et al. ||76||8 weeks CBT / general information on IBS||Bowel symptom severity, distress, disability, interference with activity, healthy worry, body image improved|
|Chilcot et al.
|64||CBT / treatment as usual||Change in cognition rather than mood mediated treatment-related improvements|
|Ljotsson et al.
|195||Internet-delivered CBT / internet- delivered stress management, 10 weeks||The improvement on the gastrointestinal symptom rating scale-IBS version was higher in ICBT when compared with ISM, both post-treatment and at 6 months follow-up|
|Andersson et al. ||85||Internet-delivered CBT/ discussion forum||Symptom improvement (> 50% reduction in the symptom score) – 36% vs. 2% of patients
Significant cost reductions for the treatment group at $16,806 per successfully treated case; reduced work loss in the treatment group. Results were sustained at 3-month and 1 year follow-up. NNT = 2.99 (95% C 2-5)
|Oerlemans et al. ||37 / 38||CBT on personal digital assistants / control , 4 weeks||CBT group showed more improvement in QoL, catastrophizing thoughts, and more pain improvement
Only improvement in catastrophizing thoughts persisted in the long-term
|Jones et al. ||34 / 36 / 35||CBT / relaxation therapy / usual medical care||CBT operates via changes in mood state (most clearly anxiety)|
|Ljotsson et al. ||85||Internet-CBT / discussion forum||Participants in the treatment arm reported 42% decrease in IBS-symptoms, while the control group reported a 12% increase in symptoms|
|Lackner et al. ||71||Wait list/10 weekly sessions CBT / four CBT sessions over 10 weeks||Rapid positive response to CBT at 4 weeks is associated with sustained IBS symptom reduction|
|64||7 weeks CBT based self- management/ treatment as usual||At 3 and 6 months post-treatment 76% of CBT patients reported symptom relief vs. 21% in the treatment as usual group|
|Hunt et al. ||28/26||5 weeks internet CBT||Improvement in symptoms score and QoL after 3 months|
|Hunt et al. ||60||CBT; selffielp||Improvement of symptoms, visceral sensitivity, catastrophization, QoL at 6 weeks|
|Ljotsson et al. ||417||Internet CBT with/without systematic exposure||10 weeks internet-delivered intervention with exposure is superior|
CBT: cognitive behavioural therapy; IBS: irritable bowel syndrome; ICBT: Internet-delivered CBT; ISM: internet-delivered stress management; NNT: number needed to treat; QoL: quality of life
Another therapeutic option in IBS is GDH. The first report about the efficacy of hypnosis in IBS dates from 1984  and since then a high number of clinical trials have supported its effectiveness. Through hypnosis patients are brought to a special mental state where they are intensely focused and receptive, followed by deep relaxation and use of gut-directed imagery and suggestions related to symptom control and normalization of gut function . The treatment requires approximately 7 sessions (each session lasts 30-40 minutes), either weekly or every other week. How GDH improves the symptoms of IBS remains a dilemma. Palsson et al. showed that hypnosis improved both IBS symptoms and general bodily symptoms, but did not affect rectal pain thresholds, rectal smooth muscle tone, and autonomic functioning (sweat gland activity, heart rate, blood pressure, skeletal muscle tension, and skin temperature). The authors concluded that the positive effect of hypnotherapy could be the consequence of the reductions in psychological distress and somatization .
A study that included 204 IBS patients reported that 71% of patients treated with hypnosis responded to treatment (they considered their symptoms very much or moderately better). Among the responders, 81% fully maintained improvement at follow-up 1 to 5 years later. Treatment responders used significantly less medication and had fewer health care visits . Hypnosis improves not only the IBS symptom score, but also the IBS-related cognitions score and the total cognitive score, suggesting that symptom improvement in IBS with GDH is associated with cognitive change . Moser et al. compared GDH (46 patients) in 10 weekly sessions within 10 weeks with supportive talks with medical treatment (SMT – 44 patients), and showed that a higher rate of patients improve after GDH, both after treatment (60.8% vs 40.9%) and after 15 months (54.3% vs. 25.0%) . The effects of GDH are long lasting. Vlieger et al. reported that after a mean duration of 4.8 years (3.46.7) 68% of patients treated with GDH were in remission vs. 20% of patients that received conventional therapy (p=0.005). The study was conducted on 52 children with IBS or functional abdominal pain. Remission was defined as > 80% improvement in pain scores compared with baseline. Pain scores and somatization scores were also significantly lower in the GDH group compared with conventional therapy group .
Two meta-analyses published in 2014 tried to establish if GDH is superior to conventional therapy in IBS patients [67, 68]. One meta-analysis included 7 RCTs and 374 patients among which 191 were in the hypnotherapy group. The studies targeted refractory IBS. There was a high variability among studies with regard to duration (30 to 60 minutes) and frequency (5 to 12 sessions) of GDH sessions. Gut-directed hypnotherapy significantly improved the following parameters at three months: the overall gastrointestinal symptom score, abdominal pain and several aspects of QOL. The authors underlined that firm conclusions cannot be drawn given the small number of studies, but GDH has beneficial short-term effects in improving GI symptoms of patients with IBS . The second metaanalysis included 8 RCTs and 464 patients. This meta-analysis showed beneficial effects of GDH compared with conventional therapy both at the end of therapy (GDH determined symptom relief and reduced global GI score) and at 1 year follow-up, when symptoms were still alleviated compared with controls in half of the patients. Number needed to treat (NNT) was 5 (at 3 months) and 3 (at 1 year follow-up) , comparable with NNT with antidepressants . Another meta-analysis from 2014 concluded that CBT, hypnotherapy, multicomponent psychological therapy, and dynamic psychotherapy were all beneficial in the treatment of IBS .
Currently, psychological interventions (CBT, hypnotherapy and/ or psychological therapy) for IBS treatment are included in 2015 NICE guidelines. The guideline recomends that patients with refractory IBS, who do not respond to pharmacological treatments after 12 months should be referred for psychological interventions .
As resumed in Table II, somatization and catastrophizing influence the severity of IBS and impact several aspects of IBS patients life. So far, the exact mechanism explaining the interaction between IBS and psychological factors, their frequent association (even with psychiatric disorders) is not known, but recent data suggest that molecular or genetic changes could be involved .
Table II. Somatization and catastrophizing in relation to irritable bowel syndrome (IBS) and psychological therapies
Somatization and factors indicative of somatization (esp. a high illness behavior score) are independent risk factors for IBS
Presence of somatization influences the way the patients perceive their illness, the outcome and the impact on their lives
Catastrophizing influences the variance of quality of life in IBS
Catastrophizing and somatization play a direct role in IBS symptom severity
CBT aims to replace maladaptive coping strategies with positive cognitions and behavior
CBT is effective in reducing IBS bowel symptoms
GDH has beneficial short-term effects in improving gastrointestinal symptoms of patients with IBS
After GDH, IBS symptoms are alleviated in 50% of patients after 1 year follow up
CBT: cognitive behavioral therapy; GDH: Gut-directed hypnotherapy
The biopsychosocial model of IBS remains valid through decades of research. Among psychological factors, two appear to influence mostly IBS outcome and the way patients with IBS interpret their illness: somatization and catastrophyzing. In addition, somatization is an independent risk factor for the development of IBS. Luckily there are methods to diminish their “effect” on IBS through psychological therapies. Although the exact mechanism is not yet known, CBT and GDH have proved to be efficient in the majority of studies and their beneficial effects last sometimes for years. For IBS patients who fail to respond to standard medical treatment, we should keep in mind these alternatives.
Cognitive-behavioral therapy for irritable bowel syndrome: A meta-analysis
Irritable bowel syndrome (IBS) is a chronic, relapsing gastrointestinal symptom complex characterized by altered bowel habits and abdominal pain and discomfort, and it affects as many as 5%-20% of individuals in the population [1,2]. The prevalence of IBS is modestly higher in women, and women are more likely to exhibit the constipation-predominant subtype and less likely to meet the criteria for the diarrhea-predominant subtype than men . IBS represents an economic burden on society and decreases IBS patients’ health-related quality of life [4,5].
The current treatments for IBS are challenging and unsatisfactory . The medical management tends to provide inadequate relief of IBS bowel symptoms , whereas the clinical trials of psychological therapies have demonstrated some improvements, especially cognitive behavior therapy (CBT). Notably, CBT has proven to be an effective therapy for both depression and anxiety disorders [8,9]. In regard to the treatment for patients with somatization and symptom syndromes, CBT appears to be a promising treatment . Although the etiology and pathogenesis of IBS remain elusive, it is recognized that patients with IBS are more likely to suffer from coexistent mood disorder, depression and anxiety than healthy controls [11,12]. Thus, CBT might also be an effective and promising treatment strategy for IBS.
The cognitive behavioral model defines how events, thoughts, emotions, actions and physiological responses interact with each other. CBT as applied to IBS includes several main steps. The first step is to educate, which consists of the explanation of IBS symptoms and the CBT model. At the same time, the patients are encouraged to find the psychological factors that are interacting with their physical symptoms. Then, the patients and the therapist work together to identify the potential associations among their thoughts, emotions and actions with IBS symptoms. Lastly, behavioral therapy, such as stress management is applied .
There has been some CBT for IBS studies published, including several separate systematic reviews or meta-analyses, that address whether CBT improved the outcome in IBS [13-17]. These systematic reviews all held the view that CBT was superior to the waiting list controls. However, the evidence of CBT for IBS is controversial when compared with different types of active controls. Shen and Nahas  found that CBT was possibly not superior to education or psychoeducational support. In contrast, Kearney and Brown-Chang  concluded that CBT was possibly better than education and support. Hutton  stated that the effect of CBT was at least as great as the medical treatment for IBS. In recent meta-analyses, Ford et al.  concluded that CBT was superior to waiting list controls or physicians’ «usual management” in IBS, and Zijdenbos etal.  also found that CBT was better than usual care or waiting list in improving symptoms and quality of life but was not superior to placebo. The evidence for the efficacy of CBT might be positive in treating IBS in these reviews [13-17]. However, these recent systematic reviews arrived at disparate conclusions, especially regarding the evidence of CBT for IBS being controversial when compared with active controls other than waiting list controls, and the validity of CBT follow-up has not been established. Finally, several important RCTs published after 2009 were not included in these previous meta-analyses. In this meta-analysis, we attempt to address these discrepancies and provide an up-to-date conclusion to establish the efficacy of CBT for IBS.
Fig. 1. Flow diagram for the assessment of studies identified in the meta-analysis.
To identify the relevant studies, we conducted a search of PubMed, Scopus, the Cochrane Library and Web of Science up to December 31, 2013. The keywords used for IBS and CBT are presented in Appendix A. Randomized controlled trials (RCTs) examining the effects of CBT in adult patients with IBS were eligible for inclusion (see inclusion criteria below).
For the full-text reading and final evaluation, we only included studies published in English. Conference abstracts were not included in our analysis because of the limited data available. Two reviewers (Li & Zhang) independently selected studies that met the predetermined inclusion criteria, and all potentially relevant papers were obtained and evaluated in detail. Any disagreement between investigators was resolved by discussion until consensus.
Manuscripts were included if they met the following criteria:
(a) adult participants (over the age of 16 years old) with underlying IBS;
(b) studies with randomized controlled research design and the cross-over study with available data of post-treatment outcomes;
(c) treatment arm with CBT (self-management CBT, CBT delivered through face-to-face, telephone or web-based, CBT organized by group or individual format, etc.; these types of CBT have been identified as having the same effect as conventional CBT [18,19]);
(d) adequate controls (waiting list, physician’s usual management, medical treatment or psychological treatment, etc.); and (e) measurable outcomes reported.
The exclusion criteria for the meta-analysis were as follows: (a) not RCT, (b) duplicated trials that included articles that used subsamples from larger studies, (c) studies that used other form of CBT as controls and (d) studies with insufficient data, unless in the studies the authors were able to provide adequate data. After inclusion and excluded, 18 studies remained for analysis (see Fig. 1).
Data collection and methodological quality
Two ofour authors subsequently collected the data from the articles meeting the inclusion criteria separately including the following items: author and year, country of origin, mean age, female (%), diagnostic criteria, intervention (method, operator, duration and length of follow-up),control categories, primary outcome measures, secondary outcome measures, intent-to-treat (ITT) data and the Cochrane Collaboration Depression and Anxiety Neurosis Review Group’s (CCDAN) scale score  (see Table 1). We calculated data such as mean age and female percentage of patients from the manuscripts as far as possible. The outcome measures that were related to our meta-analysis were extracted. The methodological qualities and the risk of bias in individual studies were independently evaluated by the two researchers using the CCDAN scale, which consists of 23 items [21 ]. The description of the CCDAN scale and the quality scores of each item of the included studies are presented in Appendix B.
The primary objective of this study was to evaluate the effect of CBT compared to controls on IBS bowel symptom severity. The effect measurements included almost all the available scales at present, such as The Composite Primary Symptom Reduction (CPSR) Score , Irritable Bowel Syndrome Severity Scoring System (IBS-SSS) , Bowel Symptom Severity Scale (BSSS) , Gastrointestinal Symptom Rating Scale–IBS version (GSRS-IBS) , IBS symptom score [26,27], abdominal pain and Rome II scores [28,29]. The secondary outcomes of this study included improvement of IBS QOL scores and psychological states as evaluated by the Hospital Anxiety and Depression (HAD) Scale [30,31], Montgomery Asberg Depression Rating Scale–Self report (MADRS-S) , Beck Depression Inventory (BDI)  or health depression score . All evaluations were finished after cessation of treatment, at short-term follow-up and at long-term follow-up with the available data.
Data synthesis and statistical analysis
We included all studies about CBT for IBS compared to controls. The variations of the controls might be a source of the heterogeneity of the meta-analyses. To sort out the sources of the potential heterogeneity, all analyses of the outcomes were classified into subgroups according to the specific types of the controls. The first subgroup included the studies with the symptom-monitoring and waiting list control groups, which included nine trials [19,34-41]. We combined the studies with the control groups of «treatment as usual”, «routine clinical care”, «standard care” and «self-help support group” as the second subgroup, which included four studies [28,36,42,43], as all of these controls allowed the patients to receive basic support from the gastroenterologist or a «fact sheet” for IBS. The third subgroup included three studies in which the control groups received medical treatments, whereas the treatment groups received CBT in addition to medical treatments [26, 29,44]. In one study , the medication used was 270 mg of mebeverine taken thrice daily. In the other two studies [26,28], drugs were prescribed according to the patient’s symptoms. The fourth subgroup included three trials that involved all relative psychological treatments, such as «stress monitoring”, «relaxation” and «psychoeducation” [43,45,46]. The majority of CBT types contain elements or procedures of cognitive therapy. On the other hand, the CBT procedures in the studies included in our study were variable and multicomponent, some of which also included the steps of relaxation [26,28,34,37,42,43], stress monitoring [35,42], psychoeducation , etc., which were applied to be the control treatments in other studies [43,45,46]. In the trials that had three treatment arms, we divided one study into two circumstances [36,43,45]. The data of the CBT arm were applied twice by being controlled with the other two control groups separately. In one study , subjects were randomly assigned to three groups: group cognitive treatment (GCT), individual cognitive treatment (ICT) and symptom monitoring waiting list, and it was verified that cognitive therapy delivered in group format was as effective as cognitive therapy delivered in an individualized manner. We only extracted the data of the intervention group delivered individually.
Characteristics of the eighteen studies included in the meta-analysis
|age||(%)||Method||Operator||Duration||Length of follow-up||outcome||outcomes||data||score|
|Payne and Blanchard||USA||40.1||88.2||Rome criteria||CBT||Therapist||8||3||SG or SMWL||CPSR||BDI||NA||24|
Blanchard et al. (2007)
|USA||49.2||73.3||Rome II||GCBT||Psychologists and advanced doctoral students||10||3||PE or SM||CPSR||NA||yes||29
|Boyce et al. (2003)||Australia||42.3||81||Rome I||CBT||Clinical psychologists||8||4.5,12||SC or relaxation||BSSS||HAD||yes|
|Ljotsson et al. (2011)||Sweden||34.9||74||Rome III||ICBT||Clinical psychologists||10||12||WL||GSRS-IBS||IBS-QOL||NA|
|Hunt et al. (2009)||USA||38.5||81.5||Self-report had been diagnosed with IBS||ICBT||Therapist||5||3||WL||GSRS-IBS||IBS-QOL||yes|
|Kennedy et al. (2005)||England||NA||NA||Rome I||CBT plus||Trained primary||6||3, 6,12||Mebeverine alone||IBS-SSS||HAD||NA|
|Greene and Blanchard (1994)||USA||38.2||75||Clinical criteria||CBT||Therapist||8||3||SMWL||CPSR||BDI||NA|
|Ljotsson et al. (2010) Ljotsson et al. (2011)||Sweden
|Rome III Rome II||ICBT
|Graduate psychology student
Psychology students and psychologists
|IBS-QOL; MADRS-S IBS-QOL; HAD||NA
|Oerlemans et al. (2011)||Netherlands||38.3||84.2||Rome III||CBT||Psychologist||4||3||SC||Abdominal pain||IBS-QOL||yes|
|Mahvi-Shirazi et al. (2012)||Iran||NA||NA||Rome II||CBT plus medical treatment||Psychologist||8||3||Medical treatment||ROME-II score||NA||NA|
|Sanders et al. (2007)||USA||51||78.3||Rome II||S-CBT||Treatment book||8||3||WL||CPSR||IBS-QOL; BSI||NA|
|Heymann-Monnikes||Germany||37.8||91.7||Rome criteria||CBT plus medical||Clinical psychologists||14||6||Medical treatment||IBS-SS||BDI||NA|
|et al. (2000)||treatment|
|Haghayegh et al. (2011)||Iran||NA||45.8||Rome II||GCBT||Clinical psychologists||8||2||WL||NA||IBS-QOL; BDI||NA||21|
|Vollmer and Blanchard||USA||43.5||56.3||Rome criteria||CBT||Therapist||10||3||GCBT; SMWL||CPSR||NA||NA||25|
Tkachuk et al. (2003)
|Canada and USA||39.5||96||Rome criteria||GCBT||Therapists||9||3||SM||Abdominal pain||BDI||NA||21|
|Moss-Morris et al. (2010)||England||39.5||73||Rome I/Rome II||S-CBT||Psychologist||8||6||Treatment as usual||IBS-SSS||HADS||yes||33|
|Jarrett et al. (2009)||USA||44.3||86.4||Rome II||CSM-IP||Two research nurses||9||3, 6, 12||CSM-T/IP; UC||IBS-SS||IBS-QOL; BSI||NA||23|
Abbreviations:ICBT = CBTdeliveredviainternet; GCBT = group-basedCBT; S-CBT = self-administeredCBT;PE = psychoeducationalsupportgroup; SM = stressmonitoringgroup; ISM = Internet-deliveredStressManagement; SG = self-help support group; SC = standard care; WL = waiting list control; SMWL = symptom monitoring waiting list control; CPSR = The Composite Primary Symptom Reduction Score; GSRS-IBS = the Gastrointestinal Symptom Rating Scale–IBS version; IBS-SS = The IBS symptom score; IBS-SSS = The Irritable Bowel Syndrome Severity Scoring System; BSSS = the Bowel Symptom Severity Scale; IBS-QOL = Irritable Bowel Syndrome Quality ofLife score; HAD = The Hospital Anxiety and Depression; BDI = Beck Depression Inventory; MADRS-S = The Montgomery Asberg Depression Rating Scale–Self report; BSI = Brief Symptom Inventory; CSM-T/IP = Comprehensive Self-Management-Telephone; CSM-IP = Comprehensive SelfManagement-In-Person; UC = usual care; NA = not available.
Stata 12.0 software was used for the statistical analyses. The SMD values with 95% CIs were calculated for the continuous data, and a random effects model was used. Studies with multiple outcomes were categorized as above (see Table 1) and then grouped together within each domain. These controlled effect sizes could then be interpreted conservatively with Cohen’s convention of small (0.2), medium (0.5) and large (0.8) effects .
We developed three methods, including IBS bowel symptoms improvement, QOL and psychological states, to evaluate the application of CBT for IBS. Analyses were performed for post-treatment, short-term follow-up (two to six months from post-treatment) and long-term follow-up (nine to 12 months from post-treatment). When the studies presented two or more data analyses at short-term or long-term followup simultaneously that had been defined beforehand, we extracted the data of the longer time point. There were different rating scales in each section. If one study reported its result with an opposite scale direction, the result of the study was multiplied by – 1. When evaluating IBS bowel symptoms and QOL, a positive effect size always indicated improvement, whereas a negative effect size of psychological state indicated improvement. We contacted the authors of the trials in which data were not fully reported or could not be calculated manually from the information presented in the article. Overall, we successfully contacted the authors of one study. When the necessary data were available, the SMD and its 95% CI were calculated. Cochrane’s Q-test and I-squared test were adopted for assessing heterogeneity. Studies are considered heterogeneous when the Cochran’s Q-test probability is lower than 0.05. Heterogeneity was classified into low (< 25%), medium (и50%) and high (>75%) levels based on I-squared test. Publication bias was examined using Egger’s test as well as the funnel plots.
Fig. 2. Effect size estimates for the efficacy of CBT compared to controls in IBS symptom improvement at post-treatment.
Our review followed PRISMA guidelines.
NOTE: Weights are from random effects analysis. The four subgroups were classified by the controls. The specific treatments are presented above.
Fig. 3. Effect size estimates for the efficacy of CBT compared to controls in IBS symptom improvement at short-term follow-up.
A total of 1008 citations were identified, of which 18 were finally selected for analysis (see Fig. 1). We handsearched the references of the previous reviews and found no additional studies. As methods for assessing three-arm trials were introduced in our study, the total number of analyzed circumstances was greater than 18. Treatment durations varied from five to14 weeks, and the follow-up periods varied from two months to one year. The proportion of female patients recruited by all trials ranged from 45.8% to 96% , and the mean age of the patients ranged from 34.5  to 51 years old. The CBT operators included therapists, psychologists, doctoral or graduate clinical psychological students, trained nurses or treatment books for self-administered CBT. Almost all selected studies recruited all types of sub-types of IBS patients except for one trial that recruited only patients with diarrhea-predominant IBS . Five studies reported ITT data at posttreatment [37,39,42,43,45]. The detailed characteristics of the individual studies are presented in Table 1.
Risk ofbias in included studies
Only two of the eighteen RCTs were performed using allocation concealments [42,43]. The other 16 studies were neither conducted using allocation concealments nor clearly reported the randomization methods.
Complete double-blinding could not be achieved in any ofthe studies. The blinding of patients was used in two studies, one of which used blinded ID number to send patients to the treatment group , and the other study avoided telling the patients about the experimental design . Three studies used the blinding of outcome evaluation [27,43,45]. The blinding of the operator cannot be conducted in a study about CBT because the specific clinical technique requires the operator to face the patient directly. Thus, the therapists were not blinded in these studies.
We also estimated several items, including adequate sample size, record of inclusion/ exclusion criteria, number and reasons for withdrawal and so on (Appendix B).
A total of 49 articles meeting the inclusion criteria were identified, but 31 studies were excluded. Five articles were notpublished in English [48-52]. Five citations only presented an abstract [53-57]. Ten studies [58-67] reported the second analysis of five other studies [39,42-45] that had already been included in our meta-analysis. Three studies developed derivative methods from RCT as nested studies [68-70]. Two studies were not RCTs [71, 72]. One study used another type of CBT as a control . Four studies did not provide extractable data for our meta-analysis [74-77]. One study was performed with patients with functional bowel disorders including IBS .
Efficacy of CBT in the treatment of IBS
Continuous post-intervention data were available for 16 studies with 19 circumstances comprising 1380 patients with IBS bowel symptoms. The remaining two trials of the included studies were ineligible, as one  did not provide extractable data, and one  had not included the IBS bowel symptom outcomes. As shown in Fig. 2, there was a medium to large significant pooled effect size of 0.678 (95% CI: 0.417, 0.939) in favor of CBT over all types of controls at the post-treatment evaluation. In the subgroup analyses, CBT outperformed waiting list and medical controls with large effects and outperformed basic support group with small effects. When CBT was compared with other psychological controls, the SMD indicated minimal and nonsignificant effects. Unfortunately, most ofthe follow-up studies had a cross-over design and could not be included in the meta-analysis, especially for the long-term follow-up. This resulted in eight trials with the data of nine circumstances at short-term follow-up and two studies with data of three circumstances at long-term follow-up remaining. The overall short-term SMD indicated medium effects 0.508 (see Fig. 3). In the subgroup analyses, the effect sizes were found in favor of CBT over waiting list and medical controls with large effects. When
Fig. 4. Effect size estimates for the efficacy of CBT compared to controls in improvement of IBS QOL.
NOTE: Weights are from random effects analysis. The four subgroups were classified by the controls. The specific treatments are presented above.
CBT was compared with other psychological controls, the SMD indicated minimal and non-significant effects. We merged all the trials with available long-term follow-up data. The SMD indicated non-significant effects. When we calculated the pooled effect size for the difference of means between pre-treatment and post-treatment of the studies that had long-term follow-up data [27,35,43,44], the SMD was 1.179 (95% CI: 0.963,1.396). The SMD between pre-treatment and long-term follow-up was 1.030 (95% CI: 0.812, 1.249). The SMD between post-treatment and long-term follow-up was – 0.078 (95% CI: –0.291,0.135), which was not significantly different. The data suggested that the efficacy of CBT for IBS may have been sustained after the treatment phase.
There were a total of seven studies comprising 558 patients that reported IBS QOL scores. One study fulfilling the inclusion criteria had to be excluded at this stage because it reported unusually high negative effect sizes for QOL, and the description of the outcomes in the article was quite different from other studies of approximately the same scale . This study was regarded as an outlier because of its distinctive outcomes. As shown in Fig. 4, there was a medium significant pooled effect size of 0.488 (95% CI: 0.237,0.740) in favor of CBT over all types of controls at the post-treatment. In the subgroup analyses, CBT outperformed waiting list controls with medium to large effects, and the SMD of CBT compared with basic support groups indicated nonsignificant effects. Only four studies comprising 425 patients provided short-term follow-up data. The pooled short-term SMD indicated small to medium effects (see Fig. 5). The long-term follow-up QOL data were available in two studies [27,35]. The pooled SMD of CBT compared with all types of controls indicated nonsignificant effects. The pooled SMD between pretreatment and post-treatment was 0.789 (95% CI: 0.471, 1.108), and the SMD between pre-treatment and long-term follow-up was 0.842 (95% CI: 0.486, 1.198). The SMD between post-treatment and long-term follow-up was 0.137 (95% CI: –0.199, 0.472), which was not significantly different. These results suggest that the efficacy of CBT for IBS QOL improvement may also have been sustained.
The data about psychological states (mainly depression and anxiety) were available for 12 studies with 14 circumstances. These studies comprised 910 patients. As shown in Fig. 6, there was a small to medium significant pooled effect size of – 0.213 (95%
CI: –0.423, –0.003) in favor of CBT overall types of controls atpost-treatment. In the subgroup analyses, CBT outperformed waiting list with large effects. When CBT was compared with all types of controls, the SMD indicated non-significant effects. At follow-up, there were a total remaining number of seven trials with eight circumstance data at short-term follow-up and three studies with four circumstance data at long-term follow-up. The overall short-term SMD for these 669 patients indicated non-significant effects (see Fig. 7), as did the subgroup analyses, and the pooled long-term SMD also indicated non-significant effects. The long-term follow-up data concerning psychological states were available in three studies [27,43,44]. The pooled SMD between pre-treatment and post-treatment of the three studies was – 0.303 (95% CI: –0.576, –0.031), and the SMD between pre-treatment and long-term follow-up was –0.171 (95% CI: –0.462, 0.120). The SMD between post-treatment and long-term follow-up was 0.125 (95% CI: –0.095,0.346), which was not significantly different. These results suggest that the efficacy of CBT for IBS may not have been sustained at long-term follow-up.
Subgroup analyses were also conducted according to the controls, whether they included the psychological treatment or not (see Table 2). The CBT effect appeared to be more significant for IBS when compared with non-psychological treatments, and the other psychological treatments appeared to be as effective as CBT.
No adverse events were reported in these studies.
This study conducted a meta-analysis to establish the efficacy of CBT for IBS and highlighted some important findings. First, the pooled effect sizes of CBT for IBS appeared to favor the use of CBT for the improvement of IBS bowel symptoms, QOL and psychological states at posttreatment evaluations. At follow-up, our study found that CBT prevailed over controls in improvementofIBS bowel symptomsandQOLatshort- term follow-ups, whereas the effect size displayed no significant difference in psychological states at the same time point. However, the test for heterogeneity of the overall included studies was significant. Thus, it is difficult to make definitive conclusions about all of the pooled average effect sizes because of high degrees of heterogeneity. Second, in the subgroup analyses, CBT was significantly more effective in improving IBS bowel symptoms, QOL and psychological states than waiting list controls at the end of the intervention and short-term follow-up. When compared with basic support controls and medical treatment, the effect sizes were in favor of CBT for the improvement of IBS bowel symptoms at post-treatment and short-term follow-up, but the effect sizes were not superior to the control groups in the improvement of QOL and psychological states. The effect sizes were almost non-significant between CBT and other psychological controls, but the effects of both CBT and other psychological controls appeared to be positive in each study. In the other subgroup analyses, the CBT appeared to be more superior to non-psychological treatment for IBS in the improvement of IBS bowel symptoms, QOL and psychological states at post-treatment and short-term follow-up. Third, when we calculated the pooled effect size of the studies that provided long-term follow-up data, the overall controlled effect sizes for CBT indicated effectiveness at post-treatment and long-term follow-up with large effects for improvements of IBS symptoms and QOL, which suggested that the efficacy of CBT for IBS may persist after the treatment phase.
Fig. 5. Effect size estimates for the efficacy of CBT compared to controls in improvement of IBS QOL at short-term follow-up.
To our knowledge, there are several systematic reviews that have examined the efficacy of CBT for IBS [13-17]. Our findings are consistent with the previous reviews, demonstrating that CBT may be highly effective in improving IBS bowel symptoms when compared with waiting lists and standard basic support groups [13-15,17]. However, the conclusions of the narrative reviews may be less convincing than meta-analyses as their conclusions were not based on the calculation of overall treatment effects but on the opinions of the authors. Importantly, our study focused on the CBT for IBS and included mostly recent studies. On the other hand, our meta-analysis developed subgroup analyses that included almost all of the possible controls for IBS. Thus, we detected the efficacy of CBT for IBS compared with waiting lists, basic support controls, only standard medical treatment and other psychological therapies. Our study also found that the efficacy of CBT for IBS might be sustained at long-term follow-up.
There are some limitations in our study. One problem is the diversity of CBT formats, which resulted in considerable heterogeneity in our meta-analysis. The operators of the CBT and the duration of the CBT were also different in the studies. As there exists no standardized treatment procedure for CBT, our meta-analysis of CBT for IBS had obvious heterogeneity. We included almost all available control arms and detected that almost none of the controls were the same except for the waiting list controls. Thus, performing subgroup analyses according to the control treatments and categorizing control treatments would be difficult. On the basis of the details reported in the articles, the controls were classified into waiting list controls, basic support controls, medicine alone and other psychological treatments. However, this classification has not been confirmed, and heterogeneity could still be detected in the subgroup analyses. Inevitably, all of the available studies were performed without double-blinding, which may have led to potential placebo effects and observation bias. We also noted that the present study was limited to materials published in English-language journals, which may have led to publication bias.
Fig. 6. Effect size estimates for the efficacy of CBT compared to controls in improvement of psychological anxiety and depression at post-treatment.
It has been proven that CBT is superior to waiting list and basic support controls in the treatment of IBS in previous reviews, and our work provides further confirmation for this conclusion. In this meta-analysis, we included almost all types of the control groups and developed the subgroup analyses according to the controls. We found that CBT had a significant advantage in all aspects when compared with a control group of non-psychotherapy, and CBT appeared to maintain the same superiority when compared to other types of psychotherapy. These results imply that the psychological treatment was, at the minimum, a good choice to treat IBS. In other subgroup analyses, we found that the combination of CBT with standard medical treatment would increase the efficacy of using drug therapy alone. We also determined that the efficacy of CBT for IBS might continue at long-term follow-up. However, the evidence is somewhat limited by the variations of CBT procedure, the relatively low quality of the included studies and the small quantity of the valid studies. More high-quality studies are needed, and the CBT should also be standardized.
In conclusion, CBT was superior to waiting lists, basic support controls or medical treatment at post-treatment and short-term follow-up in the treatment of IBS bowel symptoms but was not superior to other psychological treatments. In terms of QOL and psychological states, CBT for IBS was proven to be more effective than waiting list at the end of treatment. The efficiency of CBT may persist after the treatment phase. Our meta-analysis might be limited by the heterogeneities and small sample sizes of the included studies. Nevertheless, the superiority of CBT for the treatment of IBS should not be ignored.
Guarantor of the article: Lishou Xiong
Author contributions: LSX contributed to the study concept, design, supervision and takes responsibility for the integrity of the data and the accuracy of the data analysis. All authors approved the final version of the manuscript. LL and SHZ both contributed to the trial selection and data extraction. LL, SHZ and QY performed the analysis and interpretation of data, and LL wrote the article. MHC contributed to the supervision of the manuscript.
This study was funded by the Planned Science and Technology Project of Guangdong Province (2009B030801138).
Conflict of interest
The authors have no conflict of interests to report. Acknowledgments
Declaration of personal interests: We are grateful to Professor Roger Jones and Tom Kennedy for answering our data queries and, where applicable, providing us with their original datasets for analysis.
Fig. 7. Effect size estimates for the efficacy of CBT compared to controls in improvement of psychological anxiety and depression at short-term follow-up.
Effect sizes and the heterogeneity for the outcomes of CBT for IBS in subgroup analyses of psychological treatments
|Outcomes||Subgroup||Post-treatment||Short-term follow-up||Long-term follow-up|
|Effect size||Heterogeneity||Effect size||Heterogeneity||Effect size|
|Hedges’ g||95% CI||i2(%)||P||Hedges’ g||95% CI||i2(%)||P||Hedges’ g||95% CI||i2(%)||P|
|IBS symptom||NP||0.875||0.568,1.183*||73.5||<.001||0.597||0.330, 0.864*||53.4||.036||0.058||-0.272, 0.388||48.5||.144|
|P||0.133||-0.194,0.461||70.7||.017||0.241||-0.211,0.693||62.9||.101||— 0.255||-0.723, 0.212||/||/|
|Psychological states||NP||-0.297||-0.533, -0.060*||48.1||.031||— 0.258||-0.45, -0.067*||0.6||.412||— 0.114||-0.518, 0.29||68.2||.043|
|P||0.068||-0.192, 0.327||8.8||.295||0.164||-0.504,0.833||82.4||.017||0.395||-0.075, 0.865||/||/|
Note: * indicates that the effect sizes have significant meaning. NP: the controls were not psychological treatments.
P: the controls were other psychological treatments.
HYPNOSIS HOME TREATMENT FOR IRRITABLE BOWEL SYNDROME: A Pilot Study
Abstract: Hypnosis treatment often improves irritable bowel syndrome (IBS), but the costs and reliance on specialized therapists limit its availability. A 3-month home-treatment version of a scripted hypnosis protocol previously shown to improve all central IBS symptoms was completed by 19 IBS patients. Outcomes were compared to those of 57 matched IBS patients from a separate study receiving only standard medical care. Ten of the hypnosis subjects (53%) responded to treatment by 3-month follow-up (response defined as more than 50% reduction in IBS severity) vs. 15 (26%) of controls. Hypnosis subjects improved more in quality of life scores compared to controls. Anxiety predicted poor treatment response. Hypnosis responders remained improved at 6-month follow-up. Although response rate was lower than previously observed in therapist-delivered treatment, hypnosis home treatment may double the proportion of IBS patients improving significantly across 6 months.
Irritable bowel syndrome (IBS) is a common disorder characterized by abdominal pain and altered bowel function, which is ineffectively treated by conventional medical methods. Less than half of patients with severe symptoms report satisfactory relief 6 months after seeing doctors for their bowel condition (Whitehead, Levy, et al., 2004). The disorder is associated with substantial disability and quality-of-life impairment (Chang, 2004) and increases the average healthcare costs of individuals by 49% and nearly doubles their number of healthcare visits (Levy et al., 2001; Levy, Whitehead, Von Korff, & Feld, 2000).
A number of psychological treatments have been tested for the purpose of improving on the limited success of medical treatments for IBS. Among psychological therapy modalities tested for this purpose, cognitive-behavioral therapy and hypnosis currently have the best research evidence of success. Both of these treatments have been found in a number of studies, including randomized controlled trials (Drossman et al., 2003; Palsson, Turner, Johnson, Burnett, & Whitehead, 2002; Payne & Blanchard, 1995; Whorwell, Prior, & Faragher, 1984), to substantially benefit the majority of treated IBS patients.
Even though most published studies on cognitive-behavioral therapy and all the studies on hypnosis treatment have reported a significant therapeutic impact on IBS symptoms and these benefits have been demonstrated to last for years after treatment is completed (Gonsalkorale, Miller, Afzal, & Whorwell, 2003; Whorwell, Prior, & Colgan, 1987; Van Dulmen, Fennis, & Bleijenberg, 1996), wide implementation of these psychological treatments as adjunctive interventions for IBS faces significant obstacles. One problem is the added cost of such services. Successful psychological interventions for IBS generally require a course of clinical treatments that ranges from six to twelve or more sessions. The expense of such multiple treatment visits and the limited insurance reimbursement for psychological treatments of IBS keep these therapies unavailable to many patients.
Another barrier to widespread use of psychological treatments for IBS is the lack of availability of clinicians proficient in treating the disorder with these methods. Even though thousands of clinicians offer therapy with cognitive-behavioral therapy and clinical hypnosis, most of them do not have training and knowledge in the special gut-focused applications of hypnosis or cognitive-behavioral therapy for IBS. Further, clinicians using these techniques are unavailable in many places outside major urban population centers.
Our research team at the University of North Carolina (UNC) in Chapel Hill developed and tested a seven-session scripted protocol for the treatment of IBS and found it to be effective for improving IBS in two studies (Palsson, Turner, et al., 2002). Abdominal pain, bloating, and stool-form abnormality (hard or watery stools) all improved substantially after hypnosis treatment with this protocol. Most patients (87.5% and 94%) responded to this standardized treatment in the two studies, and improvement was well maintained at 10-month follow-up.
The availability of a written protocol that could benefit most patients when delivered verbatim opened the possibility of home treatment via audio recordings. We felt that if it were possible to deliver this treatment in an automated format, without the need for therapist contact, it would be an appealing option to reduce the cost and make hypnosis available as an adjunctive treatment to the many IBS patients who do not have access to such treatment with a clinician.
Hypnosis treatment administered via audio recordings has previously been reported to be effective for controlling a number of medical problems, including tinnitus (Brattberg, 1983), blood loss during oral surgery (Enqvist, Von Konow, & Bystedt, 1995), and postoperative nausea and vomiting (Enqvist, Bjorklund, Engman, & Jakobsson, 1997). Only one prior published study (Forbes, MacAuley, & Chiotakakou-Faliakou, 2000) has tested audio recordings as IBS therapy. The investigators found that IBS symptom scores improved in 59% of the patients who used the tape recordings at home. However, the home intervention in that study was not hypnosis but rather a mix of patient education, stress-reduction training, hypnosis-type suggestions, and coping-skills training. The present study is therefore the first study of home hypnosis treatment for IBS.
The primary aim of the study was to assess in a preliminary fashion whether hypnosis home treatment, delivered without contact with a therapist or added care by healthcare providers, improves IBS symptom severity and health-related quality of life above what may be expected from standard medical treatment. Secondary aims were to assess predictors of treatment response and the maintenance of the therapeutic effect after treatment ended.
Hypnosis participants were recruited through announcements on the Web site and in the newsletter of the UNC Center for Functional Gastrointestinal and Motility Disorders. To qualify for participation, subjects had to have been diagnosed with IBS by a physician and meet Rome II criteria (Drossman, Corazziari, Talley, Thompson, & Whitehead, 2000) for IBS. The primary physicians of enrolled patients were contacted, with the patients’ written permission, to confirm the patients’ clinical diagnosis and enrollment criteria.
To ensure that participants were in an active phase of their disorder, only subjects who reported at least one weekly episode of abdominal pain in the past 4 weeks were enrolled in the study. To minimize the risk of complications arising from unsupervised home practice of hypnosis, patients who had ever been diagnosed or treated for dissociative disorders, posttraumatic stress disorder, or borderline personality disorder or had a history of any other psychiatric disorder that included psychotic features were excluded from the study. Participating subjects could also not be on psychotropic medications (i.e., antidepressant, antipsychotic, or antianxiety medications) or be currently receiving psychological treatment for their IBS symptoms.
A total of 25 subjects were enrolled for hypnosis treatment. Of these, 19 subjects (76%) completed the study treatment protocol and evaluations, and the data from these subjects is reflected in the analyses below. One subject was disqualified during treatment due to false reporting of exclusion criteria, and two dropped out during the treatment phase due to difficulty with maintaining regular home treatment. Three subjects signed up for the study but terminated contact with the investigators prior to beginning treatment.
To obtain an estimate of the advantage of hypnosis treatment over what could be expected from standard medical care alone, the hypnosis recipients were compared to matched IBS patients not in the study. The control subjects were 57 systematically selected Rome II IBS patients from a separate observational study of standard medical care in a large health maintenance organization, described elsewhere (Whitehead, Levy, et al., 2004), who had participated in a questionnaire study of clinical outcomes from standard medical care (73% completion rate). The control patients were matched three to one to hypnosis subjects who completed participation. The triple match for each hypnosis subject was done to minimize the bias from chance individual variability in the control group due to the small size of the hypnosis group. Control subjects were matched to hypnosis subjects on (a) gender, (b) age (+/- 1 year), (c) IBS severity at baseline (to the closest degree possible), and (d) race. As Table 1 shows, excellent match was achieved between the groups on all of these variables.
Comparison of Subject Characteristics in the Two Groups at Baseline
|Variables||Hypnosis Patients (N = 19)||Group
Control Patients (N = 57)
|Agea||43.7 (16.6)||43.5 (16.3)||n.s.|
|Gender (% Female)3||78.9%||78.9%||n.s.|
|Race (% White)a||94.7%||94.7%||n.s.|
|IBS Severity Scorea||288.4 (73.8)||288.5 (104.5)||n.s.|
|IBS-QOL Score||50.3(21.0)||65.3 (21.66)||.011|
|BSI-Anxiety||54.9 (12.5)||50.2 (9.5)||n.s.|
|BSI-Depression||56.4 (13.2)||53.3 (10.3)||n.s.|
|BSI-Somatization||54.8 (8.8)||56.9 (8.7)||n.s.|
|BSI-General Severity Index||56.7(11.26)||54.4 (9.1)||n.s.|
|Number of Comorbid Medical Conditions||1.6 (1.9)||2.6 (2.4)||n.s.|
|Completion rate (% of patients completing both baseline and posttreatment outcome data)||76%||73%||n.s.|
“Control subjects were systematically selected from the HMO study pool of IBS patients to match hypnosis subjects on these variables
The IBS Severity Scale (IBSS). The IBSS (Francis, Morris, & Whorwell, 1997) consists of five questions covering the frequency and severity of abdominal pain, severity of abdominal distention, dissatisfaction with bowel functioning, and life interference from IBS symptoms. These individual responses can be analyzed separately or added together for an overall index of IBS severity. The scale has been shown to discriminate patients from controls and discriminate between patients with mild, moderate, and severe symptoms as classified by clinical assessment. Test-retest reliability was judged to be excellent (Francis et al., 1997), but no correlation coefficient was given. This scale has been found to be sensitive to treatment change from hypnosis in several prior IBS studies (Gonsalkorale, Houghton, & Whorwell, 2002; Gonsalkorale et al., 2003).
IBS-QOL. This is a 34-item disease-specific, quality-of-life measure for IBS, which has high internal consistency (Cronbach’s alpha = .95) and high reproducibility over a 7-day interval (ICC = .86) (Patrick, Drossman, Frederick, Dicesare, & Puder, 1998). Drossman, Patrick, et al. (2000) assessed the responsiveness of the IBS-QOL to treatment in 156 female patients with functional bowel disorders and showed that changes in the IBS-QOL are modestly but significantly correlated with changes in pain reports on a 14-day diary (r = .25) and with changes in total scores on the Sickness Impact Profile (r = .28). The IBS-QOL was found to differentiate responders from nonresponders in a trial of cognitive-behavioral treatment and antidepressant trial for IBS. Thus, the scale is reliable, valid, and responsive to treatment.
Brief Symptom Inventory (BSI) 18. The 18-question BSI (Derogatis, 2000) is a brief version of the Symptom-Checklist-90-Revised (SCL-90-R). Respondents rate how much each symptom has bothered them in the past week on a 4-point ordinal scale. The inventory provides scores on three symptom scales for anxiety, depression, and somatization, as well as a general severity index score. Internal consistency (Cronbach’s alphas) for the scales ranges from .74 to .89, and the correlations with corresponding scales on the SCL-90-R are .91 to .96.
Tellegen Absorption Scale (TAS). The TAS (Tellegen & Atkinson, 1974) contains 34 true/false questions and is designed to measure a person’s capacity for absorption and imaginative involvement in his or her life experiences. It has internal reliability of .88 and a test- retest reliability of between .85 and .91 (Tellegen, 1985; Tellegen & Alkinson, 1974). The TAS has good construct validity and is normally distributed in the general population. This scale was included because it has been found to partly predict hypnotizability (Dixon, Labelle, & Laurence, 1996), which may influence responsiveness to hypnosis treatment.
Comorbid Medical Conditions Questionnaire (CMCQ). This is an IBS- specific questionnaire that asks the respondent to report whether he or she has been diagnosed with each of 16 comorbid medical conditions. This questionnaire was developed and validated in our laboratory. It has an internal consistency of .67, test-retest reliability of .95, and a correlation of .64 with the Cornell Medical Index, another medical-history questionnaire (Palsson, Jones, Turner, Drossman, & Whitehead, 2002). The CMCQ was included in this study because IBS patients with comorbid medical conditions have sometimes been found in past work to respond more poorly than other patients to treatment.
Assessment. Hypnosis subjects completed the IBSS, CMCQ, BSI, IBS- QOL, and TAS questionnaires at baseline, prior to beginning their hypnosis home treatment. During the 12-week home-treatment period, they completed weekly online logs on a password-protected web page of their home practice and their symptoms (these diary symptom ratings were secondary measures for exploratory purposes and are not presented in this paper). Hypnosis subjects also completed the IBSS, IBS-QOL, and BSI at the end of the treatment period as well as 3 and 6 months after the end of treatment.
The control subjects completed the same questionnaires as the hypnosis subjects at baseline, with the exception the TAS. They completed the IBS-QOL, IBSS, and BSI again 6 months after baseline measurement (at a time point equivalent to the 3-month posttreatment followup for the hypnosis group).
Treatment. The scripted seven-session hypnosis treatment protocol used in our two prior studies (Palsson, Turner, et al., 2002) was recorded verbatim by the principal author (OSP) and duplicated on a set of five audio compact disks (CDs). The recorded protocol consisted of seven different biweekly sessions, each of which was about half an hour in length, and a shorter (13 minute) hypnosis audio exercise for daily use. The total treatment period of this structured home-treatment sequence was 12 weeks. The session frequency and total treatment duration was the same as in our previous studies. An instruction booklet that explained the nature of hypnosis and the use of the CDs was created and provided to subjects with the CD set. Hypnosis subjects were instructed to initiate the 12-week home treatment immediately following their completion of the baseline questionnaires. Each subject’s compliance with the audio sessions was tracked via their weekly online log entries. The hypnosis subjects were advised to continue medical care and treatments for their IBS provided by their physicians while participating in the study.
Control subjects received only the standard medical care provided by their physicians during the 6-month comparison interval. No control subject received psychological treatment during the time period reflected in the collected data.
All assessments of hypnosis subjects were completed through mail and the Internet to minimize contact with personnel that might have a therapeutic effect independent of the home treatment under investigation. The only direct contacts with investigators were e-mail communications with the study coordinator (and in the case of a couple of subjects, phone calls) to resolve practical matters regarding study participation and data collection. Control subjects completed all assessments (baseline and 6-month follow-up questionnaires) through mail.
The institutional review board at the University of North Carolina at Chapel Hill approved the study prior to subject enrollment.
Data analysis. Fisher’s exact tests were used to evaluate group differences in categorical data, and analyses of variance (ANOVA) tests were used to test group differences in continuous variables. Because the two patient groups were significantly different in IBS-QOL scores at baseline, analysis of covariance (ANCOVA) with the subjects’ baseline values as covariate variable was used to assess treatment changes in IBS-QOL scores.
Based on our prior analysis of different methods to measure therapeutic effects in IBS research (Whitehead, Palsson, et al., 2004), we selected a stringent criterion of a 50% or greater reduction in IBS symptom severity as the most appropriate definition of a treatment responder. Prior data show that, overall, 22% of clinical IBS patients receiving medical care can be expected to be treatment responders by this definition (Whitehead et al., 2004) at 6-month follow-up.
The primary time point selected for assessing treatment response was 3 months after the end of hypnosis treatment. This was done because it enabled us to compare outcomes directly to those of the matched control group after an equivalent time interval from baseline. However, scores on the main outcome measures were also obtained from the hypnosis group immediately after the end of treatment and at 6-month follow-up.
To test the effects of clinically significant psychological distress on treatment response, we first converted BSI scores to T-scores (scores reflecting standard deviation distance from the gender group mean) to adjust for gender differences. As suggested in the test manual (Derogatis, 2000), we classified patients with BSI T-scores of 63 or greater as clinically significantly distressed.
Comparison of Group Characteristics
Due to systematic matching, subjects in the hypnosis and control groups were nearly identical in regard to age, gender ratio, racial composition, and mean IBS severity at baseline. There were no significant group differences between the groups in psychological symptoms or comorbid medical conditions at baseline. However, baseline IBS quality- of-life scores were significantly lower in the hypnosis group compared to the control group (mean + /- standard deviation: 50.3 + /-21.3 vs. 65.3 + /- 21.6; p = .01). This difference was corrected for by using baseline scores of the subjects as covariates in the IBS-QOL outcome analysis. The comparisons of baseline group characteristics are summarized in Table 1.
The main treatment outcomes for the two groups are presented in Figure 1. Fifty-three percent (10/19) of the subjects who completed hypnosis home treatment responded to treatment by 3-month follow-up, i.e., had >50% reduction in IBS symptom severity on the IBSS. In contrast, only 26% (15/57) of controls met the same criterion for treatment responder. This group difference was statistically significant (Fischer’s exact test p < .05).
Figure 1. Changes in health-related quality of life and treatment response rates over 6 months in the hypnosis and control groups.
According to published score ranges for the IBSS (Francis et al., 1997), 6 of the 10 hypnosis treatment responders scored in the severe range (IBSS > 300) before treatment, and the remaining 4 responders were in the moderate range (IBSS > 175 and < 300). All 10 of these subjects converted to mild IBS severity scores (< 175) by 3-month follow-up.
ANCOVA results with baseline scores as covariates to control for baseline group differences showed that hypnosis subjects improved more in their IBS-QOL health-related quality of life scores compared to controls, F(1, 78) = 14.08, p = <.0001).
Maintenance of Therapeutic Effects
As shown in Figure 2, the improvements in IBS symptom severity and quality-of-life scores seen in hypnosis treatment responders at the end of treatment were fully preserved at 3- and 6-month follow-up. The nonresponders showed similar stability of their posttreatment severity levels.
Figure 2. Comparison of IBS symptom severity and quality of life at the four assessment time points for hypnosis responders vs. nonresponders.
Post-hoc Analyses of Changes in Psychological Symptoms After Hypnosis Treatment
All psychological symptom scores on the BSI were unchanged at the 3-month follow-up evaluation point compared to pretreatment scores for the hypnosis subject sample as a whole. Separate analyses of the BSI scores of the nonresponder and responder subgroups showed no changes in nonresponders. In the treatment responder group, somatization scores were significantly reduced after treatment (T-score mean + /- standard deviation: 47.1 + /- 7.7 vs. 52.4 + /- 7.6; p = .03), but all other BSI variables were unchanged.
Assessment of Predictors of Treatment Response
Baseline scores on the TAS, CMCQ, IBS-QOL, IBSS, and scores for all the scales of the BSI 18 were compared for hypnosis responders and nonresponders in an effort to identify predictors of treatment response (see Table 2). Of the tested variables, only the BSI anxiety scores alone were significantly different in these subgroups. Only 14% (1/7) of patients with clinically significant anxiety met criteria for treatment responders after hypnosis treatment, whereas 75% (9/12) of nonanxious subjects were responders.
Comparison of Predictor Variables in Responders vs. Nonresponders within the Hypnosis Group
Responders Nonresponders (n = 10) (n = 9)
|Age||46.0 (15.2)||41.1 (18.5)||n.s.|
|Years with IBS||7.6 (8.8)||13.6 (13.3)||n.s.|
|IBS Severity Score at baseline||298.5 (54.6)||277.2 (92.9)||n.s.|
|IBS-QOL Score at baseline||57.4(18.9)||42.6 (21.5)||n.s.|
|(BSI t-score > 63)|
|(t-score > 63)|
|(t-score > 63)|
|Significant General Severity||10%||33%||n.s.|
|(t-score > 63)|
|Number of Comorbid Medical||2.0 (1.0)||1.2 (0.6)||n.s.|
|Tellegen Absorption||20.5 (9.1)||20.0 (6.3)||n.s.|
Adherence to Home Treatment and Adverse Effects
Compliance with the home hypnosis treatment was good, based on the diary results. On the average, subjects completed a home practice session 5.3 times a week. There was no difference between the frequency of practice of responders and nonresponders. No adverse effects of the hypnosis treatment or practical difficulties with the audio treatment program were reported by the patients.
The findings of this study provide preliminary data that suggest that hypnosis home treatment by means of audio recordings may substantially increase the probability of favorable symptom change in IBS patients. Our data indicate that home hypnosis with the protocol we tested might double the proportion of IBS patients experiencing significant improvement in their overall symptoms across 6 months and that the therapeutic benefits in treatment responders remain undiminished beyond 6 months after treatment termination. Our results further indicate that this treatment is well received and free from practical obstacles, as no adverse effects or pragmatic difficulties were reported, and compliance with home practice was good.
It must be noted, however, that the treatment response rate observed in this study was substantially lower than seen in our previous studies of therapist-delivered treatment with the same scripts. Although the present study used a more stringent criterion to define treatment responders than our previous work, our examination of the outcome data for the individual nonresponders in the present study shows that all of them would also have been classified as nonresponders in our previous studies (Palsson, Turner, et al., 2002)
The study findings suggest that the home treatment mode of delivery may not benefit patients who have significant levels of anxiety, as indicated by the dramatic difference of 14% versus 75% response rate of anxious versus nonanxious patients. It is also likely that other measures of psychological distress will be found to attenuate the response to treatment in a larger study of this home treatment, judging from the numerically higher percentages of other psychological distress variables in the nonresponders in Table 2, even though those differences did not reach statistical significance in this comparison of small groups. Also, some patients may not have the time or inclination to complete this intensive 3-month home treatment course on their own (as represented by 2 individuals dropping out of our initial test sample of 25 for that reason).
It seems likely, if the lower response rate to home-based hypnosis treatment compared to face-to-face treatment persists in future tests, that in-person treatment by a therapist will continue to be the gold standard of psychological treatment for IBS. Home treatment may nonetheless be an appealing option for enhancing rates of clinical improvement for patients who cannot access such care by a therapist and who do not have psychological distress. However, future studies with a more definitive design and larger samples will be needed to confirm the therapeutic value of the home treatment version of our IBS hypnosis protocol.
This first test of hypnosis home treatment for IBS was only designed to provide preliminary answers to the study questions, and as a pilot study it had significant limitations. The main limitations pertain to the fact that the selection of subjects into groups was not randomized and that the patients in the control group were not participating in the same kind of study as the hypnosis patients. Because the subjects who enrolled in the hypnosis group knew that their study participation entailed treatment with hypnosis, it is possible that self-selection bias of people open to participating in hypnosis treatment affected the outcome. However, the same kind of self-selection would also determine which medical patients would accept and use a hypnosis home treatment package when offered by their healthcare providers, so this bias is unlikely to affect the conclusions that can be drawn from our results about the potential benefits of this intervention method.
The study did not control for this effect, favoring better outcomes for the hypnosis group, which may have come simply from enhanced positive expectancy of benefit due to receiving a credible intervention in additions to standard medical care alone. However, that same advantage would again apply if patients were offered this intervention as an adjunctive treatment in medical settings.
Although the two groups of patients compared in this pilot study were from different samples, the validity of the results is strengthened by the high degree of comparability of subjects in the two groups on key characteristics, due to the availability of a very large sample of control patients from which to select patients for matching. The fact that the groups were practically identical in gender, age, IBS severity, and race composition reinforces our confidence in the results of the outcome comparisons.
We conclude based on the promising results of this trial that the home treatment version of our standardized IBS treatment protocol warrants further study, especially in light of its very low cost of delivery, absence of any added burden on healthcare providers, and easy application to large numbers of patients.
HYPNOSIS FOR IRRITABLE BOWEL SYNDROME: The Empirical Evidence of Therapeutic Effects
William E. Whitehead1, 2
University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA
Abstract: Irritable bowel syndrome (IBS) is a complex and prevalent functional gastrointestinal disorder that is treated with limited effectiveness by standard medical care. Hypnosis treatment is, along with cognitive-behavioral therapy, the psychological therapy best researched as an intervention for IBS. Eleven studies, including 5 controlled studies, have assessed the therapeutic effects of hypnosis for IBS. Although this literature has significant limitations, such as small sample sizes and lack of parallel comparisons with other treatments, this body of research consistently shows hypnosis to have a substantial therapeutic impact on IBS, even for patients unresponsive to standard medical interventions. The median response rate to hypnosis treatment is 87%, bowel symptoms can generally be expected to improve by about half, psychological symptoms and life functioning improve after treatment, and therapeutic gains are well maintained for most patients for years after the end of treatment.
Irritable bowel syndrome (IBS) is a prevalent and complicated functional gastrointestinal disorder with poorly understood etiology, which has proven difficult to treat for both primary care medicine and gastroenterology. Many IBS patients experience little or no relief in their bowel symptoms from conventional medical interventions. For example, we recently found in a large HMO survey (Whitehead et al., 2004) that only 49% of patients who visited primary care doctors for IBS reported that their symptoms were at least somewhat better 6 months later.
Because of this limited effectiveness of standard medical care, many alternative or complementary treatments for IBS have been tested. These have included several psychological therapy modalities: biofeedback, relaxation training, brief psychodynamic therapy, behavioral and cognitive interventions, and various combinations of these treatments. Although all of these psychological treatment modalities
have been reported to have some impact on IBS in published work, some have shown indifferent outcomes and few have been extensively tested. Hypnosis and cognitive-behavioral therapy are the two psychological treatments that have been most investigated for IBS and show promise for making a substantial difference in the disorder.
The first study reporting successful application of hypnosis for IBS treatment was published in 1984 (Whorwell, Prior, & Faragher, 1984). Since that time, the empirical literature on the use of hypnosis in IBS has grown to a body of 15 publications, but only 11 of these (Forbes, MacAuley, & Chiotakakou-Faliakou, 2000; Galovski & Blanchard, 1998; Gonsalkorale, Houghton, & Whorwell, 2002; Gonsalkorale, Miller, Afzal, & Whorwell, 2003; Gonsalkorale, Toner, & Whorwell, 2004; Harvey, Hinton, Gunary, & Barry, 1989; Houghton, Heyman, & Whorwell, 1996; Lea et al., 2003; Palsson, Turner, Johnson, Burnett, & Whitehead, 2002; Prior, Colgan, & Whorwell, 1990; Simren, Ringstrom, Bjornsson, & Abrahamsson, 2004; Vidakovic-Vukic, 1999; Whorwell, Houghton, Taylor, & Maxton, 1992; Whorwell, Prior, & Colgan, 1987; Whorwell etal., 1984) focused on therapeutic effects–the others are studies of mechanism (Lea et al., 2003; Prior et al., 1990; Simren et al., 2004; Whorwell et al., 1992). Without exception, the 11 hypnosis studies that have reported on therapeutic outcome in IBS, summarized in Table 1, have reported hypnosis to be effective in ameliorating the bowel symptoms of large proportions of the treated patient samples. However, the research methodology and the quality of the work have varied greatly. In this article, I will briefly summarize the evidence for the effectiveness of hypnosis treatment that these studies provide and discuss what can be concluded from this body of work about the benefits that hypnosis can be expected to afford IBS patients. I will also highlight some of the current limitations and future needs in this research domain.
Summary of the Evidence
It is a generally accepted minimum standard for research assessing the effectiveness of any medical therapy to use one or more control groups and randomly assign subjects to the groups to be compared. This is important, because such a design provides controls for many kinds of potential confounding influences on outcome, such as patient expectancy of improvement, spontaneous changes in the symptom course over time, and some types of experimenter bias. To date, four studies on hypnosis for IBS have been published that satisfied these two basic design criteria, and one additional study was a controlled comparison without randomization.
Studies Reporting Therapeutic Benefits of Hypnosis Treatment for IBS
|Authors and Year||Beneficial Effects:||Hypnosis n||Hypnosis Success Rate||Control Group||Outcome|
|A. Randomized Controlled Treatment Studies|
|Whorwell et al., 1984||A, В||15||100%||Psychotherapy||Hyp. Superior|
|& Placebo Pills|
|Galovski & Blanchard, 1998||A, В||11||Waitlist||Hyp. Superior|
|Forbes et al., 2000||В||25||76%||Audiotape|
|Palsson et al., 2002 (Study II)||A, В, S||24||87%||Waiting list controls||Hyp. Superior|
|B. Nonrandomized Controlled Treatment Studies|
|Houghton et al., 1996||В, D, H, Q||25||Standard Medical Care||Hyp. Superior|
|C. Uncontrolled Treatment Studies|
|Harvey et al., 1989||В||33||61%|
|Palsson et al., 2002 (Study I)||A, B, S||18||93%|
|Whorwell et al., 1987||В||50*||95% for classic IBS 60%|
|for cases with comorbid|
|psychopathology 43% for atypical IBS|
|Gonsalkorale et al., 2002, 2003||А, В, H, Q, S||250||71%|
|Gonsalkorale et al., 2004||А, В, C, Q, S||78||Not reported|
A = Affective symptoms, В = Bowel Symptoms, C = Cognitions, D = Disability, H = Health Care Utilization, M = Medication use, S = somatization/extracolonic symptoms, Q = Quality of Life.
* included the 15 patients reported in Whorwell et al. (1984).
Two of the randomized and controlled studies compared hypnosis treatment to another therapy that was likely to carry expectancy of therapeutic benefit. The first of these was the study by Whorwell and colleagues published in the Lancet in 1984 (Whorwell et al., 1984), which also was the first study in this research domain. The investigators randomly assigned 30 patients who had severe IBS that was refractory to standard medical care to either seven sessions of hypnosis treatment, delivered over a 12-week period, or to a control group that received the same amount of psychotherapy plus placebo pills (thus receiving a double placebo, as the psychotherapy was presumed to have little effectiveness). The hypnosis group showed very substantial improvement in all the central IBS symptoms after treatment and was significantly more improved on all outcome variables than the control group, which only showed minimal improvement (slight decrease in abdominal pain and bloating). Remarkably, every single IBS patient treated with hypnosis in this study improved, and a later paper (Whorwell et al., 1987) reported that all the patients remained better at 18-month follow-up, although a couple of them had needed further hypnosis sessions to counter relapses. Even though this study was small, it was a landmark trial in this research domain both because it was the first study demonstrating the potential for hypnosis as a therapeutic modality for IBS and because it is to date the only study to use a placebo control.
Forbes and colleagues (2000) randomly assigned 56 patients to either 6 sessions of gut-directed hypnotherapy (25 patients) or an audiotape home intervention (27 patients) that contained elements of patient education, stress reduction training, hypnosis-type suggestions, and coping skills training. The gastrointestinal symptoms of the hypnosis group improved significantly more (p < .05) compared to the audiotape group. The responder rate among the hypnosis subjects was also numerically higher (76% vs. 59%), although that difference did not achieve statistical significance. Although psychological and quality-of-life scores were also measured in this trial, the investigators were unable to obtain posttreatment responses for nearly half of the sample, and the findings on these parameters (although reported as nonsignificant trends toward improved psychological well-being and quality of life) are therefore not interpretable.
The other two trials that were both randomized and controlled used symptom monitoring, i.e., patients who were waiting for the test treatment, as a control condition. This is a markedly weaker design than parallel comparisons with an alternative treatment because it does not control for patient expectancy of therapeutic gain and may in fact create negative expectancy (that is, patients may expect not to get any better while waiting for treatment). However, the waiting-control design does safeguard against spontaneous changes in symptoms over time. This may be useful in studies of disorders such as IBS that have a fluctuating-symptoms course and especially in studies on medical patients seeking clinical care. Patients with chronic disorders are likely to seek treatment (including research treatments) during peaks in symptom severity and may thus be expected to show some spontaneous improvement in the subsequent months, which might create a false impression of therapeutic effect from clinical interventions.
Palsson et al. (2002) randomized 24 patients to either a waiting list or immediate hypnosis treatment. The hypnosis patients were treated with a sequence of seven hypnosis sessions over a 3-month period, using a standardized protocol where the entire treatment course followed the same verbatim scripts for all patients. Significant reductions in all bowel symptoms were seen in the hypnosis group after treatment, but there were no improvements in the control group after an equal period of time. Once the waiting group had completed treatment, the combined response rate for the whole sample was 87% (21/24), as determined by a combination of subjective rating of improvement and reduction in bowel symptoms on symptom diaries. Abdominal pain and abnormal bowel functioning (number of days with hard or watery stools) were reduced by about half on symptom diaries after treatment, and abdominal bloating showed a smaller but significant improvement. Anxiety and somatization symptoms were also markedly reduced. All treatment responders remained improved at followup 10 months after treatment.
Galovski and Blanchard (1998) likewise randomized six pairs of IBS patients to immediate hypnosis treatment or a waiting control condition, but 1 patient was removed from analysis, resulting in data on 11 patients being reported. Patients who received hypnosis treatment were significantly more improved on a composite bowel-symptom scale after treatment compared to waiting control subjects, and all bowel symptoms were found to be improved significantly after treatment when the pre- and posttreatment values for the whole sample were analyzed.
Houghton and colleagues (1996) conducted a controlled comparison of 1-year follow-up outcomes for 25 IBS patients who completed a course of hypnotherapy in addition to receiving standard medical care versus 25 patients with equivalent bowel-symptom severity who received only standard medical care. In addition to greater improvement in their physical symptoms, the individuals in the hypnosis group showed greater reduction in extra-colonic symptoms (general physical symptoms outside the gastrointestinal tract) and greater gain in quality-of-life scores, took significantly less time off work after treatment if they were employed and were more likely to return to work if they had been unable to work (of 4 patients who were not working in the hypnosis group, 3 resumed working, whereas none of the 6 control patients not working returned to work).
Of the six uncontrolled outcome reports in the literature, two reports from the Manchester group on the same sample of patients deserve the greatest attention because of the large sample size (greater than the samples in all the other studies combined), the thoroughness and detail of outcome data collected and reported, and the long follow-up period. These publications provide the best information to date on the overall potential for hypnosis treatment in routine use for management of IBS. The first of these reports (Gonsalkorale et al., 2002) details the outcomes of a series of 250 consecutive IBS patients treated by several hypnotherapists in the investigators’ Manchester clinic with a course of hypnosis of 12 or more weekly sessions according to the well-delineated protocol developed by Whorwell. The median posttreatment bowel-symptom severity score for the group was 53% lower than before treatment; median ratings of all bowel symptoms were reduced by more than half; anxiety and depression scores were substantially decreased; and quality-of-life scores were improved. The second report (Gonsalkorale et al., 2003) complemented the previous one with long-term follow-up data on 204 patients from the clinic’s case series. The follow-up was conducted annually for up to 5 or more years after treatment. The authors reported that 71% of patients could be classified as treatment responders (defined as all patients who rated themselves very much better or moderately better after treatment). Follow-up assessments showed that 81% of treatment responders retained their improvement up to 5 or more years beyond the end of treatment. The percentage of patients remaining better did not decrease with years from end of treatment, indicating that the maintenance of benefit does not degrade much over the subsequent years. The investigators found that the great majority of patients had continued to further improve in their bowel symptoms after treatment and that most of the 19% who reported deterioration at follow-up reported only partial worsening rather than full relapse. Furthermore, patients were significantly less likely to use medications for IBS and consult physicians after treatment and had far fewer nongastrointestinal symptoms (like back pain or headache).
Palsson et al. (2002) reported an uncontrolled study on the treatment of 18 subjects that, even though small and primarily aimed at examining the mechanism of the therapeutic effects, is noteworthy because it was the first test of an entirely scripted 7-session therapy protocol, the same as used in their controlled study described above. Seventeen out of 18 patients were judged improved from treatment in this trial. Abdominal pain, bowel-functioning abnormality, and bloating all improved substantially after treatment, as did anxiety, overall psychological distress, and somatization scores.
Harvey and colleagues (1989) published a report on treatment of 33 patients that was unusual compared to other studies in this domain in two ways. First, the intervention was the shortest one tested to date, consisting of only four hypnosis sessions over 7 weeks. Second, some of the patients were treated in groups of up to 8 and some individually. The investigators found that 20 of the 33 patients (61%) improved from the treatment and reported that 11 of those had lost all their symptoms.
Whorwell, Prior, and Colgan (1987) reported on outcomes for 50 patients treated with a course of hypnotherapy in a report that included the sample of 15 hypnotherapy patients from their 1984 study. They found different success rates based on the characteristics of patients and symptoms. Patients who had classic IBS symptoms had a 100% success rate in treatment if they had no comorbid psychopathology and 60% success if they did have comorbid psychopathology; atypical IBS cases had a 48% success rate; and patients over age 50 generally did poorly, with only 25% rate of success. These early conclusions regarding the poor prognosis of older patients and those with comorbid psychiatric symptoms have been superseded by the group’s later analysis of predictors of outcome in a much larger sample (Gonsalkorale et al., 2003), which showed only very small adverse effects of age or psychopathology on outcome.
Gonsalkorale et al. (2004) investigated cognitive changes following a 12-session course of hypnotherapy in 78 IBS patients, using a scale designed to measure dysfunctional cognitions related to functional bowel disorders. Unhelpful bowel-related cognitions improved significantly, with reduction seen in the total cognitive score and all component themes related to bowel function. Overall symptom reduction correlated with improvement on the cognitive scale. The authors also described clinical outcomes for this sample and these were in line with their previous reports, but since these patients were also in the larger sample described in the investigator’s prior papers (Gonsalkorale et al., 2002), the outcomes will not be detailed here.
Finally, Vidakovic-Vukic (1999) in Holland reported “significant” improvement in 24 out of 27 IBS patients treated with a course of 10 to 14 hypnosis sessions but provided no statistical results in what appears to have been a preliminary report that was never followed by the intended report of the more complete study.
As stated earlier, hypnosis has been reported to be successful in impacting IBS symptoms in all the 11 outcome studies published to date. Where success rates have been reported, they have ranged from 61% to 100% (see Table 1) with a median success rate of 87% (for this calculation, Whorwell et al.’s 1987 report is excluded, because it included multiple success rates for different subgroups and partly included subjects from their previous 1984 study). Although investigators have not used a consistent approach to deciding who is a treatment responder, it nonetheless appears clear from these high rates of success that the great majority of IBS patients treated with hypnosis enjoy marked improvement. This is a strikingly high degree of success considering the fact that most of these studies have documented that they only enrolled subjects who had already failed standard medical treatment, and many of the treated patients had already tried multiple other interventions without benefit. This is a fact often overlooked in discussions of this form of treatment for IBS and highlights a great potential for hypnosis as an adjunct to medical treatment to enhance outcomes in management of the disorder.
Degree of Impact on Symptoms
Due to the heterogeneity of outcome methods in the published studies and lack of detail in many of them, it is hard to get an overall sense of the degree of impact on symptoms. Some of the studies have reported only the statistical significance of change rather than the actual pre- and posttreatment scores. The two research groups that have published repeated studies using consistent treatment protocols, the Manchester group in England and our team at UNC-Chapel Hill, have also been the ones providing the most detailed outcome data, but they have used very different assessment methods and different statistics (medians vs. means) to report the treatment changes. The Manchester group has used a 10-day retrospective bowel-symptom questionnaire (Francis, Morris, & Whorwell, 1997), whereas our group has used 2-week summary scores of 5-point daily symptom diary ratings where respondents are asked to rate the worst instance of each symptom each day.
It is not possible to know if the outcomes from the two teams are equivalent, because it is unknown how similarly these measures respond to change. However, regardless of how it is measured, the impact on symptoms documented by both groups has been very substantial. Both teams have consistently found abdominal pain severity and bowel-habit dysfunction to be reduced by half or more in the patient samples. The Manchester group has also found bloating severity to improve by more than half, whereas our studies show a mean reduction of one-fourth or less in bloating (Palsson et al., 2002).
The best single quantitative summary of the impact of hypnosis treatment on IBS is perhaps the reduction from a score of 334 to 155 (values calculated from data presented in the paper) on the composite bowel-symptom severity index, reported by Gonsalkorale and
colleagues (2002) in their sample of 250 patients. Published cut-offs for classifying IBS severity on this scale (Francis et al., 1997) place IBS patients with scores above 300 in the severe range and those with scores of 175 or less in the mild range. In other words, the Manchester group has demonstrated in their large case series that the median severity of IBS moves from a severe case to a mild case after a course of therapy with hypnosis.
Scope of Therapeutic Benefits
In addition to changes in bowel symptoms, the studies that have measured other outcomes have consistently found anxiety, depression, somatization, and the number of extra-intestinal symptoms to also be substantially reduced after treatment and quality-of-life scores to be enhanced. These changes correlate moderately with the degree of therapy-related bowel-symptom reduction (Gonsalkorale et al., 2002) and are sufficiently large to be a marked extra benefit to patients beyond bowel-symptom improvement. These changes in psychological symptoms associated with treatment are furthermore highly clinically significant: Using conventional clinical cut-offs for the Hospital Anxiety and Depression Scale, the Manchester group calculated that hypnosis treatment reduced by half or better the percentage of individuals who had clinical anxiety (34.3% vs. 69.3%) or depression (14.6% vs. 36.1%) in their series of 250 patients (Gonsalkorale et al., 2002).
Predictors of Treatment Response
As described above, an early report by Whorwell and colleagues (1987) indicated that patients older than 50 were substantially less likely to benefit from hypnosis intervention. Later work by this group has confirmed that there is some statistically significant negative effect of greater age, but the impact on therapy responsiveness seems to be quite small, accounting only for about 2% in outcome scores. Age is therefore unlikely to be a significant barrier to benefit from this treatment.
Whorwell and colleagues also reported in 1987 that patients with identified psychopathology were less likely to respond to treatment. Little data have been published since then to confirm or disconfirm that observation. Galovski and Blanchard (1998) reported a significant positive correlation between therapeutic response and the presence of psychiatric diagnoses, but this result is likely to be unreliable because of the shortcomings of the statistics (Pearson correlation using ordinal data from only 11 subjects). In short, the relationship between psychopathology and treatment response remains unclear and needs to be investigated in future work.
Finally, Gonsalkorale et al. (2002) found in their analyses of 250 patients that males with diarrhea-predominant bowel pattern had a significantly lower average response rate than other IBS patients. This is an interesting and unexplained observation and this interaction between gender, bowel-habit type, and therapeutic response should be investigated in future work.
Maintenance of Therapeutic Gains
All available data indicate that most patients who experience significant symptom relief after a course of hypnosis treatment can expect to stay better for 1 or more years after treatment. Two small studies (Palsson et al., 2002; Whorwell et al., 1984) showed that all patients who responded to treatment were still reporting improvement at 10 months and 18 months after treatment, respectively. Gonsalkorale et al.’s (2003) long-term follow-up of 204 patients demonstrated that approximately 4 out of every 5 patients who respond to treatment fully retain their therapeutic benefits for years (outcome assessed for 1 to 5 years after treatment) and that most continue to see further improvement in bowel symptoms after the end of the treatment course.
Limitations of Current Research and Need for Future Work
- Design limitations. The chief factor limiting the impact of the growing empirical literature on hypnosis for IBS is the design of most of the studies published to date. It is clear from review of this body of publications that more and better randomized controlled trials are sorely needed in this domain, especially studies comparing hypnosis treatment to other interventions with demonstrated impact on IBS (Whitehead, 2004). For most of the time period that this research covers, no IBS-specific medications were available, and there were no particular medications with recognized efficacy in treating IBS. Now, however, there are two IBS-specific drugs on the market and more on the way, and a recent meta-analysis confirms that antidepressants help IBS (Jackson et al., 2000). Hypnosis intervention should be compared side-by-side to these medications in randomized trials. The effects of hypnosis and cognitive-behavioral therapy (the best established other psychological treatment for IBS) should also be compared in a parallel fashion. Furthermore, combinations of hypnosis and medications that are effective for IBS are likely to achieve even better results than either hypnosis alone or medications alone, since hypnosis presumably achieves its impact on symptoms via different mechanisms than medications, and trials should therefore be conducted on such combined interventions.
- Varied and unclear nature of the tested interventions. The hypnosis interventions that have been tested for IBS in the 11 studies described above have varied considerably. The number of treatment sessions has ranged from four to twelve or more. The study by Harvey and colleagues (1989) that used the shortest course of treatment (four sessions over 7 weeks) was also the one with the lowest success rate. At the other end, the use of 12 or more sessions, which is now the norm for the Manchester group, does not seem to add anything to success rates, as the studies with the highest success rates have used 7-session protocols (Palsson et al., 2002; Whorwell et al., 1984). However, the Manchester group’s outcome data with a 12+ session therapy course (Gonsalkorale et al., 2002) suggest that they may be achieving greater general impact on symptoms than other investigators (although this is not possible to determine directly due to different measurement methods), so there could be some advantage to a long sequence of sessions. Determining the optimal “dose” that achieves the fullest benefit for most patients while remaining cost-effective is an important goal for future research in this area.
- Heterogeneity of outcome measures. Comparison of the effectiveness of therapy across studies requires basic comparability in outcome measures. This has not been achieved in research on hypnosis for IBS. Bowel symptoms, emotional symptoms, and quality of life have all been measured in several different ways. Success rates have also been determined in ways that vary from study to study and make direct quantitative comparison questionable. It is important that investigators make a concerted effort to achieve uniformity in outcome measurement, as this is a major weakness in achieving a cumulative body of evidence in this research area. Our team has recently adopted Francis et al.’s (1997) IBS severity scale as an outcome measure in our hypnosis work, partly to make our outcomes directly comparable to those of the Manchester group, which has published most research in this area. It would be highly desirable for other groups to do the same.
d Lack of detailed description and standardization of treatment. The description of the hypnosis treatment tested in the published studies has often been entirely inadequate to give other investigators or therapists an idea of the therapy approach that was used. This problem is somewhat lessened by the fact that the bulk of the studies have followed Whor- well’s Manchester protocol or our North Carolina protocol, both of which are available in detail from the investigators. However, published studies that do not follow such well-delineated and generally available methodologies need to contain a thorough description of the intervention, detailing the kind of therapeutic suggestions and imagery that were used, if they are to be of value to the field.
Summary and Conclusions
The literature on hypnosis treatment for IBS is to date a collection of studies that are generally small and sometimes poor in design. However, these studies report consistent IBS symptom reduction that is very substantial, is associated with broad improvement in psychological well-being and life functioning, can be just as readily achieved by patients who have not had relief from standard medical care, and may generally be expected to last for years. The cumulative weight of this research, which now includes a placebo-controlled trial, four other controlled trials, and a very large case series with long-term follow-up, leaves little doubt that hypnosis is the best choice, rivaled only by cognitive-behavioral therapy, for effective treatment of severe IBS patients who fail to respond to medical management.