Comprehensive self-management for irritable bowel syndrome: randomized trial of in-person vs. combined in-person and telephone sessions.

There is increasing evidence that psychological treatments including cognitive behavioral therapy (CBT) are effective strategies for the management of patients with irritable bowel syndrome (IBS) ( 1 ). The use of CBT is based on the hypothesis that IBS symptoms are due, at least in part, to dysfunctional cognitions about visceral sensations ( 2 , 3 ). CBT has been compared to standard medical care using both an individual presentation ( 4 , 5 ) and small groups ( 6 ). When compared to education alone, CBT was shown to decrease gastrointestinal (GI) symptoms to a greater degree ( 7 ). In a study using group therapy, the psychoeducation support group was as effective as the cognitive therapy group in reducing GI symptoms, and both were more effective than daily stress monitoring ( 6 ). In Great Britain, a nurse-delivered CBT plus mebeverine intervention when compared to mebeverine alone was found to produce greater reductions in symptom severity and that the response could be sustained for 12 months ( 8 ). In prior work our team combined CBT with education, relaxation training, and diet management into an 8-session comprehensive self-management (CSM) program for patients with IBS ( 5 ). When compared to usual care, this full 8-session CSM delivered by an advanced practice nurse resulted in improvements in GI symptoms and quality of life (QOL) that were sustained at 12 months post- intervention ( 5 ). One problem noted in this study was that the need for weekly travel to the intervention site imposed a burden on patients. Alternatives to face-to-face approaches for the effective delivery of therapist-intensive interventions are prompted by concerns over travel time, clinical availability, and disruptions in patients’ daily activities. Telephone, web-based, and video-conferencing may all yield benefits in terms of access and convenience ( 9 , 10 ). Previous research has demonstrated the feasibility of using the telephone to delivery interventions for depression ( 11 – 14 ), anxiety disorders ( 9 , 15 ), cancer- symptom management ( 16 , 17 ) and smoking cessation ( 18 , 19 ). In most cases cognitive behavioral interventions delivered by telephone or a combination of telephone and in-person sessions were as effective as those relying exclusively on in-person groups ( 15 ), and more effective than usual care ( 13 , 17 , 20 ), attention control ( 14 ), or care management approach ( 12 ). However, in one study, cancer-symptom management by telephone did not differ from usual care ( 16 ). The goal of the current study was to modify the previously tested CSM intervention for delivery by telephone, and compare the efficacy of CSM delivered mainly by telephone (CSM-T/IP) to in-person delivery (CSM-IP) and to Usual Care. We hypothesized that those in the CSM-IP and CSM-T/IP groups would demonstrate decreased GI symptoms and enhanced QOL compared to those in the UC group, and that outcomes in the CSM-T/IP group would be comparable to the CSM-IP group.

WHAT IS CURRENT KNOWLEDGE Irritable Bowel Syndrome is common but treatments are limited. Psychological interventions can be an alternative to multiple medications. WHAT IS NEW HERE Nutrition and cognitive behavioral therapy is effective in decreasing the distress caused by IBS. Therapy is effective whether delivered by telephone or in-person.

Table 2 is the consolidated standards of reporting trials (CONSORT) diagram showing numbers of subjects screened, randomized, and followed. One hundred and eighty eight subjects were randomized and only 12 of these (6%) failed to provide any follow-up data. Eighty-seven percent of subjects assigned to CSM received at least seven out of nine sessions. Homework was assigned between treatment sessions and 90% of CSM-T/IP subjects and 91% of CSM-IP subjects meet our expectations for homework completion, i.e., at least 75% of assigned work was completed on at least 75% of the sessions. Demographics and baseline clinical characteristics are given in Table 3 . Subjects in this sample were mainly female, white and relatively well educated. Seventy-five percent of the sample was working, 11% were retired, 20% were not working when they enrolled in the study, and 5% were stay-at-home parents. Over half of those participants who were working gave job titles that were classified as professional. Overall, participants reported having typical IBS symptoms for 10 years (SD 14) prior to their diagnosis: 31% started in childhood (4 to < 17.99 yrs), 62% started between 18 – 49 years, and 8% reported that their symptoms started at 50 years or older. Half of the subjects were diagnosed by a gastroenterologist and the rest by a primary care provider (i.e., family practice physician, internist, nurse practitioner, physician’s assistant, or other). Average age of first IBS diagnosis was 27 years (SD 14). Neither demographic or IBS characteristics differed statistically among groups. Twelve participants in this study did not provide any follow-up date. The participants who had no follow-up data (6%) or partial follow-up data (13%) did not differ on demographic variables or IBS characteristics from those participants that had complete data (86%). Table 4 shows the baseline means and change scores of the primary and secondary outcome variables in the three treatment groups. The two CSM groups show a large improvement from baseline to 3 months in the primary outcomes which persists at 6 months and 12 months. Note that for the QOL an increase indicates improvement, while for the other variables a decrease (minus value) indicates improvement. There was no significant difference in the amount of improvement in CSM-T/IP compared to CSM-IP. Figure 1 , 2 , and 3 further illustrates these results for the two primary outcome variables. Results are similar for most secondary outcomes, with both CSM groups being better than Usual Care and the CSM groups not differing from each other. However, for psychological distress there is a trend towards better outcomes in the CSM-T/IP than in CSM-IP while for CSFBD there is a trend towards better outcomes in CSM-IP than in CSM-T/IP. Note also that the treatment differences for psychological distress are due to the worsening of symptoms in the Usual Care group more than the improvement in the intervention groups. The intervention effect was stronger for the WPAI measure of interference with daily activities than on the WPAI measure of interference with work, partially due to the smaller sample of people who were employed at both baseline and follow-up. These results are based on models with only main effects. Additional analyses tested for interaction effects. The interaction between follow-up time and treatment group was not significant for any outcome, though it was nearly significant (p = .088) for CSFBD. Overall, these results indicate that the treatment effect seen at 3 months persists at 6 and 12 months with little diminution. Several ancillary analyses were conducted in an effort to determine which particular GI symptoms were most affected by CSM. Table 5 presents the baseline means and change scores of the individual GI symptoms that were included in the IBS Symptom score. The two CSM groups show a large improvement in abdominal pain/discomfort and intestinal gas which persists through 12 months. The changes in bloating, constipation, diarrhea, and urgency were greater in CSM than UC but not significantly different in the mixed model test. Table 6 presents the percent of subjects with at least 50% improvement in the IBS symptom score and abdominal pain/discomfort score, and odds ratios of CSM-T/IP and CSM-IP relative to UC. The odds ratios were all large, greater than 2.3, and almost all are statistically significant. Adverse events include one participant who experienced suicidal thoughts, which lead to her withdrawing from the study.

Every day subjects rated 26 symptoms on a scale of 0 (not present), 1 (mild), 2 (moderate), 3 (severe) or 4 (very severe). Of these, six were GI symptoms related to IBS: abdominal pain or discomfort, bloating, constipation, diarrhea, intestinal gas, and urgency. An IBS symptom score was computed by first determining the severity of the worst IBS symptom on each day to get an IBS severity for that day, then collapsing across the diary days for each subject to determine the percentage of days with moderate to very severe GI symptoms ( 30 , 31 ). The IBS-Quality of Life (IBSQOL) questionnaire is a 42-item questionnaire with nine scales: sleep, emotional, mental health beliefs, energy, physical functioning, diet, social role, physical role, and sexual relations ( 32 ). Example questions are “ How often did your IBS make you feel fed up or frustrated ” e.g., 1 (always), 2 (often), 3 (sometimes), 4 (seldom) to 5 (never); or “ My IBS affected my ability to succeed at work/main activity ” e.g., 1 (strongly agree) to 5 (strongly disagree). The scales are transformed to a standard 0 to 100 scale. A total score was computed by averaging all but two of the scales (eating/diet and sexual relations). The eating/diet scale was omitted because participants in the CSM groups were encouraged to avoid foods that cause problems for them. The sexual relations scale was omitted because it was missing for a large fraction of the sample had not been sexually active in the previous 4 weeks. Extensive and acceptable validity and reliability tests have been conducted ( 32 ). Internal consistency (Cronbach’s alpha) for the scales ranged from α = 0.73 to 0.93 for this study. The Brief Symptoms Index (BSI) includes 53 symptoms that are collapsed into 9 subscales: depression, hostility, somatization, interpersonal sensitivity, psychoticism, obsessive-compulsive, anxiety, phobic anxiety, paranoid ideation, and a mean score of all items (Global Severity Index). The subject is asked to consider the last 7 days then rate each symptom from 0 (not at all) to 4 (extremely) distressing. However, for the follow-up assessments the BSI-18 was used, which includes 3 subscales anxiety, depression, and somatization. This BSI is based on the Symptom Checklist-90-Revised (SCL-90R), which has been normed on men and women and adolescents in non-patient and adult psychiatric outpatient and inpatient samples. Internal consistency of the subscale scores are reported to range from α = .77 to .90 and test-retest reliability (one week interval) was .78 to .90 ( 33 ). For this study the internal consistency was α = .88 for Global Severity Index, α = .80 for Anxiety, α = .85 for Depression, α = .69 for Somatization at Baseline. The Cognitive Scale for Functional Bowel Disorders (CSFBD) describes 25 cognitive beliefs related to functional bowel disorders ( 34 ). The items are rated from 1 (strongly disagree) to 7 (strongly agree). A typical item is, “ I often worry that there might not be a bathroom available when I need it. ” The CSFBD has high concurrent criterion validity, acceptable convergent validity, and high content validity and face validity with minimal social desirability contamination ( 34 ). The internal consistency for this study was α = .937. The summary score was the mean of all items. The Work Productivity and Activity Impairment Questionnaire (WPAI) has been adapted for persons with IBS ( 35 ). It includes 9 questions related to the impact of IBS on work and other regular activities. Construct validity is acceptable when tested against known measures in employed individuals affected by a health problem. The test-retest (1-day) ranged from r = .71 to .95 for the items ( 35 ). For this analysis two scales will be used, the Overall Work Productivity Loss (missed work and work impairment due to IBS) and the Daily Activity Impairment scales (impairment while working due to IBS). Quality control and safety measures included the use of a standardized protocol manual and standardized training. The two advanced practice nurses who delivered the intervention were aware of the group assignment and provided both interventions. A third nurse collected the follow-up data and was blinded to the participant’s group assignment. Weekly staff meetings were held to review the study implementation. The CSM sessions were audio recorded to assess adherence to the intervention protocol. Three randomly selected recordings per participant were reviewed using a checklist for compliance with the protocol. Our Data Safety Monitoring Board met prior to the onset of the study and yearly until the study was completed. The goal of the study was to finish with 180 subjects with analyzable follow-up data. With that sample size there would be 81% power if the change in QOL or IBS symptom score was 0.5 SD higher in the two CSM than in the usual care group, or 93% power if the change in QOL or IBS symptom score was 0.6 SD higher. These power calculations are based on ANOVA at one time point, but actual analyses use data from three time points and control for baseline, hence power will be somewhat higher. Our earlier study showed a difference of about 0.7 SD between CSM and usual care. The two primary outcome variables (IBS symptom score and QOL) were analyzed separately. For each, data from the three follow-up time points were analyzed together using a mixed model, with subject as a random effect and treatment group (three levels, CSM-IP, CSM-T/IP, UC) and measurement time (three levels, 3-mo., 6-mo., 12-mo) as fixed factors. Analyses controlled for covariates that might be related to the outcomes: baseline levels of the outcome variable as well as QOL and psychological distress at baseline. The main effect for group was used to test whether the three treatment groups differ with respect to GI symptoms at follow-up. Whenever this overall main effect was significant, univariate ANOVAs were done at each follow-up time point and appropriate contrasts were used to test all pairwise comparisons (CSM-T/IP versus Usual Care, CSM-IP versus Unusual Care, and CSM-T/IP versus CSM-IP). Similar analysis were done for secondary outcomes. Two binary symptom improvement variables were created, defined as a 50% decrease from baseline in the IBS symptom score (% of days with at least moderate symptom) and a 50% decrease in abdominal pain/discomfort score. Logistic regression was used to estimate the odds ratios for probability of improvement in each CSM group relative to Usual Care, controlling for baseline. These analyses will allow comparison of our results to those reported in two recent meta-analyses ( 1 , 36 ) which reviewed a number of studies using psychological interventions for IBS where 50% improvement in symptoms was used as the primary outcome. Some pain researchers would regard a 50% criterion as excessively stringent and prefer 30% threshold for clinical significance ( 37 ), however the 50% threshold is what has been reported in the IBS literature. An intent-to-treat approach was used in this study. That is, every effort was made to collect follow-up data on every subject who was randomized, regardless of how many intervention sessions were attended, and all subjects on whom follow-up data were available were included in the analysis.

The results in this trial demonstrate that comprehensive self-management (CSM) therapy is effective, whether delivered primarily by telephone or delivered entirely in-person. Both approaches were more effective in decreasing GI symptoms and increasing QOL than usual care. The magnitude of the treatment effect was virtually the same for the two delivery modalities. Moreover, these improvements persisted through the 12 months post-randomization follow-up (i.e., 9 months post intervention). The GI symptoms most strongly impacted by the intervention were abdominal pain/discomfort and intestinal gas. This is interesting in light of an earlier report ( 38 ) showing that these two symptoms were most strongly associated with reduced QOL. There was also a strong effect of CSM on cognitive beliefs and work loss and impact on daily activities. Given the large financial impact of IBS due to lost productivity, this finding has important implications for potential societal cost-savings from CBT ( 39 , 40 ). Further research using more extensive measures of productivity loss and health care costs would be useful in evaluating the cost effectiveness of CSM therapy. It should be noted that the CSM-T/IP delivery mode was not designed to be cheaper to deliver than CSM-IP in terms of therapist time, since all sessions were approximately 60 minutes in length whether delivered by telephone or in-person. The main benefit of CSM-T/IP over CSM-IP was the decreased travel time to the therapist’s office. The CSM intervention is multi-faceted, consisting of education and reassurance, nutritional counseling, relaxation strategies, and cognitive restructuring ( 5 ). While all participants received all components, the therapists observed that individuals differed in which components they found most helpful. The use of a workbook allowed participants to continue using strategies that they found helpful after the end of treatment sessions. One challenge with delivering CSM by telephone became apparent early on in the study: some participants were multitasking while on the telephone with the therapist. The study protocol was quickly changed so that prior to the first phone therapy session, the therapist discussed with the participant the importance of being in a quiet place without distractions during the telephone session, and helped the subject problem-solve about how to arrange that. Several studies have shown the efficacy of CBT delivered in-person for IBS ( 1 , 5 – 7 ). Recently studies have investigated alternative delivery modes for CBT in IBS. An earlier study by our group ( 5 ) found a workbook plus one in-person session to be less effective overall than eight in-person CBT sessions. Lackner ( 41 ) found that four CBT sessions in combination with a self-help book were as effective as ten sessions of in-person CBT. Delivery of CBT by trained clinic nurses was also effective ( 8 ). Although telephone delivery of psychotherapy has been studied for depression and anxiety ( 9 , 10 ), this is the first study of a CBT-based intervention for IBS delivered primarily by telephone. As in any psychological intervention study, it was not possible to blind subjects as to treatment assignment. It is possible that some of the observed treatment effect could be due to the placebo effect. But the fact that treatment differences persisted nine months after the end of treatment implies long-term changes in behaviors, habits, attitudes, or beliefs among subjects in the CSM intervention groups. It should be noted that the CSM-T/IP intervention was not entirely by telephone. The first two sessions were in-person, in order to build a therapeutic relationship with the therapist and teach breathing exercises before starting the telephone sessions. The final session was also in-person, in order to provide a transition to self care. We feel that these three in-person sessions were important to the success of the CSM-T/IP intervention. An intervention that was totally conducted by telephone might not be as effective. Anecdotally, the research staff noted that quite a few subjects volunteered a preference for mode of delivery of CSM, with some having a preference for CSM-T/IP because of the reduced travel time, while others indicated a preference for CSM-IP due to the personal interaction. Since this study found both versions to be effective with no evidence that one is better than the other, a flexible delivery model could be the best approach, with patients allowed a choice in delivery mode for sessions 3 through 8. In conclusion, this report has shown that the mulitifaceted CSM intervention delivered primarily over the phone by psychiatric nurse practitioners is an effective option for treating IBS. Further investigation is warranted into how best to incorporate such an intervention into clinical practices.