Therapist telephone-delivered CBT and web-based CBT compared with treatment as usual in refractory irritable bowel syndrome: the ACTIB three-arm RCT

Study design

We conducted an open, pragmatic randomised controlled trial (RCT) in primary and secondary care to determine the clinical effectiveness of TCBT and WCBT in patients with refractory IBS.

Setting

Patients were recruited from general practices from Wessex and South London Clinical Commissioning Groups (CCGs), and from one Southampton and two London secondary care sites [Southampton University Hospital Trust (SUHT), King’s College Hospital (KCH) and Guy’s and St Thomas’ NHS Foundation Trust (GSTT)] between May 2014 and March 2016. Research was co-ordinated by two academic centres: the University of Southampton and the Institute of Psychiatry, King’s College London. Treatment took place at participants’ homes via telephone and the internet. Therapists were based at the South London and Maudsley NHS Foundation Trust (SLAM).

Ethics approval and research governance

Ethics approval was awarded by the National Research Ethics Service (NRES) Committee South Central – Berkshire on 11 June 2013 (reference number 13/SC/0206) and local research governance approval was obtained from all participating CCGs.

The trial was registered with the International Standard Randomised Controlled Trial Number (ISRCTN) under the reference number 44427879.

Summary of any changes to the protocol

A total of four substantial amendments were approved by the ethics committee and included a change to the primary outcome, the addition of a 24-month outcome time point and an increase in the recruitment target. Two minor amendments were submitted and acknowledged (Table 1).

Training of co-ordinating centres

Research teams from the co-ordinating centres in Southampton and London were trained in January 2014 at the University of Southampton. The processes covered included telephone screening of patients accepting the invitation and the information required for fully informed consent, therapist procedures, setting up sites (site files, logs and packs), completing the research team database and record-keeping, communication with sites and patients (texts for e-mails and letters) and monitoring LifeGuide (https://lifeguidehealth.org/; accessed 27 January 2019) queries and patient e-mails.

Recruitment of general practices

Expressions of interest from general practices were gathered by Wessex and South London Clinical Research Networks (CRNs) and passed on to the research teams, who then contacted the practices to provide further information on the trial. The number of general practice sites required was estimated based on information from the MIBS feasibility study. 1,15 An assumption of 30–80 patients with a computer diagnosis of IBS per general practitioner (GP) was made (2–5% prevalence and 1600 registered patients per GP). In addition, it was estimated that 5–10% of those would fulfil the inclusion criteria and be willing to participate in the trial (three to eight per GP). Thus, we initially anticipated recruiting approximately 30 general practices.

Recruitment of secondary care sites

Secondary care sites were identified for participation in ACTIB by expert contacts of the research team and comprised SUHT, KCH and GSTT. It was estimated that over 500 patients with IBS attend Southampton hospital GI clinics each year and over 1000 attend the London GI clinics, providing a large population of patients attending secondary care to invite to the study.

Training of sites

Training was undertaken by providing a written training manual and a training log was signed by all site personnel with a delegated role. The training manual included a list of study materials to be supplied by the co-ordinating centre. It covered procedures for identifying patients from the practice records using Read codes to identify adult patients with an IBS diagnosis and preparation of the invitation letters using a Microsoft Excel^® (Microsoft Corporation, Redmond, WA, USA) spreadsheet pre-populated with identification (ID) numbers, recruitment of patients from an opportunistic consultation using a pre-numbered pack, recruitment of patients using a poster (see Report Supplementary Material 1) in receptions, instructions for the blood test including postage of the blood samples and completing paperwork for the co-ordinators and pathology and instructions for completing a serious adverse event (SAE) form if needed. Prior to starting the search, all personnel were added to a delegation log and signed off by the principal investigator at the site. All sites received a site file and a copy of the NICE guidelines for IBS. Sites were kept up to date with trial progress and their individual performance by monthly e-mail updates from the co-ordinating centres.

Patient recruitment

Participating sites were asked to recruit patients by sending invitation letters, by opportunistically recruiting during consultations and by displaying posters in receptions. Because of the limited availability of therapy sessions, site initiation was staggered and the response rate monitored to ensure adequate recruitment and a steady workload for the therapists.

Primary care patients were identified by searching GPs’ lists for those with a diagnosis of IBS, by opportunistically recruiting patients presenting with symptoms consistent with IBS and by displaying posters in receptions. General practice administration staff ran searches for eligible patients using appropriate clinical diagnosis and symptom codes. GPs checked the lists of patients to be invited prior to the invitation letters being sent out to ensure that it was appropriate to contact them.

Secondary care patients were identified from gastroenterology clinics at SUHT, KCH and GSTT. When possible, clinic administration staff searched clinic lists for patients with a diagnosis of IBS. Potential participants were contacted by letter (sent from the clinic), which informed them about the trial and invited them to take part. When clinics needed more support, researcher administration staff were available to hand out packs and answer patient queries. The consultants checked the lists of patients to be contacted prior to the invitation letters being sent out to ensure that it was appropriate to contact them. Invitation letters were sent in batches.

Patients received an invitation letter (see Report Supplementary Material 2) and a patient information sheet (PIS) (see Report Supplementary Material 3) by post from the site, or from the GP/consultant, or from the research team following interest from the poster. Patients were asked to return the reply slip in a pre-paid envelope with contact details to the researchers either indicating that they were interested in being contacted or stating that they were not interested by ticking a list of potential reasons (and with an option for free-text responses). The researchers attempted to contact the patients who indicated that they were interested in participating for screening, either by telephone (including attempts during the evening and at weekends) or via text/e-mail, to arrange a convenient day and time to telephone.

Patient screening

Screening was conducted to ensure that potential participants met the inclusion criteria (Box 1), including being aged ≥ 18 years with refractory IBS [clinically significant symptoms defined by an Irritable Bowel Syndrome Symptom Severity Score (IBS SSS) of > 75], fulfilling Rome III criteria19 and having already been offered first-line therapies (e.g. antispasmodics, antidepressants or fibre-based medications) but still had continuing IBS symptoms for ≥ 12 months. Potential participants aged > 60 years were included only if they had undertaken a consultant review in the previous 2 years to confirm that their symptoms were related to IBS and that other serious bowel conditions had been excluded. This was because NICE guidelines9 advise that a new change in bowel habit in those aged > 60 years should be investigated further, as there is an increased risk of bowel cancer in this age group.

During the telephone screening, researchers followed a script and completed a paper screening questionnaire (see Appendix 1). Screening was conducted to assess the patient’s full eligibility for the study, check their understanding of the study procedures (including the need for a blood test), answer any questions to ensure that informed consent could be undertaken online and discuss the time commitment and the constraints around availability for therapy; it included the validating measures IBS SSS and Rome III. Screening also checked participants’ access to the internet.

Potential participants were excluded if they had unexplained rectal bleeding or weight loss; had a diagnosis of inflammatory bowel disease (IBD), coeliac disease, peptic ulcer disease or colorectal carcinoma; were unable to participate in CBT because of speech or language difficulties; had no access to an internet-enabled computer to be able to undertake the WCBT; had received CBT for IBS in the past 2 years; had had previous access to the MIBS website; or were currently participating in an IBS/intervention trial (Box 2).

General practitioners were notified of any of their patients who failed screening for unexplained weight loss or rectal bleeding so that the symptoms could be followed up.

Patient consent

Patients meeting the screening criteria were e-mailed instructions for the website address/uniform resource locator (URL) to log in to the LifeGuide online consent form (see Report Supplementary Material 4). The research team received an automated e-mail once the form was completed, which prompted the researchers to send another e-mail to the patient, customised for each general practice or secondary care site, with an instruction to make a blood test appointment for the next stage of screening.

Blood test

Patients telephoned their general practice surgery or gastrointestinal (GI) clinic to make an appointment for the blood test. When patients already had recent (within the previous 3 months) blood test results, these were requested from the site so that patients were not required to undergo another set of tests. The function of the blood tests was to exclude alternative diagnoses to IBS, as recommended for IBS diagnosis in the NICE guidelines. 9 The blood tests undertaken were full blood count (FBC) for anaemia, tissue transglutaminase antibodies (TTG) as screening for coeliac disease and C-reactive protein (CRP) for inflammation (marker for IBD). The blood tests were undertaken by practice nurses, by GPs within the surgeries or by phlebotomists or research nurses at the secondary care sites. Sites were pre-supplied with vacutainers, a form to fax back to the research team and a form to post with the sample. Patients were instructed to arrive with their ID number so that all forms could be identified by ID number only. Samples were sent to the SUHT pathology laboratory for testing and were then destroyed. The results were posted to the research team and the GP and were checked by the chief investigator. Blood sample results were stored in a locked filing cabinet.

Abnormal blood test results

A patient with abnormal blood test results (e.g. indicating anaemia or a positive TTG) was excluded from the study and referred back to his or her GP for further assessment and the GP was informed of the abnormal results by post. Patients with a CRP level above the normal laboratory range were phoned by the team and given the option to have a second test after 4 weeks (as CRP can be raised temporarily because of a minor intercurrent illness). A second high CRP result excluded them from the trial and their GP was informed of the test result.

Baseline questionnaire

Patients with acceptable blood test results were e-mailed instructions to log on to LifeGuide to complete their baseline questionnaire (see Appendix 3) and, on completion, the team received an automated e-mail that prompted them to initiate patient randomisation.

Randomisation

Randomisation was provided by the randomisation service at the UK Clinical Research Collaboration-registered King’s College London Clinical Trials Unit (CTU) and accessed by the research team via a web-based system. Randomisation was at the level of the individual, using block randomisation with randomly varying block sizes stratified by centre (Southampton general practices, Southampton secondary care, London general practices, London secondary care). The CTU procedure was as follows: on completion of the baseline questionnaire, the trial manager or research assistant electronically submitted details of each participant to the CTU. This included participant ID number, site, initials and date of birth. The system immediately notified the unblinded researchers and recorded the randomisation outcome. Research staff were allocating patients to therapists so could not be kept blinded.

Once a patient was randomised to the study, a letter was sent to their GP (see Report Supplementary Material 5) to inform them of the patient’s participation, their allocated group and their blood test results. Patients were e-mailed instructions to receive their group by logging on to LifeGuide as promptly as possible to enter one of three codes. This directed them to one of three alternative web pages within LifeGuide, allowing them access to the relevant intervention. Patients were asked to register, which triggered future automated reminder e-mails to them at each of the follow-up collection points and directed them to further instructions for each arm of the trial. Having separate sites ensured non-contamination of treatments. Patients in the TCBT arm were asked not to share their therapy manual with others to avoid cross-contamination.

Interventions

Two active interventions were assessed in the study: TCBT with a detailed patient manual; and low-intensity WCBT (the Regul8 programme developed in the MIBS trial20), with some therapist support.

Those in the control arm received TAU. Patients in the two active intervention arms also received TAU.

The CBT content of the two treatments was the same. The CBT was based on an empirical cognitive–behavioural model of IBS,20,21 which specifies that factors such as stress and gastric infection trigger the symptoms of IBS, which are then maintained by patients’ cognitive, behavioural and emotional responses to the symptoms. For instance, if a patient becomes anxious (emotion) about the symptoms, believes that he or she has no control over them (cognitions) and responds by avoiding social situations (behaviour), this can increase anxiety and maintain symptoms through the link between a heightened autonomic nervous system and the enteric nervous system. This model was used to structure the content of the therapy sessions.

The therapy consisted of education, behavioural and cognitive techniques, aimed at improving bowel habits, developing stable healthy eating patterns, addressing unhelpful thoughts, managing stress, reducing symptom focusing and preventing relapse (Table 2).

The two main differences between the therapy trial arms were:

1. The amount of therapy contact was predetermined (TCBT participants received 8 hours and WCBT participants received 2.5 hours of therapy). Those in the TCBT arm had six 1-hour telephone sessions with a CBT therapist at weeks 1, 2, 3, 5, 7 and 9 and homework tasks. They also received two 1-hour booster sessions at 4 and 8 months. WCBT participants received three 30-minute telephone therapy support calls at weeks 1, 3 and 5 and two 30-minute booster sessions at 4 and 8 months.

2. The TCBT patients used a self-management CBT manual and the WCBT patients had access to an interactive website.

Treatment as usual

Patients in all three arms received TAU, with the control arm being TAU alone. TAU was defined as continuation of current medications and usual GP or consultant follow-up with no psychological therapy. All general practice sites or secondary care sites involved in the study received a copy of the NICE guidance for IBS9 at the start to ensure that all clinicians had the standard best practice information on IBS management. They also received a deskside reminder (see Report Supplementary Material 7) to remind them of the guidelines, protocol guidance on prescribing psychological therapies and inclusion criteria. All participants received a standard information sheet on lifestyle and diet in IBS based on the NICE guidance (see Report Supplementary Material 6). Information was collected on any changes in IBS treatment/management during the study, and numbers of GP and consultant consultations were recorded for all three arms. The TAU-alone participants had access to the WCBT website (but with no therapy support) at the end of the 12-month follow-up period.

Treatment adherence for intervention arms

Treatment adherence was defined separately for the two active treatment arms. Participants allocated to TCBT who completed at least four of the initial telephone CBT sessions were deemed as having adhered to treatment. Those who were offered WCBT and completed four or more website sessions and at least one telephone support session were considered treatment adherent.

Withdrawal from treatment and/or follow-up

In accordance with good clinical practice, patients were free to withdraw from the treatment and/or follow-up at any time without this affecting their medical care. Therapists made two attempts to contact non-attenders of the therapy sessions before withdrawing patients. All information on the event was collected in the drop-out report form (see Report Supplementary Material 8) and added to the CTU commercial data entry system [InferMed MACRO (InferMed, London, UK)].

Blinding

As with any therapy trial, blinding was not possible for participants or therapists. It was also impractical for the research assistants, who liaised with the participants and therapists regarding administration tasks. However, blinding was pre-planned for the outcome assessors and the trial statisticians, as described below.

After the database was locked, a decision was taken to use multiple imputation (MI) to deal with missing data in the formal statistical analysis (see Statistical methods), which meant that the trial statistician would become aware of treatment allocation when carrying out these analyses. However, these analyses were undertaken at the end so that any that could be carried out without revealing treatment allocations were carried out by a blinded statistician.

Outcome measures were collected via the web (when possible) and were patient reported. The researcher who contacted patients by telephone to capture primary outcome data in a short questionnaire (see Appendix 4), for those patients who did not complete follow-up questionnaires after reminders, was kept blinded to the participant’s treatment group to avoid bias. Statisticians, the principal investigators and all oversight committee members were also kept blinded to treatment group.

Baseline data

The baseline data included sociodemographic details, current medication, past medical history and medications, duration of IBS symptoms, previous or current psychiatric diagnoses, Rome III, IBS SSS, Work and Social Adjustment Scale (WSAS) , EuroQol-5 Dimensions (EQ-5D), Client Service Receipt Inventory (CSRI), Hospital Anxiety and Depression Scale (HADS), Cognitive scale for Functional Bowel Disorders (CG-FBD), Brief Illness Perception Questionnaire for IBS (IPQ), Irritable Bowel Syndrome Behavioural Responses Questionnaire, Beliefs about Emotions Scale (BES), impoverished emotional experience (IEE) factor of the Emotional Processing Scale-25 and Positive and Negative Affect Schedule (PANAS).

Primary outcome measures

The clinical effectiveness of the intervention was assessed by two co-primary measures: IBS SSS24 and WSAS. 25

The IBS SSS is widely used in IBS studies (a 50-point within-participant change from baseline is regarded as clinically significant). 24 We powered this trial to detect a 35-point difference between groups at 12 months for the sample size calculations. This was to account for a 15-point placebo response in the TAU arm (the placebo response is known to be important in IBS, and the MIBS trial1,15 showed a 24-point difference in the no-website group from baseline to 12-week follow-up; thus, allowing for a 15-point placebo response at 12 months was prudent). The IBS SSS24 is a five-item, self-administered questionnaire measuring severity of abdominal pain, duration of abdominal pain, abdominal distension/tightness, bowel habit and QoL. It has a maximum score of 500: a score of < 75 indicates normal bowel function, 75–174 indicates mild IBS, 175–299 indicates moderate IBS and 300–500 indicates severe IBS.

The WSAS measures the effect of the IBS on people’s ability to work and manage at home and to participate in social and private leisure activities and relationships. 25 WSAS has been shown to be sensitive to change in IBS trials. 7,20 It has five aspects, each scored from 0 (not affected) to 8 (severely affected), with a possible total score of 40.

Secondary outcome measures

The Subject’s Global Assessment (SGA) of Relief26 is frequently used in treatment trials to identify IBS responders for therapy. 26 Participants rate their relief from IBS symptoms on a scale of 1 to 5, ranging from ‘completely relieved’ to ‘worse’. Scores are dichotomised so that patients scoring 1–3 are considered responders and those scoring 4–5 are considered non-responders.

The HADS27 is a well-validated, commonly used self-report instrument for detecting depression and anxiety in patients with medical illnesses.

The Patient Enablement Questionnaire (PEQ)21 assesses participants’ ability to cope with their illness and life.

The CSRI28 and EQ-5D29 were used to gather information on use of health services and QoL.

Patients’ GP notes were reviewed at 12 months to assess GP and other consultations in the year prior to entering the study and in the 12 months since entry into the study. Other studies20,30 have shown an impact on GP contacts from patient self-management programmes.

Patients’ adherence to the CBT treatments was measured by recording the number of telephone sessions and an automated count of web sessions accessed. A patient completing four or more sessions of the website and one or more of the telephone support calls was deemed compliant with the WCBT arms. In the TCBT arm, a patient completing four or more of the initial telephone CBT sessions was deemed compliant. Patients kept a simple log of homework tasks to complete.

Process/mediator variables

Process and mediator variables were collected in this trial and will be used to inform a process evaluation that will be undertaken in a funded extension agreed by the HTA programme, and will be reported separately from this HTA report.

The measures collected were:

• The CG-FBD,31 a 31-item scale assessing unhelpful cognitions related to IBS.

• The IPQ,32 an 8-point scale to assess participants’ perceptions of their illness.

• The Irritable Bowel Syndrome-Behavioural Responses Questionnaire,33 a 26-item scale that measures changes in behaviour specific to managing IBS symptoms.

• The BES,34 a 12-item questionnaire that measures beliefs about the unacceptability of experiencing and expressing negative emotions. These beliefs are likely to have implications for emotion regulation and processing. Principal components analysis identified one factor and the scale had high internal consistency (0.91). 34

• The IEE factor of the Emotional Processing Scale,35 composed of five items and related to the labelling and awareness of emotional events, which influence the way people process their emotions. The subscale has high internal consistency (0.82). 35

• The PANAS36 measures both positive and negative affect. The reliabilities of the PANAS, as measured by Cronbach’s alpha, were 0.89 for positive affect and 0.85 for negative affect. 37 Participants completed only the positive affect subscale because the HADS measures negative affect.

Adverse event reporting

An AE was defined as any clinical change, disease or disorder experienced by the participant during their participation in the trial, whether or not it was considered related to the intervention. A SAE was defined as an AE that was life-threatening or that resulted in inpatient hospitalisation, a disability/incapacity, a congenital anomaly/birth defect in the offspring of a subject, or another medical event requiring intervention to prevent one of these outcomes. All sites and therapists were supplied with SAE forms (see Report Supplementary Material 9) to complete and a SAE standard operating procedure (SOP).

In addition, patients were asked to self-report any medical events at each assessment.

Patients were asked the following online questions at the 3-, 6- and 12-month follow-up points:

• Since you started the study have you had any of the following events – a life-threatening event, admission to hospital where you had to stay overnight, permanent disability/incapacity, a congenital anomaly/birth defect in a child of yours, or other medical events requiring medical attention to prevent one of the above? If yes, please give details.

• Has your health been adversely affected since the start of the study? If yes, please give details.

The chief investigator, on behalf of the sponsor, assessed each completed SAE form for relatedness to the intervention and expectedness, to identify serious adverse reactions (SARs) and suspected unexpected serious adverse reactions. The co-ordinating centres’ SOPs were followed with respect to reporting to the sponsor, Research Ethics Committee (REC), Data Monitoring and Ethics Committee (DMEC), Trial Steering Committee (TSC) and local governance offices. Annual safety reports were submitted to the REC. All AEs and SAEs were entered onto MACRO.

Economic evaluation measures

These are discussed in Chapter 3.

Follow-up procedures

Participants completed follow-up measures online at 3, 6 and 12 months after baseline using the LifeGuide website (see Appendix 5). Those who were unable to complete the outcome measures online received a paper copy of the questionnaires. If this was not completed, then they received a telephone call from a blinded researcher, who took the participant through a limited selection of the main outcome measures. These were the IBS SSS, WSAS, SGA, PEQ and HADS.

The follow-up procedure was as follows: the baseline assessment was conducted within 90 days of screening (to allow time for the blood tests) and randomisation was conducted within 3 days of baseline (to allow for weekends). Patients received two automated reminders from LifeGuide to complete their follow-up assessments at 3, 6 and 12 months. In addition, the researchers received an automated message from LifeGuide after 3 weeks if the questionnaires had not been completed, which prompted the team to send a personalised e-mail to the patient. This was followed by a further text, a paper copy of the full questionnaire sent to the patient’s home address and then, if still not complete, a telephone call from a blinded member of the research team to complete a short questionnaire.

All baseline and most 3-, 6- and 12-month outcome data were self-completed by the patient on a data collection section of the Regul8 programme website (LifeGuide). LifeGuide was maintained and hosted by the University of Southampton. The research team was notified when LifeGuide data were missing, and actions were taken to collect these data via paper questionnaires in the post and by telephone. Primary outcome data and some secondary outcome data that were collected by paper questionnaires or short telephone questionnaires were entered by the research team onto MACRO; these included PEQ, IBS SSS, WSAS, SGA and HADS. Other paper-collected outcome measures, including CSRI, IPQ, Behavioural Responses Questionnaire, IEE-EPS (Emotional Processing Scale), BES, PANAS and EQ-5D, were entered onto a spreadsheet. In addition, all questionnaire completion dates and sources of data were entered onto MACRO, as were economic data, AEs and SAEs and withdrawals.

Error checking was carried out in the full questionnaire database of the ACTIB trial. This was to check that the error rates for the primary outcomes were < 1% and the error rates for the secondary outcomes were < 5%; 20% checks were carried out on all variables for the 3-, 6- and 12-month questionnaires. Data checking was carried out by the London data entry research assistants, with Stephanie Hughes, Alice Sibelli and Sula Windgassen ensuring that the checks were carried out by a different person from the original data entry:

• 3 months – 89 questionnaires, every fifth ID selected = 18 patients

• 6 months – 103 questionnaires, every fifth ID selected = 21 patients

• 12 months – 139 questionnaires, every fifth ID selected = 28 patients.

A spreadsheet of any errors and changes needed was kept, listing variable name, original data entry, change, date, initial and reason.

Patient retention

In late 2014, follow-up monitoring revealed that follow-up rates were not as high as the 80% anticipated in the original sample size calculation. Follow-up procedures were very carefully scrutinised and every effort was made to increase rates using the follow-up methods described in the previous section. In December 2015, the REC approved an increase in the number of participants to be recruited, with a new target of up to 570 participants [as per an updated sample size calculation to allow for the lower follow-up rates (see Sample size)]. The REC also approved sending vouchers to participants, as evidence from the literature38 suggested that non-conditional vouchers could help to incentivise patients to complete follow-up questionnaires. Cards with an enclosed £10 Love2shop voucher were posted to all participants before their 3-month questionnaire date (and immediately for any patients who had passed this date).

Sample size

A 35-point difference between therapy groups and TAU on IBS SSS at 12 months was regarded as the minimal clinically important difference (MCID) (assuming a 15-point placebo response in the TAU arm in the trial). 15,24 Assuming a within-group IBS SSS standard deviation (SD) of 76 (taken from the MIBS pilot study1), this equated to an effect size of 0.46. To achieve 90% power to detect such an effect or larger using a two-sided independent-samples t-test at the 2.5% significance level (adjusting for two primary outcomes), it was estimated that the trial would require 119 subjects per group. Based on each of 10 therapists delivering therapy to 17 patients in the WCBT and TCBT groups, and an intraclass correlation of 0.02, taken from Baldwin et al.,39 this sample size was increased by an inflation factor of 1.32 to take account of therapist effects. We measured IBS SSS at baseline and assumed that baseline values were predictive of post-treatment values (correlation 0.4). Accounting for this in our statistical analysis model allowed us to decrease the sample size by a deflation factor of 0.84. Finally, we applied a further inflation factor of 1.25 on the assumption that attrition would be < 20%. The final sample size requirement was thus calculated as 165 patients per group or 495 patients in total.

In terms of our second primary outcome (WSAS), the initial 495 sample size was calculated to be sufficient to detect a difference between the WCBT (or TCBT) and TAU groups of ≥ 3.7. Specifically, we assumed inflation factors of 1.32 for correlation of outcomes within therapists and of 1.25 for attrition and a deflation factor of 0.84 for correlation between baseline and follow-up measures. Based on this, a moderate effect size of 0.46 could be found with 90% power at the 2.5% significance level, given 119 participants per group. Assuming a SD of 8.0 (as estimated in a study of CBT for IBS7), this would equate to a treatment difference of 3.7 on this scale. This is less than the difference of 5.4 in change of means in WSAS that was found in a trial of a CBT-based self-management intervention for IBS. 20

As the trial progressed, we found that the attrition rate was closer to 30% (November 2014 estimate). The sample size was recalculated using the same group size of 119 subjects, with inflation and deflation factors of 1.32 and 0.84 kept constant. The updated attrition rate of 30% gave a sample size of 189 patients per group and a total of 567 patients. We gained ethics approval to increase recruitment to the trial within the same recruitment time frame to aim for this larger calculated sample size.

Statistical methods

A statistical analysis plan (SAP) (see Appendix 6) was developed by the statisticians (KG, RH and SL), discussed by the trial management team and the DMEC and approved by the chief investigator (HE) and chairperson of the TSC before database lock. The following approach was used to formally compare the primary and secondary outcomes between a CBT arm and TAU: an intention-to-treat (ITT) approach was used for all primary and secondary outcomes, that is, participants were analysed in the groups to which they were randomised. For each outcome and assessment time point (3, 6 or 12 months), we estimated the effect of treatment (TCBT or WCBT) compared with TAU to assess treatment effectiveness. The TCBT and WCBT arms were not formally compared.

As we had two primary outcomes (IBS SSS and WSAS at 12 months), significance testing for these two variables was conducted at the Bonferroni-adjusted significance level of alpha = 0.025 to account for two outcome comparisons. Furthermore, because we carried out two planned comparisons per outcome measure, further adjustments were not necessary.

The primary outcome measures and the secondary outcome measure, HADS score, were continuous variables. Their modelling relied on normal assumptions for error terms and treatment effects were quantified by trial arm differences (and standardised differences). We had planned to analyse PEQ under a normal assumption but found inflated floor and ceiling effects for this variable. To facilitate modelling, the original PEQ was reclassified as a binary responder measure (0 = non-response, 1 = response). SGA measures were also reclassified as binary responders, as originally described in the SAP. A ‘PEQ responder’ was defined as a participant achieving a score of ≥ 6 at the post-randomisation time point. A ‘SGA responder’ was defined as a participant achieving a score of between 1 and 3, as planned in the SAP. Binary outcome variables were analysed within a logistic regression framework and treatment effects quantified by odds ratios (ORs).

Software

Recruitment

Sites were entered into the study in stages to ensure a continuous and steady feed of patients into the available therapy slots. The first sites to participate were given the instruction to search databases in February 2014 and the first invitation letters were sent in March 2014. The patient recruitment rate was monitored and subsequent sites were started each month over the following 23 months.

A total of 15,065 invitations were given out by 77 sites (74 general practice surgeries and three secondary care centres). There were 14,908 invitation letters (98.96%) delivered by mail-outs following a note search of potential participants. Forty-four invitation letters (0.29%) were sent after potential participants saw displayed posters and contacted the research teams, and 113 invitation letters (0.75%) were handed out by consultants or GPs during GI consultations as opportunistic recruitment (Table 3).

Summary of recruitment

Table 12 and Figure 1 show recruitment by centre and over time.

TABLE 12 - Details of patient recruitment pathway

LPC, London primary care; SPC, Southampton primary care; UHS, University Hospital Southampton NHS Foundation Trust.

Pathway	Site (n)					All centres (n)
	GSTT	KCH	LPC sites	SPC sites	UHS
Sites	1	1	28	46	1	77
Mailout	203	208	5482	8841	174	14,908
Poster	0	0	2	40	2	44
Opportunistic	80	0	7	25	1	113
Patient invitations	283	208	5491	8906	177	15,065
Invitations
‘Yes’	100	44	414	938	29	1525
‘Declined’	18	28	633	1719	25	2423
Screened	96	41	370	916	29	1452
Randomised	59	19	141	324	15	558

FIGURE 1.

Participant recruitment against target.

Consolidating Standards of Reporting Trials

A CONSORT flow chart has been constructed. This includes the number of potential patients contacted and screened, the number of eligible patients, the number of patients agreeing to enter the trial and the number of patients refusing to enter the trial, and then, by treatment arm, the number of patients adherent to treatment, the number of patients continuing through the trial, the number of patients withdrawing, the number of patients lost to follow-up and the number of patients excluded/analysed.

Treatment adherence was defined separately for the two active treatment arms. Participants allocated to TCBT who completed at least four of the initial telephone calls were deemed adherent to treatment. Those who were offered WCBT and completed four or more website sessions and at least one telephone support call were considered as treatment adherent.

Details of the patient pathway through the recruitment process are included in the CONSORT flow diagram (Figure 2), with details of the inclusion and exclusion criteria given in Table 13.

FIGURE 2.

The CONSORT flow diagram. 42

TABLE 13 - Ineligibility after screening details (CONSORT flow diagram detail)

Ineligibility criteria	Number of participants
Failed inclusion criteria	547
Not aged ≥ 18 years	2
Does not have refractory IBS (clinically significant symptoms defined by a IBS SSS of > 75)	54
Does not fulfil Rome III criteria	43
Has not been offered first-line therapies (e.g. antispasmodics, antidepressants or fibre-based medications) and does not have continuing IBS symptoms for ≥ 12 months	297
If > 60 years old, patient has not had a consultant review in the in the previous 2 years	151
Exclusion criteria	76
Unexplained rectal bleeding or weight loss	5
Diagnosis of IBD	8
Diagnosis of coeliac disease	2
Diagnosis of peptic ulcer disease	0
Diagnosis of colorectal carcinoma	0
Unable to participate in CBT because of speech or language difficulties	4
No access to an internet computer to be able to undertake the lower-intensity WCBT	53
Received CBT for IBS in the past 2 years	3
Had previous access to the MIBS website	1
Currently participating in an IBS/intervention trial	0

Baseline data summary

A total of 558 (38.4%) out of 1452 patients screened for eligibility were randomised, with 186 randomly allocated to TCBT, 185 to WCBT and 187 to TAU.

Table 14 shows that, as expected from randomisation, the three trial arms were well balanced with regard to clinical and demographic variables. In particular, the distribution of recruitment centre (stratification factor) was almost identical across trial arms.

Participants had a mean age of 43.1 years at baseline and the majority were female (75.8%). Over 90% of participants were white, 64.7% of participants were married or living with a partner and 35.3% were single, separated, divorced or widowed. The median duration since diagnosis was 7.4 years and the median duration of symptoms prior to diagnosis was 3.0 years. At baseline, only 10.2% of participants had seen a consultant. Approximately 30% of participants had received treatment for anxiety and nearly 40% had received treatment for depression. In accordance with the HADS, approximately 60% of participants were not distressed at baseline.

A very small proportion, approximately 1%, of participants had joined a self-help organisation. Most of those who had done so had joined an online group.

Participants were asked their highest level of formal education. The largest single group (54.6% of participants) were those with at least a degree. Approximately 20% of participants had only General Certificate of Secondary Education (GCSE) or equivalent-level qualifications and a similar proportion had postgraduate qualifications. The majority of participants, 68.6%, owned their own property, and 27.5% lived in rented accommodation, of whom only 6.3% rented their accommodation from a local authority. In addition, 14.9% of our sample lived alone, 41.4% lived with their spouse only and a further 23.3% lived with their spouse and children. The mean deprivation (IMD) score was 17.1 (range 0.8–54.0).

At screening, the mean IBS SSS was 292.3 (range 75–500) but at baseline the mean IBS SSS was 265.0 [range 14–500 (see Table 26)].

At baseline, participants had a mean HADS anxiety score of 10.7 (range 1–21), a mean HADS distress score of 16.4 (range 1–39), a mean HADS depression score of 5.7 (range 0–19) and a mean WSAS score of 12.5 (range 0–40) (see Table 26).

The majority of participants, 74.2%, would have preferred one of the CBT groups, with 36.2% preferring the online CBT. Only 25.8% expressed no preference or preferred the standard care.

Therapy summary

All therapists were available to work in both therapy arms (TCBT and WCBT) and with any participant regardless of recruitment centre. There were 13 therapists, of whom 10 (77%) were female. Six therapists were clinical psychologists (46%) and seven were CBT psychotherapists (54%). They had worked for a median of 7 years in their professions (minimum 4 years, maximum 24 years), with clinical psychologists having worked for a median of 7.5 years in their profession (minimum 4 years, maximum 12 years) and CBT psychotherapists for a median of 5 years (minimum 4 years, maximum 24 years). They had worked within medically unexplained symptoms for a median of 4 years (minimum 1 year, maximum 20 years). Their mean age was 42 years (SD 5.9 years), with a minimum age of 34 years and a maximum age of 52 years. Table 15 shows the number of participants allocated to each therapist.

Withdrawal from treatment

Forty-five participants withdrew from treatment at some point during the trial (Table 20). More than half of those withdrawing from treatment also withdrew from follow-up and did not supply any further outcome data after ceasing therapy sessions.

Compliance with treatment

For those in the TCBT arm, treatment compliance was defined as taking part in four telephone sessions out of the six initial sessions offered. Approximately 84% of participants completed at least four telephone calls; 16% did not comply. The take-up of therapy and booster sessions offered is described in Table 21.

For those in the WCBT arm, treatment compliance was defined as taking part in one telephone session out of three sessions offered and accessing at least four web sessions. Approximately 88.1% of participants allocated to the WCBT arm completed at least one telephone call, and 69.8% completed four web sessions (Table 22).

A larger percentage of participants failed to comply with WCBT (30.8%) than with TCBT (16%).

Follow-up summary

Follow-up assessments were scheduled at 3, 6 and 12 months after randomisation. Follow-up rates in terms of primary outcomes were good, with 136, 124 and 131 participants completing both final outcome measures in the TCBT, WCBT and TAU arms, respectively (see Figure 2). These numbers constitute follow-up rates at 12 months of 73% for TCBT, 67% for WCBT and 70% for TAU. Table 23 shows how many of the measures were recorded within the prespecified time windows (ranging from 1 week before the intended time point until 4 weeks after). At 12 months, approximately 75% of recorded outcomes were obtained within the prescribed window.

The overall follow-up rates were 76.5% at 3 months (427/558), 72.9% at 6 months (407/558) and 70.3% at 12 months (392/558) (see Figure 2 for a summary of patient follow-up).

We tested whether or not compliance with treatment offered was predictive of missing primary outcomes at 12 months using Fisher’s exact test. The test found that non-compliance with treatment offered predicted missing outcomes in both the TCBT (p < 0.001) and the WCBT (p < 0.001) arms. Logistic regression tests found that higher baseline IBS SSS (p < 0.001) and higher IMD scores (p < 0.001) were associated with more missing values in the primary outcomes at 12 months. Thus, all of these variables were included as predictors of missing values in the imputation step of our MI approach (for details see Methods). All results reported here are based on MI with k = 100 imputations to avoid missing data biases.

Outcome measures

Outcome measures are summarised by assessment time point and trial arm in Tables 24–26, and the results of the formal trial arm comparisons are provided in Tables 27 and 28.

Table 24 shows that improvements over time were observed for both primary outcomes (IBS SSS and WSAS) and in all trial arms. Table 26 illustrates IBS SSS changes in terms of severity categories and also shows that the percentage of participants with no or only mild symptoms increases after CBT treatment in both the TCBT and the WCBT arms, and also somewhat in the TAU arm.

Primary outcome measures

Table 27 shows the results of the formal statistical analyses of the continuous outcomes.

Irritable Bowel Syndrome Symptom Severity Score

The estimated IBS SSS differences for TCBT versus TAU were –69.2, –58.3 and –61.6 at 3, 6 and 12 months, respectively. For WCBT versus TAU, the differences were lower, at –53.0, –35.7 and –35.2, respectively (Figure 3). The comparison of the CBT arms with the TAU arm at all times showed statistically significant benefits in terms of IBS SSS. At 12 months, moderate-sized effects were found for TCBT (p < 0.001, standardised effect = 0.65) and WCBT (p = 0.002, standardised effect = 0.37) (see Table 29). The estimated differences between the TCBT arm (a reduction of 61.6) or the WCBT arm (a reduction of 35.2) and the TAU arm were larger than our stated minimum clinically significant difference (MCID = 35, as per our sample size calculation; see Sample size).

FIGURE 3.

The IBS SSS means by trial arm. IBS SSS possible range is 0–500.

Therapist effects were only detected and modelled in the TCBT arm at 12 months.

The number needed to treat (NNT) to improve the IBS SSS for one case by 50 at 12 months was 3.5 for TCBT and 4.6 for WCBT.

Work and Social Adjustment Scale

The estimated WSAS score differences for TCBT versus TAU were –3.4, –2.7 and –3.5 at 3, 6 and 12 months, respectively. For WCBT versus TAU, the differences were –3.0, –2.5 and –3.0, respectively. The comparison of the CBT arms with the TAU arm at all time points showed statistically significant benefits for both TCBT and WCBT. At 12 months, small to moderate effects were found for TCBT (p < 0.001, standardised effect = 0.4) and WCBT (p < 0.001, standardised effect = 0.3) (Figure 4). Therapist effects were not detected for any time point or CBT treatment arm.

FIGURE 4.

The WSAS means by trial arm. WSAS possible range is 0–40.

The NNT to improve the WSAS score for one case by 5.9 at 12 months was 5 for both TCBT and WCBT.

Secondary outcome measures

Descriptive summaries for the two binary secondary outcomes, PEQ responder and SGA responder, are provided in Table 26. The responder percentage increased over time for both responder types and in all trial arms. The results of the formal trial arm comparisons are shown in Table 28.

TABLE 28 - Estimated trial arm ORs for binary outcomes (ratios above 1 indicate an improvement)

The 12-month model included therapist effects in the TCBT arm. Thus, these effects are conditioned on therapist.

The corresponding (lower) marginal ORs within site are 9.0 (Southampton general practices), 9.2 (Southampton secondary care), 9.3 (London general practices) and 9.3 (London secondary care).

The corresponding (lower) marginal ORs within site are 5.8 (Southampton general practices), 6.0 (Southampton secondary care), 5.9 (London general practices) and 5.9 (London secondary care).

Note

All inferences were derived by MI as described in Methods. Each model used k = 100 imputations.

Responder	TCBT vs. TAU		WCBT vs. TAU
	Estimated OR (95% CI)	Test (degrees of freedom); p-value	Estimated OR (95% CI)	Test (degrees of freedom); p-value
PEQ responders
3 months	13.3 (6.9 to 25.8)	t = 7.7 (1347); p < 0.001	6.6 (3.3 to 13.2)	t = 5.4 (918); p < 0.001
6 months	11.6 (6.4 to 20.9)	t = 8.1 (1134); p < 0.001	4.8 (2.7 to 8.7)	t = 5.2 (889); p < 0.001
12 monthsa	9.3b (4.5 to 19.3)	t = 6.0 (2369); p < 0.001	3.5 (2.0 to 5.9)	t = 4.5 (1079); p < 0.001
SGA responders
3 months	10.2 (5.8 to 18.1)	t = 8.1 (1289); p < 0.001	6.4 (3.6 to 11.3)	t = 6.3 (824); p < 0.001
6 months	6.2 (3.7 to 10.4)	t = 6.9 (1637); p < 0.001	5.0 (2.8 to 8.8)	t = 5.6 (630); p < 0.001
12 monthsa	6.1c (2.5 to 15.0)	t = 3.9 (2957); p < 0.001	3.6 (2.0 to 6.3)	t = 4.4 (600); p < 0.001

Patient enablement

For the TCBT arm, the odds of being a responder on the PEQ (defined as a score of ≥ 6) were 13.3, 11.6 and 9.3 (3, 6 and 12 months, respectively) times those in the TAU arm. The ORs for WCBT compared with TAU were lower, at 6.6, 4.8 and 3.5 at 3, 6 and 12 months, respectively, but still indicated large improvements in responder rates (Figure 5). These results were statistically significant for both CBT arms and at all time points, with p-values of < 0.001 at all times. Therapist effects were detected only in the TCBT arm at 12 months.

FIGURE 5.

The PEQ responders by trial arm.

Subject’s Global Assessment of Relief

For SGA responders, the ORs were 10.2, 6.2 and 6.1 for TCBT and 6.4, 5.0 and 3.6 for WCBT (3, 6 and 12 months, respectively) (Figure 6). These results were also statistically significant for both CBT arms and at all time points. Therapist effects were again detected only in the TCBT arm at 12 months.

FIGURE 6.

The SGA responders by trial arm.

As mentioned above, therapist effects were detected at 12 months for the primary outcome, IBS SSS, and the secondary outcomes, SGA and PEQ. Such effects were seen only in the TCBT arm. To understand these effects, we tabulated the outcome improvements by 12 months by therapist in the TCBT arm. Table 29 indicates that this therapist variability was largely due to lower responses on all three measures in the patient groups that were seen by therapist T2 or tended not have a therapist assigned at all. Participants were randomised only to treatment group, not to therapist. It cannot be assumed that the groups seen by an individual therapist were equivalent.

TABLE 29 - Response (over time) in the TCBT arm, by therapist

This set of participants was either assigned a therapist who saw very few participants or not assigned a therapist at all.

Therapist	Number assigned	Number of participants completing outcomes	Mean baseline IBS SSS (SD)	IBS SSS mean difference (SD)	Responders, n (%)
					PEQ	SGA
T1	46	38	288.3 (13.3)	–156.8 (15.2)	35 (92.1)	36 (94.7)
T2	9	9	268.3 (48.7)	–37.8 (31.9)	6 (66.7)	6 (66.7)
T3	24	15	279.3 (23.8)	–133.1 (36.1)	13 (86.7)	13 (86.7)
T4	22	17	284.3 (14.7)	–132.6 (20.1)	15 (88.2)	14 (82.4)
T5	11	9	266.5 (19.6)	–140.2 (23.6)	7 (77.8)	9 (100.0)
T6	13	9	278.5 (30.7)	–116.4 (30.1)	6 (66.7)	9 (100.0)
T7	8	5	247.0 (39.6)	–108.6 (42.0)	3 (60.0)	4 (80.0)
T8	36	29	253.3 (15.1)	–84.6 (18.5)	19 (65.5)	21 (72.4)
Mixeda	17	7	256.9 (22.1)	–69.9 (48.5)	4 (57.1)	5 (71.4)
Total	186	138	272.3 (7.0)	–119.2 (8.8)	108 (78.3)	117 (84.8)

Hospital Anxiety and Depression Scale

For HADS distress, the estimated differences for TCBT or WCBT compared with TAU were found to be –2.1 for TCBT to TAU at 3 months, –2.2 for TCBT to TAU at 6 months and –2.8 for TCBT to TAU at 12 months, and –2.5 for WCBT to TAU at 3 months, –2.9 for WCBT to TAU at 6 months and –2.3 for WCBT to TAU at 12 months (Figure 7 and see Table 27). These results were statistically significant at all time points and in both CBT arms (Table 29). The effect sizes for HADS, once standardised, were small to moderate. Therapist effects were not detected for any time point or CBT treatment arm.

FIGURE 7.

The HADS means by trial arm. HADS possible range is 1–39.

General practitioner visits

The number of GP visits recorded in the GP records in the year before and after randomisation were recorded on a notes review form (see Appendix 7). The numbers of visits did not noticeably change and were similar across the three trial arms. The mean number of visits for all arms was between four and five per year both pre and post randomisation (see Appendix 8, Table 41).

Sensitivity analysis

Four sets of sensitivity analyses were conducted. The first sensitivity analysis assessed the impact of excluding participants who had an IBS SSS at baseline below the inclusion threshold of 75 from the analysis set. (The IBS SSS eligibility criterion was determined at screening.) The second sensitivity analysis looked at the impact of using only observations that were recorded within the prespecified assessment time windows. The third sensitivity analysis evaluated the impact of defining PEQ responders according to an alternative threshold. Finally, the fourth sensitivity analysis explicitly targeted the efficacy rather than the effectiveness of the CBT treatments and assessed the impact of this on effect size estimates for the two co-primaries.

Sixteen participants had scores that were lower than the eligibility threshold when this was reassessed at baseline. In the TCBT arm, three participants were below 75 on the IBS SSS at baseline; in the WCBT arm, eight participants were below this threshold; in the TAU group, five participants were below this threshold. These participants were dropped from the analysis and the co-primaries were analysed in the same way as before. For IBS SSS at 12 months, the estimated trial arm differences were –63.0 (95% CI –91.9 to –34.2; p < 0.001) for TCBT compared with TAU and –35.5 (95% CI –58.2 to –12.7; p = 0.002) for WCBT compared with TAU. For WSAS at 12 months, the estimated trial arm differences were –3.5 (95% CI –6.1 to –0.9; p = 0.02) for TCBT compared with TAU and –3.4 (95% CI –4.7 to –2.1; p < 0.001) for WCBT compared with TAU. Comparisons of these findings with the original analysis results shown in Table 29 demonstrate that the analyses were robust regarding the timing of the IBS SSS eligibility assessment.

Participants were asked to record outcomes at 3, 6 and 12 months after randomisation within the prespecified time window, ranging from 7 days before the intended assessment date to 4 weeks afterwards. Treatment effects were expected to be reasonably constant over the 5-week period. For primary outcomes at the 12-month assessment time point, 98 out of 391 (25.0%) IBS SSS values and 98 out of 394 (24.9%) WSAS values were recorded outside this assessment window (see Table 25). These participants were dropped from the analysis and the co-primaries were analysed in the same way as before. For IBS SSS at 12 months, the estimated trial arm differences were –64.3 (95% CI –93.2 to –35.5; p < 0.001) for TCBT compared with TAU and –37.2 (95% CI –64.1 to –10.4; p = 0.007) for WCBT compared with TAU. For WSAS at 12 months, the estimated trial arm differences were –3.3 (95% CI –4.9 to –1.8; p < 0.001) for TCBT compared with TAU and –3.3 (95% CI –3.8 to 10.4; p = 0.20) for WCBT compared with TAU. The results demonstrate that, apart from some loss of power (wider CIs), the IBS SSS analysis was not sensitive to outcomes being recorded within the specified time period. However, for the WSAS, power loss was associated with dropping one-quarter of the sample, leading to the WCBT effect becoming non-significant (see Table 29).

As we had not anticipated analysing PEQ scores on a binary scale, we confirmed that the substantive analysis results were not sensitive to the choice of threshold for defining a participant to be a ‘PEQ responder’. In the main analysis, we considered those reporting a PEQ score of ≥ 6 to be responders. For this sensitivity analysis, we considered responders to be those reporting a PEQ score of ≥ 4. The ORs of being a responder reported at the primary outcome time of 12 months using these criteria were estimated to be 12.7 (conditioned on therapist and stratification site, 95% CI 4.9 to 33.0; p < 0.001) for TCBT versus TAU. For the comparison of WCBT with TAU, which was not affected by therapist effects, the OR was estimated to be 4.0 (95% CI 2.3 to 6.9; p < 0.001). Comparing these findings with those reported in Table 30 shows that the definition of a ‘PEQ responder’ does not affect the substantive conclusions.

An instrumental variables analysis was carried out to assess the efficacy of TCBT and WCBT, based on complete cases. The 12-month efficacy (quantified by CACE) for IBS SSS was estimated to be 71.8 (95% CI 93.7 to 59.9; p < 0.001) for TCBT compared with TAU and 50.4 (95% CI 75.2 to 25.5; p < 0.001) for WCBT compared with TAU. For WSAS, the estimated efficacy was –4.5 (95% CI –6.1 to –3.0; p < 0.001) for TCBT compared with TAU and –4.2 (95% CI –5.9 to –2.4 points; p < 0.001) for WCBT compared with TAU. This suggests that the intervention might be more efficacious (the effects of actually receiving the intervention stronger) than the effects suggested by the ITT analyses.

Adverse events

There were 77 recorded AEs in the TCBT arm, 61 in the WCBT arm and 55 in the TAU arm (Table 30). We would expect to see an increased rate of reported events in the CBT arms, particularly in the TCBT arm, as the therapists completed AE forms for any AEs that were mentioned during the therapy sessions.

Of these, 18.1% were gastrointestinal related (see Table 11). Fourteen of these events were seen in the TCBT arm, 10 in the WCBT arm and 11 in the TAU arm. Of the AEs, 23.3% were psychological events, of which 18 were seen in the TCBT arm, with 17 and 10 events seen in the WCBT and TAU arms, respectively. The other 14 body system codes made up 58.6% of AEs, most individual system codes accounting for < 10% of events.

There were four GI AEs in the TAU arm, one psychological event in the TCBT arm and no events related to the treatment in the WCBT arm that were recorded prior to unblinding as possibly related to the intervention (Table 31). There were two events in each of the TAU, WCBT and TCBT arms for which, prior to unblinding, it was recorded as unknown whether or not they were related to the intervention.

Protocol deviations

Therapist-related protocol deviations are outlined in Appendix 9.

A full list of protocol deviations can be found in Appendix 9.

Summary of results

The ACTIB trial was a RCT of TCBT with a patient manual and web-delivered CBT with minimal therapist input (WCBT), both compared with TAU in adults with refractory IBS with a 12-month follow-up. Both therapy arms received TAU as well as CBT. A total of 558 out of 1452 (38.4%) of patients screened for eligibility were recruited: 186 were randomised to TCBT, 185 were randomised to WCBT and 186 were randomised to TAU. The mean IBS SSS was 265.0 at baseline, indicating moderately severe IBS symptoms. The trial arms were well balanced at baseline.

Follow-up rates were 76.5% at 3 months (427/558), 72.9% at 6 months (407/558) and 70.3% at 12 months (392/558). ITT analysis was undertaken with MI.

Compared with the TAU arm, both the TCBT and WCBT intervention arms showed a significant improvement in primary (IBS SSS and WSAS) and secondary outcomes over the trial period. In terms of the two primary outcomes, at 12 months, compared with the TAU arm (IBS SSS of 205.6), IBS SSS were 61.6 lower for the TCBT arm (95% CI 89.5 to 33.8; p < 0.001) and 35.2 lower for the WCBT arm (95% CI 57.8 to 12.6; p = 0.002). The WSAS score in the TAU arm was 10.8 (SD 9.3) at 12 months, 3.5 lower with TCBT (95% CI 5.1 to 1.9; p < 0.001) and 3.0 lower with WCBT (95% CI 4.6 to 1.3; p = 0.001). Thus, both primary outcomes were clinically and statistically significant at 12 months.

In terms of secondary outcomes, SGA of relief of symptoms had a 84.8% response rate in the TCBT arm at 12 months compared with 41.7% in the TAU arm (OR 6.1, 95% CI 2.5 to 15.0; p < 0.001) and 75.0% in the WCBT arm (OR 3.6, 95% CI 2.0 to 6.3; p < 0.001). PEQ had a 78.3% response rate in the TCBT arm compared with 23.5% in the TAU arm (OR 9.3, 95% CI 4.5 to 19.3; p < 0.001) and 54.8% in the WCBT arm (OR 3.5, 95% CI 2.0 to 5.9; p < 0.001). For HADS, compared with the TAU arm [mean HADS score of 16.4 (SD 6.9) at 12 months], scores were 2.8 (95% CI 4.1 to 1.5) lower (p < 0.001) in the TCBT arm and 2.3 (95% CI 3.7 to 1.0) lower (p < 0.001) in the WCBT arm at 12 months.

Strengths and limitations

Interpretation of ACTIB results

Objectives of the qualitative work

The objectives were to identify the factors that facilitate or impede adherence to web-delivered and TCBT in patients with refractory IBS, to provide insight into the quantitative results of the ACTIB trial and to identify social and psychological processes of change that took place during the trial.

Background

Face-to-face CBT has been shown to improve IBS symptom severity and QoL. However, its availability in primary care is limited and its cost-effectiveness is uncertain. Furthermore, it has been associated with high dropout rates. Both telephone- and web-based CBT are becoming widely used psychological interventions to treat a diverse range of medical and mental health conditions. Web-based CBT affords individuals the opportunity to engage in a therapy by overcoming some of the known barriers to traditional face-to-face CBT.

Methods

Semistructured interviews were undertaken with 52 ACTIB trial participants at 3 months and 42 interviews were undertaken at 12 months post baseline (see Appendix 10 for interview schedules). Inductive thematic analysis was used.

Findings

Key themes related to the perceived benefits of CBT, the role of the therapist and the processes by which CBT may have elicited benefits. Perceived benefits of CBT include reduced symptoms and, arguably more importantly, an increased capacity to cope with symptoms, negative emotions and other challenges of daily life. Therapists have an important role to play in supporting patients to engage with CBT and to make sense of the therapy and their IBS. Patients valued having therapist support available alongside Regul8 and this may have helped enhance their engagement, adherence and outcomes. The perceived high quality of the intervention itself also facilitated adherence. CBT appeared to help patients make changes to their ways of thinking and behaving in relation to their IBS, broadly consistent with the theoretical mechanisms of change underpinning this intervention.

Conclusion

Telephone- and web-based CBT for IBS appears to help patients change how they think and feel about IBS, and can produce benefits that go beyond symptom reduction to more fundamental changes in coping and self-management strategies. If rolled out in practice, such interventions would benefit from offering therapist support alongside any web-based intervention, could be offered to patients despite initial scepticism regarding psychological interventions and could be augmented with longer-term support for maintenance.

Design

This qualitative study was nested within the main ACTIB trial. Semistructured interviews were subjected to thematic analysis, incorporating techniques from grounded theory and framework analysis. Data collection and initial analysis proceeded iteratively, with coding beginning after completing the first few interviews and informing subsequent interviews. The analysis was data driven (i.e. inductive) in that the researchers sought to identify themes in the data rather than imposing any pre-existing interpretive framework. Subsequently, inductively identified themes were compared across subgroups, mapping the qualitative findings against prespecified quantitative data. This approach enabled the strengths of qualitative data and inductive analysis to be realised while also relating these findings to their context, that is, the main trial.

Ethics consideration

At the beginning of the trial, all trial participants completed an online consent form (see Appendix 2) and agreed to be contacted to discuss their involvement in a qualitative interview at 3 months post baseline (i.e. post treatment) and at 12 months post baseline. On subsequently being invited to take part in interviews, the purpose and scope of the interviews were reiterated. Before commencing each interview, interviewers obtained and recorded oral consent. All interviews were anonymised on transcription; pseudonyms and/or participant numbers have been used throughout the reporting of this study, and to minimise the risk of participant identification, combinations of multiple participant characteristics are not reported at an individual participant level.

The qualitative study was included in the main ethics application approved by the NRES Committee South Central – Berkshire (REC reference 13/SC/0206) on 11 June 2013. The interview topic guides were approved by the same committee on 4 February 2014.

Sampling and recruitment

The aim was to interview a diverse sample of approximately 17–20 interviewees per arm at baseline (i.e. 10–12%). Interviewing participants from each active arm allows factors to be identified that relate to adherence and change processes; including participants from the TAU arm provides insight into the quantitative results. Interviewing the same individuals at 3 and 12 months allows greater depth to explore change processes over time and the potential to better understand any differences in the quantitative results between 3 and 12 months. Interviewees were sampled purposively to encompass variety in gender, age, ethnic background, baseline symptom severity scores, recruiting site (London/Southampton) and setting (primary/secondary care). Sampling for variety on such key characteristics helps ensure that the qualitative findings capture the breadth of participants’ experiences and are not dominated by any one subgroup.

Participants were invited in batches to permit iterative sampling, interviewing and analysis. The researchers attempted to contact participants via two e-mails, one text message and two telephone calls, after which no further attempts were made. In total, 100 trial participants were contacted in this way. Recruitment to interviews ended when no new themes emerged and existing themes were well developed within a diverse sample. The ACTIB trial recruited a total of 558 participants, 52 of whom were interviewed.

Interviews

Semistructured interviews were used to elicit interviewees’ experiences of taking part in the trial and IBS more broadly. Separate topic guides for the 3- and 12-month interviews (see Appendix 10) were developed collaboratively by the research team and refined after being piloted with two people with IBS. The topic guides comprised a series of open-ended questions and prompts used by the interviewer to elicit participants’ experiences of, reflections on and thoughts and feelings about the trial within the broader context of managing IBS. The questions were open ended and designed to elicit concrete descriptions of interviewees’ experiences and reflections thereon. The questions were grouped into four sections: taking part in the trial and trial treatments, other treatments tried for IBS, experiencing and managing feelings and thoughts about the future. Topic guides were used flexibly to ensure that interviewers covered all required topics while allowing for interviewees to introduce unanticipated issues and to have some control over the flow of the interviews. Interviewees were offered the choice of being interviewed by telephone or face to face. All interviews were audio-recorded using a digital audio-recorder.

Participants

Fifty-two trial participants took part in 3-month interviews; their characteristics are shown in Table 33. Ten individuals declined to take part in follow-up interviews: five from each of the CBT arms. Forty-two interviewees were interviewed again at 12 months; their characteristics are shown in Table 34. The reasons participants did not take part in the second interview were that they had no time available in their schedule (n = 3), they had childcare responsibilities (n = 2) or they were not contactable (n = 5).

Facilitators to and barriers of web-delivered and therapist telephone-delivered cognitive–behavioural therapy for irritable bowel syndrome

The facilitators to and barriers of engaging with CBT for IBS can be best understood within the broader context of interviewees’ previous experiences of IBS treatments and their reasons for entering the ACTIB trial. A focused analysis of 3-month interviews was conducted to explore interviewees’ experiences of seeking and appraising treatments for IBS before entering the trial. 68 In summary, this analysis demonstrated that interviewees entered the trial having previously tried a diverse range of treatments for IBS without satisfactory effects on their symptoms or QoL. Accounts of treatment seeking were characterised by a sense of being trapped within a ‘vicious cycle’ of alternating hope (for new treatments) and despair (on finding them ineffective). A desperation and willingness drove interviewees to try any treatment modality available if it offered potential relief. Interviewees derived hope that a new treatment would resolve their symptoms (at the extreme, provide a cure) from various sources, including word-of-mouth recommendations, internet sources, marketing claims and advice from health-care professionals. Treatments that had been tried included various medications (prescribed and over the counter), special diets and complementary or alternative therapies. Interviewees appraised treatments for their effects on symptoms and QoL while also considering, but rarely prioritising, other aspects, including convenience of the regimen itself in the broader context of one’s personal social and working life, whether or not it addressed the perceived root causes of IBS, perceived side effects and adverse impacts on QoL, and cost. On finding a treatment to be ineffective, disappointment often ensued before the search resumed for a more effective remedy for ongoing symptom flare-ups. Repeated disappointing experiences contributed to some interviewees feeling exploited by marketing companies, having reduced confidence in health-care professionals and feeling negatively about themselves and their QoL. Escaping the vicious cycle was helped when individuals found some symptom relief and began to accept that their IBS might need ongoing self-management rather than continuing to hope for a cure.

The invitation to take part in the ACTIB trial thus came to people within a complex context of treatment-seeking behaviours and disappointing experiences. Some of the reasons people gave for enrolling in ACTIB strongly reflected this context. For example, interviewees talked about wanting help with their IBS, being willing to try anything that might help or provide new insight into their IBS, wanting to try a non-medical approach in general or CBT in particular (having tried many medications previously), feeling that their GP had nothing else left to offer them and feeling frustrated or down about their ongoing symptoms and reduced QoL. Less commonly described reasons for participating were altruistic (e.g. to help research and other patients in the future) or passive (e.g. because they received an invitation from their GP):

I was interested when I first heard about it because – of it being about CBT as a treatment because I’ve tried all the medications that my GP had suggested.

ID40370, TAU, 3 months

I was just feeling really miserable about my symptoms and – and anything would help, really.

ID28570, WCBT, 3 months

Typically, interviewees were commencing CBT with some hope that it would provide either a cure or effective relief, but this hope was often tempered by a long history of disappointing experiences of treatments. At the most extreme, some interviewees described fearing that CBT would become another in a long line of disappointing treatments, whereas others were simply desperate for some relief:

I did the programme at a time when I was desperate to find a solution, so I was really willing to – give it a try; so I was motivated, but same time I really didn’t believe it would.

ID40015, TCBT, 12 months

Facilitators

Two key factors appeared to facilitate engagement with both forms of CBT in the trial: the perceived high quality of the intervention itself and the relationship that patients developed with their therapist. Interviewees reported finding the materials and sessions informative and well structured, and liked particular features, such as having some flexibility to self-pace, the follow-up sessions and reminders, the online/telephone format and the ability to refer back to materials at a later date. The relationship that patients developed with their therapist encouraged interviewees to continue engaging with both therapist telephone-delivered and web-delivered CBT. Especially (but not only) in the TCBT arm, this relationship helped interviewees to overcome challenges and to make sense of their experiences over the course of therapy; interviewees particularly valued working with the same therapist across all sessions and praised their support, compassion and professionalism:

Well I liked the fact that – I was not left on my own, I was just – I had someone to talk to and actually follow the progress and then – it was really good because there were some really – there were some times when it was quite hard, like emotionally – I mean things happened last year and at some point I think – I think the symptoms got – they actually began to get worse again and then the fact that I could speak to [therapist] about this, she explained and we went through everything, I think it just really helped, because I think – a month into the programme I would have given [up] because of what happened. And because [therapist] was there, she really gave me the support I needed at that point, so it was really good. I mean otherwise I think I would have stopped the programme before I finished it.

ID40015, WCBT, 3 months

Interviewees from the WCBT arm appeared to value the therapist interactions (which took the form of telephone support calls) predominantly in practical terms (as reminders and/or opportunities to clarify and ask questions about the materials) or in terms of feeling supported and listened to (i.e. having someone to talk to about their IBS):

Except that it made me feel a bit better, to talk to somebody who understood, that I had the freedom to talk about these things that people don’t normally talk about and other people don’t really want to hear. So it was nice to be able to talk to someone about that and be listened to with some understanding; that was great and that was helpful in that sense.

ID39958, WCBT, 12 months

A sense of progression during the trial also appeared to encourage some participants to continue engaging with CBT. This could take the form of experiencing early improvements in symptoms and/or revising and then introducing new topics in each session:

Literally after the second week of doing it, sort of reading through the books and then – talking to [therapist] for the hour and going through everything, it was brilliant and the fact that it did really help, you know, week by week we were talking about different behaviours and – and I think, literally, I sort of saw improvement quite quickly really.

ID24527, TCBT, 3 months

Barriers

An initial scepticism about having CBT for IBS was a potential barrier to engagement mentioned by interviewees from both CBT arms. This scepticism appeared to be driven by having a physical model of IBS, in which there was no logical place for mental processes such as thoughts and feelings. However, many of the interviewees who described initial scepticism also typically reported overcoming this during therapy, suggesting that the intervention materials successfully addressed this issue for most people:

To be honest, when I started I was very sceptical, I couldn’t see how thought processes and things would actually affect your tummy, but when it’s explained through the literature and when you speak to a therapist, you can really see the connection between how you think and how your tummy reacts and – I think it just takes somebody to tell you . . .

ID25119, TCBT, 3 months

The demands that the interventions made on the interviewees and their time presented an ongoing barrier or challenge to engagement with both web-delivered and therapist telephone-delivered CBT. Participants in both therapy arms referred to the need for self-discipline to complete the homework tasks in the CBT programme. A few interviewees felt that the web-delivered sessions were repetitive, which exacerbated the sense of the intervention making excessive demands on their time:

I found it hard um . . . I’m not very good at doing homework and never have been and I don’t suppose I was, um . . . so where it’s given my homework to do, I’ve not – I’ve not been, um let’s say a grade A student.

ID21339, WCBT, 3 months

What did you dislike about being in this group?

I think probably the discipline of having to do the homework, but then I wanted to do the – it’s kind of a bit of a paradox; I wanted to do the homework because I’m keen to participate and kind of make the best of it, but it’s kind of remembering to do it and – having something else to do during the week.

ID25044, TCBT, 3 months

Some interviewees from the WCBT arm reported disengaging with therapy when they felt that the materials were not right for them. This manifested in three main ways:

1. Some interviewees felt that the materials did not provide for their particular level of symptom severity (typically, they felt that their symptoms were more severe than those described in the materials).

2. Some felt that they were not learning anything new about IBS, often describing how they had had it for many years and had already tried the strategies recommended by Regul8.

3. Some felt that the topics were covered in the wrong order for them; for example, they wanted to address stress earlier on so that they would then feel better able to engage with other topics:ID20066, WCBT, 3 months

   I’ve followed all the little sections on the trial, looking at your diet, looking at your stress, looking at your activity and I’ve kind of gone through all of those on my own in the last few, you know, over the years. So from – from my point of view, I didn’t get an awful lot out of it because it was already telling what I already knew.

A sense of reluctance about, or discomfort with, talking about IBS symptoms was an initial deterrent for some interviewees, particularly, but not only, those allocated to the TCBT arm. It may have been that the increased frequency and intensity of telephone conversations in the therapist-delivered group helped these individuals to become more accustomed to speaking about their symptoms and to be more at ease with their specific therapist than those in the web-delivered group, who received fewer telephone calls:

I did think it is odd to do counselling over the phone, but now I think – actually – it doesn’t matter, it doesn’t really matter at all as long as the counselling is good.

ID40210, TCBT, 3 months

The phone calls were OK but I wasn’t as comfortable discussing stuff over the phone.

ID40567, WCBT, 12 months

Insight into the quantitative results of ACTIB

The quantitative results showed that, compared with the TAU arm, both the TCBT and WCBT intervention arms showed significant improvement in primary (IBS SSS and WSAS) and secondary (PEQ and SGA of relief of symptoms) outcomes. Expanding on these results, interviewees described, in their own words, the benefits that they perceived as resulting from their participation in ACTIB and, specifically, from the CBT that they received. The majority of participants in the therapy arms reported improvements that they attributed to CBT. Perceived benefits included symptom-related improvements (e.g. reduced symptoms and less severe symptoms), reduced use of medications/other health services and broader improvements in coping (e.g. with symptoms, emotions, stress, life events, interpersonal relationships) and QoL (e.g. confidence in daily life, re-engaging with social activities):

I think it is much improved really; I’ve not had as much sort of constipation as what I used to have, so – so yes – for me, it has been really, really good.

ID24547, TCBT, 3 months

Previously I would maybe see the doctor every 6 months maybe or when I was about to have a flight or a journey or some stressful situation and I’d end up with some omeprazole or something, to stop acid and IBS-type symptoms. And I haven’t had to do that for all of the time since the course finished, the ACTIB course finished.

ID20822, TCBT, 12 months

I think, in a way, it has helped me cope with some of the emotional side of that better than if I hadn’t had any therapy. I mean actually I think probably if I hadn’t had the CBT at all and I’d gone through the same problems, I would be even worse than it is at the moment.

ID40210, TCBT, 12 months

I’ve learned some new things about the condition and things that you can do to – help any symptoms when they come on and they were things that were different, from things I’ve been told by the doctor.

ID20068, WCBT, 12 months

But I mean overall I’ve seen a huge difference in who I am, how I am, what I say, think, how I behave and my outlook on life; there has been a lot of changes in that sense.

ID10074, WCBT, 12 months

. . . [it has changed] the way that I feel about my IBS, it’s just something there and if it happens it happens and I can deal with it, whereas it’s not kind of overtaking my life any more.

ID28570, WCBT, 12 months

Yes, I think it’s really helped me to – not ignore my symptoms, but to see them for what they are and so knowing that, OK, I know what’s happening now, I know why it’s happening, but still [carry on with] my day-to-day life much more confidently. I think I’ve become more confident and will just go out anyway and stuff like that.

ID25044, TCBT, 12 months

It is possible that some of the subjective perceived effects of CBT had not been fully captured by the quantitative outcome measures. A small number of interviewees described struggling to express their experiences within the confines of the questionnaires:

Some questions I felt you could elaborate on, um [Interviewer: Uh huh], for instance, if it said – ‘does it stop you going out?’. I mean, I could put yes, but I – I said no because I force myself to go out, but just to answer no, it makes you feel like you’re not worrying about it.

ID21140, TAU, 3 months

So it’s been very frustrating and it’s frustrating because – my experience of IBS – does not seem to fit the questions.

ID39958, WCBT, 3 months

A few interviewees reported not experiencing benefits from CBT, particularly in the longer term, when interviewees talked about forgetting or neglecting newly learned practices and finding old habits returning:

I would like to say that – things changed and it was helpful . . . but . . . but I don’t think so.

ID39958, WCBT, 12 months

I found it very helpful at the beginning, and obviously as time [fades], you forget about the practices.

ID40192, TCBT, 12 months

Potential social and psychological change processes

In addition to the changes in symptoms, QoL and other ‘outcomes’ described above, interviewees talked about changes that might have contributed to these outcomes. Notably, many of these changes mapped well onto the theoretical model underpinning CBT. For example, interviewees described changes in the attitudes, beliefs and behaviours that are broadly consistent with the CBT model of IBS:

So I think – there isn’t enough – promotion of anxiety and stress and things that can make your IBS worse; so I think understanding that more and [learning] techniques to help manage those situations has helped.

ID40567, WCBT, 12 months

I have now come to understand that the physical is just the manifestation of what is going on in the head.

ID10074, WCBT, 12 months

I don’t know what your name for them was but there are reminders of things to say or do just to remind yourself of – of some of the thought processes. And I had three or four of those as tools to use, so if I did feel the anxiety coming on then a little bit of deep breathing, diaphragm-type breathing, I felt that that was the solution and if it was a solution or not, it worked. So that gives you confidence and that works. Also looking – looking at things from a much wider perspective, so – look at the evidence, that was another one, a real little trick. You’re sitting in an aeroplane, everyone around you is just chatting, reading their books, having drinks, nobody’s fretting and getting unwell or anything; it’s just a journey. So the evidence all around you is overwhelming and positive, compared to the tiny little bit that your brain is trying to manoeuvre you into, to feeling negative or anxious. So that’s another little tool I’ve remembered and set aside to use if I need to. All those things like eating well, going to bed earlier, exercise. Another one: if I feel that I’m going to have a situation where stress has built up to a point, I can sense, if you like, the triggers of maybe being a bit forgetful or whatever and your mind being too much on stressful things, to get some exercise and get the cortisol levels down and do something physical. So I’ve maintained physical exercise that I hadn’t had before the CBT, which I think is probably favourable, as well.

ID20822, TCBT, 12 months

I started thinking to myself, you know, I don’t need to stress and worry about it, there is a toilet here, it’s there if I can use it, which has helped, because I used to go home early from parties and things like that with a tummy ache. But now I can just think, you know, I’m working myself up about it and then, more often than not, the feeling goes away and then I’m absolutely fine.

ID25119, TCBT, 3 months

I think like with the thoughts, just being aware of the sort of things that can kind of perpetuate the cycle of stress. I think catastrophising or black-and-white thinking and things like that, I think I can see them in myself and I think just being aware of that, you can kind of try and take a back step and see it in another way and re-evaluate the situation.

ID26417, TCBT, 3 months

Interviewees also described gaining greater insight into the nature of IBS. Crucially for some, this entailed the realisation that they should no longer be seeking a single curative treatment. In other words, it helped them to break out of the vicious cycle of treatment seeking that had characterised their approach before enrolling in ACTIB. This was made easier by having gained effective coping strategies from the CBT:

Well managing stress and obviously helping to reduce or minimise the flare-ups of the IBS; that’s the benefits that I got out of it; understanding what’s happening, I suppose is a benefit, because when you understand something a lot better, you’re better prepared to deal with it, so that’s obviously very important, very useful.

ID20850, TCBT, 12 months

I think that I understand a bit more about my IBS and stop thinking that there is going to be one solution to this, because before I used to think that.

ID16045, TCBT, 12 months

I just think that I can cope much better with my symptoms, which is maybe not that – has helped me with not having so many bad stomach aches, because I sort of do know how to manage my symptoms now. Before, I think my mindset was – I’m never going to get over this, it’s not going to go away; whereas now I can understand – it doesn’t stop me – my symptoms don’t stop me from doing stuff.

ID24547, TCBT, 12 months

For some participants, the CBT offered during the trial seemed to improve their understanding of emotional states and triggers:

I do think that – the therapy [received during the trial] I guess, because I think sometimes you – you don’t – know what triggers your emotions and I think, you know, going on the sort of different courses has helped understand how to control and trigger.

ID24547, TCBT, 3 months

The CBT has helped untangle what upsets me sometimes [Interviewer: OK] or what – like – how I feel about things and writing it down. Previously it’s been quite difficult to figure out how I feel about things. [Interviewer: OK] I just can’t figure it out – but now I can figure it out a bit better, because I think – if it’s manually written down, I felt like this – this made me feel this – and then I’ve looked down the sheet and then I’m like, OK, so I feel anxious a lot of the time.

ID33746, WCBT, 3 months

Furthermore, only participants from the CBT groups tended to describe their personalities as strong or resilient when talking about problem-solving emotions, suggesting that they felt more empowered than the TAU group. A sense of empowerment to manage negative emotions was a common pattern reported by the CBT participants when asked about their experiences of the main trial. The way CBT participants talked about a perceived increased control over their IBS symptoms, gained through new self-management techniques learned during the study, may have had an impact on the way they subsequently identified and described active coping strategies related specifically to managing their negative emotions. Furthermore, emotional regulation may have been addressed with some individuals if it arose as a pertinent issue in the therapy sessions:

I think it’s just when I feel very negatively about something, I actually do something to correct it. So I’m not someone who just accepts the situation; I’ve got quite like a strong fighting personality, so – so when things get too negative, then I just – I just change them.

ID40015, TCBT, 3 months

Negative feelings? [. . .] so I’m not a negative person and when I deal with negative things, I try and turn it around to positive.

ID40102, TCBT, 3 months

Participants receiving CBT also explained in more detail the analytical processes that allowed them to reframe a situation. Some explicitly acknowledged the positive role CBT played in changing their thought processes:

Yes. I think I am getting better because – so yes – they’re going down because I’m more aware of – that the feelings that I’m getting, I can recognise a bit more when I’m feeling anxious or when I’m feeling stressed. And I am trying to put into use the things I’ve learned from my online sessions and my sessions with the therapist [. . .] Yes, so the sessions have helped with that, the way you have to think about what’s causing you to feel like that and what’s the worst that can happen; kind of breaking it down to try and think better or more helpful thoughts. So that’s what I’m trying to do now.

ID40567, WCBT, 3 months

I think, in a way, it has helped me cope with some of the emotional side of that better than if I hadn’t had any therapy. I mean actually I think probably if I hadn’t had the CBT at all and I’d gone through the same problems, I would be even worse than it is at the moment.

ID40210, TC, 12 months

Summary

• The benefits of CBT included reduced symptoms and, arguably more importantly, an increased capacity to cope with symptoms, negative emotions and other challenges of daily life.

• Therapists had an important role to play in supporting patients to engage with CBT and to make sense of the therapy and their IBS. Patients valued having therapist support available alongside Regul8 and this may have helped enhance their engagement and outcomes.

• The CBT appeared to help patients make changes to their ways of thinking and behaving in relation to their IBS, which was broadly consistent with the underpinning theoretical mechanisms of change underpinning this intervention.

Strengths and limitations

Implications for practice

If the interventions were rolled out in practice, the qualitative data suggest that it would be important to:

• offer therapist support alongside Regul8

• consider augmenting the intervention to provide additional support for sustaining behaviour changes over time

• help GPs/consultants to encourage even those patients who are initially sceptical about a talking therapy to try CBT for IBS.

Service use and costs by time period

In the 6 months prior to baseline, > 80% of participants had contact with GPs (Table 35). Relative to other services, there were high rates of contact with other doctors, pharmacists and practice nurses. Relatively few had inpatient stays. Medication was received by just over half of each group and investigations (usually blood tests) were carried out in around three-quarters of participants. The mean number of contacts in Table 37 (and similar tables for other time periods) is only for those with at least one contact (i.e. excluding those not using the service). If a service was used, then there were usually < 10 contacts during the period. Health service costs were highest for inpatient care (despite its low use), GP contacts and other doctor contacts. There were few notable differences between the three groups prior to baseline. The mean total health-care costs were £681 for TCBT, £620 for WCBT and £802 for TAU. Although the TAU costs are relatively high, the SDs around the mean costs are substantial, as is common with cost data, at £122 more than for TCBT (bootstrapped 95% CI –£93 to £351) and £182 more than for WCBT (bootstrapped 95% CI –£27 to £393). During the baseline period, about one-fifth of each group received care from family or friends because of health problems. For those in receipt of this informal care, the number of hours was, on average, between 9 and 13 per week. This time, valued using the average hourly wage of £15, resulted in costs that were relatively high compared with the health-care costs. Lost work days were experienced for about half of the TCBT and TAU groups and 40% of the WCBT group. An average of about 10 days were lost from work for those for whom this was the case. Combining the informal care and lost employment costs with the health-care costs results in mean societal costs at baseline of £1995 for TCBT, £1965 for WCBT and £2352 for TAU. TCBT had societal costs that were £357 lower than those for TAU (bootstrapped 95% CI –£607 to £1214) and WCBT had costs that were £387 lower than those for TAU (bootstrapped 95% CI –£683 to £1349).

In the period prior to the 3-month follow-up, around half of participants in each group had GP contacts (Table 36). The next most used service was pharmacist contacts. No clear differences in service use were observed between the groups. Inpatient use was low in each group, but in the WCBT group the cost of inpatient care was very high because of one participant having an extended period of time in hospital. Again, costs tended to be high for GP and other doctor contacts. Inpatient costs for the WCBT group were substantially higher than for the other two groups. Medication was used by around one-third of the TCBT group and slightly more of the other two groups. Investigations were received by around one-third of each group. Overall, the groups were similar in terms of service use. The mean total health-care costs excluding the intervention were £271 for TCBT, £346 for WCBT and £227 for TAU. Controlling for baseline, the TCBT group had costs that were £52 higher than those for TAU (bootstrapped 95% CI –£60 to £170) and WCBT had costs that were £152 higher than those for TAU (bootstrapped 95% CI –£66 to £508). Informal care was received by similar proportions in each group, but TAU participants received more hours per week, which resulted in higher informal care costs for this group. Lost work days were similar in each group. The mean societal costs were £682 for TCBT, £723 for WCBT and £836 for TAU. Controlling for baseline, TCBT had mean costs that were £28 higher than those for TAU (bootstrapped 95% CI –£309 to £361) and WCBT had costs that were £218 higher than those for TAU (bootstrapped 95% CI –£226 to £649).

During the 3 months prior to the 6-month follow-up, similar patterns of service use as before were observed, with around half of all participants receiving GP care, around one-quarter seeing a pharmacist and small numbers having inpatient care (Table 37). Costs tended to be highest for GPs and other doctors. Mean non-intervention health-care costs were £281 for TCBT, £224 for WCBT and £206 for TAU. After controlling for baseline costs, it was shown that TCBT had £84 higher costs than TAU (bootstrapped 95% CI –£41 to £230) and WCBT had costs that were £41 higher than those for TAU (bootstrapped 95% CI –£70 to £165). Informal care costs were very different during this period. Although the proportions in receipt were similar, the mean number of hours per week was far greater for the TAU group. The number of lost work days was lowest for the WCBT group. The mean societal costs were £840 for TCBT, £503 for WCBT and £1344 for TAU. Controlling for baseline resulted in TCBT having costs that were, on average, £350 lower than those for TAU (bootstrapped 95% CI –£332 to £996) and WCBT having costs that were £407 lower than those for TAU (bootstrapped 95% CI –£74 to £923). It is of interest that none of these differences was statistically significant. This is due to the large SDs around the costs and the pre-existing baseline differences.

Finally, in the 6 months prior to the 12-month follow-up, we see highest use for GPs, other doctors, pharmacists and practice nurses (Table 38). Inpatient use was slightly higher than before and this resulted in relatively high inpatient costs, especially in the TCBT and TAU arms. The mean total health-care costs were £519 for TCBT, £325 for WCBT and £393 for TAU. After controlling for baseline, TCBT was shown to have non-intervention health-care costs that were, on average, £141 higher than those for TAU (bootstrapped 95% CI –£118 to £474) and WCBT had costs that were £43 lower than those for TAU (bootstrapped 95% CI –£115 to £205). The higher level of informal care for TAU was maintained during this period. TAU also had higher lost employment costs than the other two arms. The mean societal costs were £1055 for TCBT, £1029 for WCBT and £2103 for TAU. Controlling for baseline, TCBT had costs that were £850 lower than those for TAU (bootstrapped 95% CI –£262 to £2181) and WCBT had costs that were £748 lower than those for TAU (bootstrapped 95% CI –£175 to £1953).

Costs over follow-up

The mean total health-care costs, including the intervention, over the 1-year follow-up period were £1650 (SD £1931) for TCBT, £943 (SD £955) for WCBT and £715 (SD £884) for TAU. Adjusting for baseline differences, TCBT was on average, £943 more costly than TAU, a difference that was statistically significant (bootstrapped 95% CI £572 to £1363), and WCBT was £278 more costly than TAU, which was also a statistically significant difference (bootstrapped 95% CI £11 to £514). The cost differences for the participants for whom EQ-5D-5L data were available (which is relevant for the complete-case analysis) were £956 (bootstrapped 95% CI £601 to £1435) for TCBT and £224 (bootstrapped 95% CI –£11 to £448) for WCBT.

The mean societal costs over the follow-up period, including therapy costs, were £3065 (SD £5179) for TCBT, £2094 (SD £3069) for WCBT and £4374 (SD £11,843) for TAU. After controlling for baseline, it was found that TCBT cost £858 less than TAU (bootstrapped 95% CI –£1063 to £2812) and WCBT costs £1028 less than TAU (bootstrapped 95% CI –£404 to £2626).

As stated above, medication costs were not included in the totals because of the quality of the data. We did know what medications were taken but the data on frequency and dose were not complete. What is evident from Tables 35–38 is that there was a slight reduction in medication use after baseline and participants in the TAU group were slightly more likely to be on medication. The majority of prescriptions were for very inexpensive oral drugs and so any cost saving would be very small.

Quality-adjusted life-years

Table 39 shows that EQ-5D-5L scores were relatively high for each group at each time point. It is evident that improvements were greater for the two therapy groups than for TAU. Controlling for baseline EQ-5D-5L utility scores, TCBT resulted in 0.0414 more QALYs than TAU (95% CI 0.0194 to 0.0635 QALYs) and WCBT resulted in 0.0269 more QALYs than TAU (95% CI 0.0041 to 0.0497 QALYs). The QALY differences for those with follow-up cost data are 0.0429 QALYs (95% CI 0.0205 to 0.0653 QALYs) and 0.0290 QALYs (95% CI 0.0063 to 0.0518 QALYs), respectively.

Cost-effectiveness results: complete-case analyses

Dividing the health-care cost differences by the QALY differences for those with both sets of data results in the following ICERs: TCBT versus TAU, £22,280; and WCBT versus TAU, £7724. The ICER for TCBT versus WCBT was £52,662.

Figures 8–10 show the uncertainty around the ICERs. The points on these cost-effectiveness planes represent a pair of incremental costs and QALYs from 1000 bootstrapped samples. Figure 8 shows that TCBT is certain to result in higher costs than TAU and to produce more QALYs. From Figure 9, which compares WCBT with TAU, we see that there is a 96.2% likelihood that WCBT is more expensive than TAU and produces more QALYs, a 3.2% likelihood of lower costs and more QALYs, a 0% likelihood of lower costs and fewer QALYs and a 0.6% likelihood of higher costs and fewer QALYs. Figure 10 reveals that, compared with WCBT, there is a 91.4% likelihood that TCBT is more expensive and produces more QALYs, a 0.0% likelihood that it is less expensive and produces more QALYs, a 0% likelihood it is less expensive and produces fewer QALYs and a 8.6% likelihood it is more expensive and produces fewer QALYs. Finally, the CEACs in Figure 11 reveal that, at very low values placed on a QALY gain, TAU is most likely to be cost-effective. With higher values placed on a QALY, WCBT becomes the most likely treatment to be cost-effective. The probability that TCBT is most cost-effective does increase but only exceeds the probability for WCBT at £55,000 per QALY. At the £20,000 threshold commonly used in evaluations in the UK, WCBT is most likely to be cost-effective, followed by TAU and then TCBT.

FIGURE 8.

Cost-effectiveness plane for TCBT vs. TAU (health-care perspective, complete-case analysis).

FIGURE 9.

Cost-effectiveness plane for WCBT vs. TAU using QALYs.

FIGURE 10.

Cost-effectiveness plane for TCBT vs. WCBT.

FIGURE 11.

Cost-effectiveness acceptability curves based on QALYs.

Using complete data, the difference in IBS SSS at 12 months compared with TAU, controlling for baseline, was 84 for TCBT and 56 for WCBT. The corresponding cost differences are £962 and £281, respectively. This leads to an ICER for TCBT compared with TAU of £11 per unit improvement on the IBS SSS, and an ICER for WCBT compared with TAU of £5. CEACs based on the IBS SSS outcome measure are shown in Figure 12. As with the QALYs, TAU is most likely to be cost-effective for very small values placed on a 1-point improvement on this measure. However, if the value is around £5 per 1-point improvement, then WCBT becomes the option that is most likely to be cost-effective. TCBT becomes most likely to be cost-effective if the value placed on a 1-unit improvement exceeds £24.

FIGURE 12.

Cost-effectiveness acceptability curves based on IBS SSS.

The difference at 12 months on the WSAS between TCBT and TAU was 5.1 and between WCBT and TAU was 4.0. The corresponding cost differences are £943 and £278, respectively. The ICERs based on these figures are £185 for TCBT and £70 for WCBT. In Figure 13, it can be seen that, when the WSAS is used, the probability that WCBT is the most cost-effective option occurs for values placed on a 1-unit improvement on this measure in excess of £60. TCBT is unlikely to be the most cost-effective option for values placed on a 1-unit improvement on the WSAS within the range £0–200.

FIGURE 13.

Cost-effectiveness acceptability curves based on WSAS scores.

Cost-effectiveness results: sensitivity analyses

When therapy costs were reduced by 25% and 50%, the ICERs (health-care perspective) compared with TAU were reduced for both groups and brought TCBT well below the £20,000 threshold (Table 40). Increasing therapy costs had the opposite effect, but even with a 50% increase the WCBT option was still well below the £20,000 threshold.

As stated previously, there were many cases with missing EQ-5D-5L data and, consequently, with missing QALYs. After imputation, the number of QALYs over the follow-up was 0.8519 for TCBT, 0.8300 for WCBT and 0.8184 for TAU. After controlling for baseline, the incremental QALY gain for TCBT compared with TAU was 0.0280 and for WCBT compared with TAU was 0.0172. The mean service costs after imputation were £1650 for TCBT, £1151 for WCBT and £886 for TAU. The incremental costs compared with TAU were £812 for TCBT and £337 for WCBT. This results in an ICER for TCBT of £29,000 per QALY and an ICER for WCBT of £19,593 per QALY.

When the minimum wage was used to value informal care and lost work days, the mean societal costs were reduced to £2362 for TCBT, £1520 for WCBT and £2518 for TAU. After controlling for baseline, TCBT cost £45 more than TAU (bootstrapped 95% CI –£1145 to £1124), whereas WCBT cost £389 less than TAU (bootstrapped 95% CI –£391 to £1283). When informal care was valued using the cost of a home-care worker, the mean societal costs were £3281 for TCBT, £2275 for WCBT and £5098 for TAU. After controlling for baseline, we found that TCBT cost £1228 less than TAU (bootstrapped 95% CI –£1061 to £3917) and WCBT cost £1241 less than TAU (bootstrapped 95% CI –£498 to £3351).

Limitations

There were limitations to this economic evaluation. First, and most importantly, the number of missing EQ-5D-5L data was a concern. Those with missing data had worse IBS SSS and WSAS scores at each time point and so imputation from these resulted in smaller QALY gains for TCBT and WCBT relative to TAU. Second, service use data were provided by participants themselves and there may have been recall accuracy problems. However, this would not be likely to affect one group more than another and it was the only option for collecting comprehensive data. Third, medication data were not of high quality. We did know what medications were taken but quantities and durations were not complete. However, these costs would be a small proportion of the total and their inclusion would have only a marginal effect. Finally, this was a trial and the implementation and delivery of interventions outside the trial setting may be less than optimal.

Trial Steering Committee

Professor Peter White, Professor of Psychological Medicine (Chairperson); Professor Else Guthrie, Professor of Psychological Medicine and Medical Psychotherapy; Professor Qasim Aziz, Professor of Neurogastroenterology; Professor Tom Sensky, Emeritus Professor in Psychological Medicine; and Ms Jill Durnell, Owner and Managing Director at JayDee Solutions (PPI member). Observers: Dr Hazel Everitt, Professor Trudie Chalder, Professor Rona Moss-Morris, Dr Gilly O’Reilly, Dr Martina Prude (Sponsor Representative) and Dr Jenny Baverstock (Primary Care Research Network Representative).

Data Monitoring and Ethics Committee members

Professor Astrid Fletcher, Professor of Epidemiology of Ageing (Chairperson); Dr Charlotte Feinman, Consultant Psychiatrist; Professor Ronan O’Carroll, Professor of Psychology. Observers: Dr Hazel Everitt, Professor Trudie Chalder, Professor Rona Moss-Morris and Dr Gilly O’Reilly.

Introduction

• Thank you for agreeing to take part in this interview [short hello].

• I would like to record the conversation we have today so that I can refer back to it at a later date. It enables me to listen to you better. Is that OK?

• Before we start there are a few things I would just like to mention.

• What we talk about will be used as part of the study, but anything said will remain anonymous. We are going to ensure this by not using your real name when we type up the interview. Is that OK?

• If I ask a question that you do not want to answer, that is absolutely fine. Just say so and I will ask you a different question. If at any point you would like to stop participating then please just tell me and we will stop the interview.

• I am currently contacting a number of people who are taking part in the ACTIB study to find out about their experience of taking part in the trial, the treatments used in the trial, other IBS treatments they have tried in the past and the way they feel.

• Anything you can tell me about your experiences including good and bad points would be useful.

• Do you have any questions before we start? Are you happy to continue?

Section A

First I would like to ask you about the trial you are taking part in.

1. Could you tell me all about your experience of taking part in the trial so far?

Prompt: I am interested in why you agreed to take part in the trial and what you were initially expecting from it.

I am very interested to hear about the treatment that you have had as part of being in the trial (website delivered, therapist delivered, treatment as usual).

1. Could you tell me all about the treatment that you had for your IBS as part of the trial?

Additional prompts to be used flexibly if necessary:

1. What were your expectations of this group before initially trying it?

   1. How did you feel about being allocated to this group in the beginning?

   2. How did you feel about the website-delivered group after a while?

2. What did you like about being in this group?

3. What did you dislike about being in this group?

4. How has your IBS been since you have been using this treatment?

Can you tell me about anything that you feel changed while being in this group? (Symptoms, thoughts, feelings, lifestyle, social/relationships.)

1. Are there any other things that have changed?

Looking back on this treatment now . . .

1. What do you think about this treatment for IBS?

2. Do you have any particular feelings about the use of this treatment for IBS?

Section B

In this section of the interview, I would like to find out about other treatments that you have tried for your IBS.

1. Could you tell me about the most helpful treatments you have tried for your IBS?

   1. What did you like/dislike about these treatments?

2. Could you tell me about the least helpful treatments you have tried for your IBS?

   1. What did you like/dislike about these treatments?

Prompt: I’m interested in the treatments you have tried before taking part in the trial or any that you have tried since.

For the treatment:

1. What led you to try out this treatment?

2. How would you compare these treatments to each other?

3. How do the past treatments compare with the treatment received during the trial?

Section C

Until now, we have talked about your experience of taking part in the trial and the different treatments for your IBS. In the following section of the interview, I would like to find out more about how you feel in general and the way you manage your feelings. There are no right or wrong answers; I am just interested in your own personal experience.

1. Could you talk me through how you feel on a typical day?

   1. Has it always been like this for you? Prompts: have there been any changes recently? If yes, which ones?

2. How do you express your emotions or feelings to other people? Prompt: why do you think this is? What about other people like colleagues of friends?

   1. Has it always been like this for you? Prompts: have there been any changes recently? If yes, which ones?

3. Could you describe a time that you have experienced negative feelings?

   1. Can you tell me what you do/did when you experience negative feelings?

   2. Is there anything else that you do to cope with negative feelings? If yes, please describe.

4. Could you now describe a time that you have experienced positive feelings? Prompt: can you tell me what you do when you experience positive feelings?

Additional prompts to be used flexibly if necessary:

1. How easy/difficult is to work out how you feel?

2. Do you feel in control of your emotions? Why do you think this is?

Section D

In the final section of the interview, I am interested in your thoughts about what happens after the trial.

1. How have you been doing since you finished the first set of telephone support sessions? (For active arms only.)

2. What do you think will happen next with your IBS?

3. Is there anything that we could do differently to improve our treatments for people with IBS?

Closing and ending

• Thank you very much for sharing your thoughts and experiences with me today.

• What you have told me will really help us to understand patients’ experiences and hopefully to improve our treatments for IBS.

• Before we finish, is there anything else you want to tell me? Is there anything you want to ask me?

• Offer the participant a copy of the transcript and/or a summary of the findings.

Introduction

• Thank you for agreeing to take part in this interview [short hello].

• I would like to record the conversation we have today so that I can refer back to it at a later date. It enables me to listen to you better, is that OK?

• Before we start there are a few things I would just like to mention.

• What we talk about will be used as part of the study, but anything said will remain anonymous. We are going to ensure this by not using your real name when we type up the interview. Is that OK?

• If I ask a question that you do not want to answer that is absolutely fine, just say so and I’ll ask you a different question. If at any point you would like to stop participating then please just tell me and we will stop the interview.

• I’m currently contacting a number of people who are taking part in the ACTIB study to find out about their experience of taking part in the trial, the treatments used in the trial, other IBS treatments they have tried in the past and the way they feel.

• Anything you can tell me about your experiences including good and bad points would be useful.

• Do you have any questions before we start? Are you happy to continue?

Section A

First I would like to ask you about the trial you are taking part in.

1. Now that it has been a year since you started in the trial, can you tell me what you think about your experiences with the study?

As part of the trial, you received [treatment as usual/website + telephone support/telephone CBT]. Now that it has been a little while since your treatment finished, could you tell me what you think about it?

Additional prompts to be used flexibly if necessary:

1. Can you tell me about anything that you feel changed because of this treatment? (Symptoms, thoughts, feelings, lifestyle, social/relationships.)

   1. Are there any other things that have changed?

2. How has your IBS been since we last spoke?

3. Do you think the treatments had any long-term effects on you?

4. Looking back on this treatment now . . .

   1. What do you think about this treatment for IBS?

   2. How do you feel about the use of this treatment for IBS?

   3. What do you like about this treatment?

   4. What do you dislike about this treatment?

Section B

In this section of the interview, I would like to find out about other treatments that you have tried for your IBS.

1. Could you tell me about the most helpful treatments you have tried for your IBS since we last spoke?

   1. What did you like/dislike about these treatments?

2. Could you tell me about the least helpful treatments you have tried for your IBS since we last spoke?

   1. What did you like/dislike about these treatments?

For the treatment:

1. What led you to try out this treatment?

2. How would you compare these treatments to each other?

3. How do the past treatments compare with the treatment received during the trial?

Section C

Until now, we have talked about your experience of taking part in the trial and the different treatments for your IBS. In the following section of the interview, I would like to find out more about how you feel in general and the way you manage your feelings. There are no right or wrong answers; I am just interested in your own personal experience.

1. Could you talk me through how you feel on a typical day?

   1. Has it always been like this for you? Prompts: have there been any changes recently? If yes, which ones?

2. How do you express your emotions or feelings to other people? Prompt: why do you think this is? What about other people like colleagues of friends?

   1. Has it always been like this for you? Prompts: have there been any changes recently? If yes, which ones?

3. Could you describe a time that you have experienced negative feelings?

   1. Can you tell me what you do/did when you experience negative feelings?

4. Is there anything else that you do to cope with negative feelings? If yes, please describe.

5. Could you now describe a time that you have experienced positive feelings? Prompt: can you tell me what you do when you experience positive feelings?

Additional prompts to be used flexibly if necessary:

1. How easy/difficult is to work out how you feel?

2. Do you feel in control of your emotions? Why do you think this is?

Section D

In the final section of the interview, I’m interested in your thoughts about what happens after the trial.

1. How have you been doing since you finished the first set of telephone support sessions? (For active arms only.)

2. What do you think will happen next with your IBS?

3. Is there anything that we could do differently to improve our treatments for people with IBS?

Closing and ending

• Thank you very much for sharing your thoughts and experiences with me today.

• What you have told me will really help us to understand patients’ experiences and hopefully to improve our treatments for IBS.

• Before we finish, is there anything else you want to tell me? Is there anything you want to ask me?

• Offer the participant a copy of the transcript and/or a summary of the findings.