Summary of main findings
The trial has demonstrated that it is feasible to recruit and randomise patients with IBS from many GP practices to a complex trial and achieve high follow-up rates.
The results must be interpreted with caution as this is an exploratory study with relatively low numbers and power. No statistically significant differences were seen between the medication groups for the main outcome measures (IBS SSS and IBS-QOL) at 6 or 12 weeks. Nor were there any statistically significant differences between the website groups for the IBS-QOL at 6 or 12 weeks or IBS SSS at 12 weeks. However, there was a statistically significant difference seen in the IBS SSS for the website at 6 weeks, with the ‘no website’ group having a lower IBS SSS score than the website groups, but this difference disappeared by 12 weeks.
There were no differences in the secondary outcomes in the medicine groups.
In the website groups, there was significantly increased Enablement at 6 and 12 weeks and a significantly more participants scored Subjective Assessment of Global relief (SGA) as improved at 12 weeks compared to the No Website group.
Strengths and limitations
This exploratory trial was the first to look at providing CBT as a web-based self-management programme for patients with IBS. The factorial design allowed the feasibility of undertaking a trial to assess the effectiveness of common IBS medications to be explored at the same time.
Participants were asked to take 7 large over-encapsulated tablets a day for 6 weeks. Despite this, compliance with the medication taking was good.
The lack of difference between the medication groups may be due to low power with the limited numbers in this trial. A larger trial will be able to provide more definitive results. Ideally, smaller trial medication would be sourced for a larger trial to further improve compliance and reduce the placebo effect of taking the tablets.
Expectancy effects (i.e. people's expectancy of web treatment and drug treatment) and the placebo effect may have influenced the trial results. We did not measure treatment expectancy effects in this study.
Over 5700 patients were invited by their GP to participate in this study with 135 being randomised. Thus a small subset of the patients recorded as having IBS by their GPs entered the study. This low recruitment rate is not unexpected in a study requiring patients to be on no initial medication, to take many large tablets a day and participate in a self-management programme for 6 weeks, but it may reflect a more committed group of participants and affect the generalisability of the study findings. Reasons for declining to enter the study have been sought and recorded and the study sample is similar in gender and age to the invited participants.
Comparisons with other studies
There are a limited number of published trials of CBT for IBS but they indicate a positive effect for CBT on IBS over a range of outcome measures [5, 9, 18, 30] in the short-term (up to 3 months). However, information is lacking for longer term follow up. For instance, a trial of face to face CBT plus mebeverine (6 × 50 minute CBT sessions over 6 week delivered by general practice nurses) against mebeverine alone in a primary care setting in London  showed a significant reduction in IBS-SSS in the CBT arm at 3 months but not at 12 months follow up.
The lack of improvement in the IBS SSS and IBS-QOL in the website self-management groups in this study is disappointing as the paper-based self-management programme from which the Regul8 website was developed  did show a significant improvement in IBS SSS scores in the self-management group compared to the control group. It may be that an effect of the website was masked by the general improvement in all groups during the length of the trial. It is acknowledged in many trials that IBS symptoms have a large placebo response to any intervention and this can make it difficult to distinguish true treatment effects, particularly in a small study. However, we found a trend for continued improvement in the self-management group (particularly those who had the telephone support) whilst those in the No Website group and the medication groups appeared to lose some of these gains at 12 weeks (see Figures 2 and 3). Longer term follow up may reveal differences between the website groups. Two other reasons why the Regul8 programme may be less successful than the paper-based manual are that: the earlier study used an experienced therapist, and had significantly more therapist contact time. In this study a nurse with only minimal CBT training delivered the intervention. In addition, of those that had access to the Regul8 programme, half had no therapist support at all and half had one, 30 min telephone support session, whereas in the trial of the paper-based manual, one face to face and two telephone support sessions were offered to participants (3 hours in total). Therapist support is expensive compared to a stand-alone website and thus was deliberately kept to a minimum for this trial. However, recently published trials of internet interventions for IBS [18, 30] and qualitative data from this study suggest that it is important to assist engagement with the website.
The finding that Enablement scores were higher in those that had access to the Regul8 programme suggests that these participants felt better able to cope with their symptoms. Additionally, those that received the telephone session had higher Enablement than those with access to the website without support. This suggests benefit from the website which is enhanced by therapist support and strengthens the postulation that the amount of support offered in this study may have been insufficient. Ability to cope with IBS symptoms and continue a full life is clearly an important outcome irrespective of symptom severity. Improvements in the Work and Social Adjustment Scale, a measure of the ability to continue work and other activities with IBS, was shown to be improved by CBT in the long term (12 month follow up) in the Kennedy trial .
The lack of any significant effect on IBS outcomes in the medication arms of this study, may be due to limited power, and would need to be confirmed in a larger trial but are not unsurprising as a Cochrane review  has highlighted the poor evidence of their effectiveness.
Implications for future research and clinical practice
The factorial trial design for this study had the benefit of providing data for both the medication and the website groups but this may have limited the ability of the trial to show an effect for the CBT self-management programme, since the key to self-management is being able to attribute symptom improvement to cognitive and behavioural changes. In a blinded trial participants would have had some difficulty knowing whether any change in their symptoms was due to the medication they were taking or to the CBT self-management principles they were trying to employ. This issue was raised in discussion with patients in the qualitative interviews. The effect of including medication the factorial design, may have been to reduce their confidence in employing the CBT techniques. Thus a definitive trial of the website should probably not include a blinded medication component.
When recruiting participants from primary care for this trial, it was found to be important to screen potential participants with the ROME III and exclusion criteria as 98 of the 265 potential participants who completed the screening questionnaire either did not fulfil the ROME criteria or met exclusion criteria.