Key mechanisms of cognitive behavioural therapy in irritable bowel syndrome: The importance of gastrointestinal related cognitions, behaviours and general anxiety

Highlights

• Gastrointestinal related cognitions, behaviours and general anxiety mediate treatment effect of CBT in IBS

• Change in gastrointestinal cognitions precede reduction in general anxiety

• Change in gastrointestinal safety behaviours precede reduction in general anxiety

• Change in gastrointestinal avoidance behaviour was not a mediator of treatment effect

• Psychological treatments in IBS should target gastrointestinal specific responses for change in IBS

Abstract

Background

Irritable bowel syndrome (IBS) is a functional gastrointestinal disorder characterised by abdominal pain and altered bowel movements. Cognitive behavioural therapy (CBT) has been shown to be effective in reducing symptom severity in IBS and enhancing quality of life/functioning. The present study sought to identify how CBT achieves change in these outcomes.

Method

Secondary analysis was conducted on 149 patients with irritable bowel syndrome who had been randomised to cognitive behavioural therapy plus an antispasmodic medication or antispasmodic alone. Single and sequential mediation was modelled using structural equation modelling. Gastrointestinal (GI) related avoidance behaviour, safety behaviour, cognitions and general anxiety were included as mediators.

Results

GI safety behaviours, cognitions and general anxiety mediated treatment effect on the outcomes of symptom severity and work and social adjustment. Avoidance behaviour was not a significant mediator for either outcome. Sequential mediation models indicated that unhelpful GI related cognitions reduced before anxiety did, and this sequential path (R➔GI related cognitions➔anxiety➔outcome) was significant for both symptom severity (b = −0.22, CI [−0.40 to −0.90], p = .005) and work and social adjustment (b = −0.26, CI [−0.44 to −0.11], p = .003) where ‘R' is randomisation. Reduction in GI safety behaviours also preceded reduction in anxiety. This sequence (R ➔GI safety behaviours➔anxiety➔outcome) was significant for both symptom severity (b = −0.11, CI [−0.24 to −0.01], p = .049) and work and social adjustment (b = −0.12, CI [−0.23 to −0.03], p = .03).

Conclusion

Results suggest that it is important for psychological treatments to target IBS specific factors for change.

Introduction

Irritable bowel syndrome is a functional gastrointestinal disorder characterised by abdominal pain and associated changes in bowel habits [1]. The prevalence of IBS is estimated to be between 10 and 22% in the UK [2,3] with a similar global prevalence rate [4]. There are no physiological diagnostic markers for IBS and it has long been established as a ‘biopsychosocial illness’ [5,6]. As such physiological factors such as genetics or infection interact with psychological and social factors such as unhelpful gastrointestinal (GI) related cognitions, anxiety and life stress to result in bowel symptoms and abdominal pain.

Cognitive behavioural therapy (CBT) has been shown to effectively reduce symptom severity and enhance quality of life in IBS [[7], [8], [9]]. However, there is not one generic CBT model or approach in IBS. A number of CBT models for IBS have been developed [[10], [11], [12]] utilising different cognitive and behavioural strategies to target the key mechanisms identified by the respective models. The three system's model for IBS, as based on Lang's three system's model of panic disorder [13], identifies gastrointestinal (GI) related cognitions and behaviours as key factors in maintaining symptoms [12]. Therapeutic strategies such as challenging unhelpful thinking patterns and goal setting are used to change cognitions and behaviours. More recently, an interoceptive CBT model of IBS (CBT-IE) has been developed [11]. This model posits that symptoms are maintained by GI specific anxiety and hypervigilance to symptoms [11]. Exposure based techniques (interoceptive and in vivo) are used in CBT-IE to reduce GI specific anxiety and hypervigilance.

The term ‘mechanism’ refers to the processes through which treatments have effect on the desired outcome [14]. Understanding what the key mechanisms of treatment are, is important for modifying treatments so that they can more effectively target identified mechanisms of action [15]. This has the potential for improving the efficacy and efficiency of treatments. Insight into mechanisms of change in treatments can also inform the understanding of the nature of IBS. Evidence for key processes in the maintenance of IBS can be inferred from research demonstrating how treatments reduce symptom severity through change in specific mechanistic processes [14,16]. I.e. if treatment works by reducing unhelpful GI cognitions we can infer that these cognitions are likely maintaining factors of IBS.

Both three system's CBT and CBT-IE may be classified as ‘endogenous’ stress models of IBS [17]. Both identify stress-inducing processes arising specifically from IBS symptoms (e.g. GI specific cognitions, behaviours, anxiety) as key maintaining processes of IBS [17,18]. In contrast ‘exogenous’ stress models of IBS postulate that symptom-exacerbating stress arising from factors external to IBS (e.g. stress or general anxiety), are central maintaining factors of IBS [17]. While there is evidence to show that anxiety exogenous to IBS may have a causal role in the onset [19] and that those with IBS have increased stress reactivity [20,21], there is less evidence suggesting that changing such exogenous factors precede change in IBS symptoms [16]. A systematic review of treatment mechanisms in CBT for IBS suggested that across different CBT treatment protocols change in IBS-specific processes, particularly GI related cognitions, were key in reducing symptom severity and improving quality of life (QoL) [16].

Cognitive and behavioural responses have been shown to have a mechanistic role in treatment effect across CBT protocols [17,[22], [23], [24], [25]]. In CBT-IE, avoidance was shown to mediate a reduction of symptom severity [25]. Similarly, the mechanistic role of both GI related cognitions and behaviours was demonstrated in a randomised controlled trial assessing the efficacy of a three system's CBT treatment when added to an antispasmodic versus an antispasmodic alone [22]. The data from this trial is used in the present paper. The trial recruited adults from primary and secondary care sites in South London and assessed outcomes at baseline, 1.5, 3, 6 and 12 months post-randomisation [12]. CBT plus antispasmodic medication were superior to medication alone in reducing symptom severity and improving functioning at 1.5 and 3 months follow-up [12,26]. There was some waning of response at 6 and 12 months. The trajectories for both outcomes are illustrated in panel A and B in appendix C.

Previous mediation analysis using this data conducted sequential modelling in order to identify whether GI related cognitions or behaviours changed first in a sequence of change from treatment to the outcomes of symptom severity, work and social adjustment and anxiety [22]. Separate models were run for each outcome and GI related behaviours were assessed as one construct that included both avoidance behaviours (e.g. avoiding foods or situations) and safety behaviours (e.g. checking stools or access to the toilet). Change in behaviour was found to precede change in cognitions for all outcomes. However, the analysis did not include mediators measured at consecutive time points, limiting inferences about causality.

The present study sought to build on the previous study and wider research to explore in more detail the order of change during therapy. Based on the endogenous models of IBS, it is hypothesized that for change in outcomes to occur, it is primarily important for GI related cognitions and behaviours to change. Change in these IBS specific processes is hypothesized to lead to reduction in general anxiety. In contrast, exogenous models would lead us to expect that general anxiety would change prior to GI related cognitive and behavioural processes, as general anxiety (or stress) is the key target for change in such treatment approaches.

The present analyses advances the analytic strategy used previously by using consecutive time points (from baseline to 12 months) to aid inferences about causality. Although the significant treatment by group effect was lost by 6 months, it has been argued that it is important to conduct mechanistic research in the absence of significant treatment effects. This allows information to be gathered about what may have been responsible for this the lack of sustained effect. For example, whether the treatment did not target the treatment mechanisms as expected [27,28]. Given that a previous factor analysis found two factors, namely control (or 'safety behaviours') and avoidance behaviour were distinguishable, we also wanted to explore these potential mediators separately [29]. The aims of the present study were (1) to identify whether CBT produced change in symptom severity and work and social adjustment via a change in anxiety, cognitions, avoidance behaviours and/or safety behaviours (2) to identify whether mediating effects found in cognitions and/or behaviours preceded or were preceded by change in anxiety to produce change in outcomes.