The specific effect of systematic exposure in irritable bowel syndrome: complier average causal effect analysis using growth mixture modeling

H. Hesser, E. Hedman, P. Lindfors, E. Andersson, B. Ljótsson

We reanalyzed data from a previously published randomized component study that aimed to test the incremental effect of systematic exposure in an internet-delivered cognitive behavioral treatment (ICBT) for irritable bowel syndrome (IBS). Three hundred and nine individuals with IBS were randomly assigned to either the full treatment protocol (experimental condition) or the same treatment protocol without systematic exposure (control). Participants were assessed weekly for IBS symptoms over the active treatment phase. We used a complier average causal effect (CACE) analysis, in the growth mixture modeling framework, to (1) examine the specific effect of exposure among those who received the intervention (i.e. compliers), and (2) explore the associations of pre-treatment patient characteristics with compliance status and outcome changes. Fifty-five per cent of those assigned to the experimental condition were classified as compliers. The CACE analysis that took into account compliance status demonstrated that the magnitude of the incremental effect of systematic exposure on IBS symptoms was larger than the effect observed in an intention-to-treat analysis that ignored compliance status (d = 0.81 v. d = 0.44). Patients with university education showed more improvement during the exposure phase of the treatment. Pre-treatment patient characteristics did not predict compliance status. The effect of systematic exposure on IBS symptoms is of substantial magnitude among those individuals who actually receive the intervention (CACE). Studying the subsample of individuals who discontinue treatment prematurely and tailoring interventions to improve compliance may increase overall improvement rates in ICBT for IBS.