Effectiveness of Disease-Specific Cognitive Behavioral Therapy on Anxiety, Depression, and Quality of Life in Youth With Inflammatory Bowel Disease

A Randomized Controlled Trial

Luuk Stapersma, MSC; Gertrude van den Brink, MD; Jan van der Ende, MSC; Eva M. Szigethy, MD, PHD; Ruud Beukers, MD, PHD; Thea A. Korpershoek, MANP; Sabine D. M. Theuns-Valks, MD; Manon H. J. Hillegers, MD, PHD; Johanna C. Escher, MD, PHD; Elisabeth M. W. J. Utens, PHD


J Pediatr Psychol. 2018;43(9):967-980. 

In This Article


The current study, which had very low attrition (< 3%), tested the effect of a disease-specific CBT compared with CAU in reducing subclinical anxiety and/or depressive symptoms and in improving HRQOL, in adolescents and young adults with IBD. At the immediate post(-treatment) assessment, disease-specific CBT added to standard medical care did not perform better than standard medical care. Overall, both the PASCET-PI and CAU group significantly improved over time, on all three outcomes, 3 months after baseline (i.e., at the immediate post[-treatment] assessment). Furthermore, in subgroup analyses, we did not find indications for differences between age-groups, boys versus girls, nor between CD and UC/IBD-U regarding the effect of the PASCET-PI on anxiety, depression, or HRQOL.

Our results are in contrast to results of earlier trials with positive findings of CBT treatment for youth with IBD (Szigethy et al. 2014) but are in accordance with some of the evidence from studies in adults with IBD (McCombie et al. 2013). There are several explanations for our findings.

First, just by participating in the study, patients in the CAU group did not exactly receive standard medical care. They were psychologically assessed at two points in time with questionnaires and interviews. This is not done in routine practice, and therefore, it provided additional exposure to attention from professionals. Usually, only if psychological problems are obvious, the medical team refers patients for mental health care. CAU was chosen as the comparison condition, because it resembles the current care for youth with IBD in our institute best. However, mere participation in the trial may have had a positive effect on all patients due to increased awareness and (unintended) psychoeducation. It has been described before that merely answering questions or participating in a trial can influence behavior or emotions. For example, McCambridge (2015) recently described that the "question-behavior effect" can occur in randomized trials. Moreover, Arrindell (2001) has described the re-test effect: In patients with psychiatric problems, mean scores of psychopathology often decrease at follow-up (without any formal intervention). A first assessment can heighten awareness of anxious or depressive symptoms, which can cause a respondent to try to deal with these symptoms (by talking more about it or try to think different) or lead to more introspection or self-monitoring (Arrindell, 2001). The awareness caused by receiving information about the study and receiving the psychological assessment may have contributed to the fact that all patients improved. It can be perceived as some form of support, like in the control conditions of earlier trials in youth with IBD (Levy et al., 2016; Szigethy et al., 2014).

Second, the overall patient group had a low disease burden, both psychologically as well as somatically. Included patients experienced only subclinical anxiety and/or depressive symptoms, as randomization was not ethical for patients with clinical mental health disorders. We mainly included patients in clinical remission, because for patients with severe disease activity, adherence to the CBT protocol might have been complex. For the subclinical anxiety and/or depressive symptoms, mere participation may have been enough to improve. This raises the question: Which IBD patients should receive psychological treatment? When we analyzed those patients with the highest levels of subclinical anxiety and/or depressive symptoms, still no differences between CBT and CAU were observed. However, a recent trial showed a significant effect of CBT compared with a waiting list on QOL, anxiety, and depression in adult IBD patients, of whom 70% met criteria for a psychiatric disorder(Bennebroek Evertsz' et al., 2017). This implies that IBD patients with severe psychological problems can actually benefit from CBT. Furthermore, for adolescents with IBD, when compared with supportive therapy, CBT has been shown to improve somatic depressive symptoms as well as clinical disease activity and erythrocyte sedimentation rate, but only in patients with CD and moderate clinical disease activity (Szigethy et al., 2015). This suggests that patients with active disease can benefit from CBT. For these patients, however, sessions should be delivered with great flexibility, as they may not be able to adhere to weekly "live" sessions.

Third, the PASCET-PI may not be suited enough to improve subclinical anxiety and/or depressive symptoms. However, an earlier study using the original PASCET-PI protocol in a group of patients selected on elevated depression did find an effect on these subclinical depressive symptoms, and also on comorbid anxiety disorders (Szigethy et al., 2007). In our trial, patients experienced more anxiety symptoms than depressive symptoms, which may have influenced the results. Nevertheless, CBT is the most evidence-based psychological therapy for both anxiety and depression (Compton et al., 2004). In general, CBT techniques do have an effect on both anxiety and depressive symptoms, with even higher effect sizes found for anxiety than for depression (Weisz et al., 2017). This implies that PASCET-PI may be effective, as to both anxiety and depressive symptoms. In adults with IBD, mixed results are found with respect to the effectiveness of CBT on psychological as well as somatic symptoms (Gracie et al., 2017; McCombie, Mulder, & Gearry, 2013). Several recommendations are made (McCombie et al., 2013; Mikocka-Walus et al., 2016) to focus on patients with, for example, decreased HRQOL or experiencing psychological problems and to take into account high attrition rates in power and sample size calculations. In our trial, these recommendations were covered by selecting patients on anxiety and/or depression and by having very low attrition. The mixed findings in IBD are consistent with mixed findings on the effect of preventive CBT programs for subclinical anxiety and/or depression in youth (Bennett et al., 2015). As our patients experienced subclinical psychological and somatic symptoms, the treatment can be considered as preventive (for the development of clinical disorders). Further studies are needed to examine this type of preventive effects, especially in patients with IBD, as psychological problems can also affect the disease course (Alexakis, Kumar, Saxena, & Pollok, 2017; Van Tilburg et al., 2017).

Fourth, although a sample size of 70 participants should be large enough for the expected effect sizes for CBT on anxiety and/or depression, perhaps we would have found a significant group difference with a larger sample size. Originally, to take into account possible attrition, we aimed to enroll 100 patients (Van den Brink & Stapersma et al., 2016), which we could not achieve. Revised power calculations still indicated that we had sufficient power to investigate the effect of the PASCET-PI, using n = 70. With this sample size, one would expect to see at least a trend toward a difference between the two groups, but this was not the case. Moreover, compared with earlier trials, a strength of the present study was the very low attrition rate and that almost all (95%) patients completed disease-specific CBT.

Fifth, it may be possible that the effect of the PASCET-PI sustains on the longer-term, whereas the effect of the control group diminishes over time. The course of IBD can be fluctuating, and perhaps the knowledge and skills taught in the PASCET-PI can be more useful when patients suffer from more disease activity or flares during a longer period of follow-up. Patients themselves often expressed that this was a motivation to participate in the therapy ("I have no complaints now, but the CBT skills can be useful in the future, when I have a flare"). Data on longer follow-up assessments will be available for analyses later.

In summary, strengths of the current study are that we included patients with a broad and clinical relevant age, with both anxiety and/or depressive symptoms, and that our study had very low attrition. Moreover, no patients in the control group sought mental health care. Furthermore, as our study sites encompass both rural and urban hospitals, this strengthens the generalizability and external validity of our findings. Although the age-specific instruments were most appropriate for the patients in our study, statistically it was a limitation that using different instruments made it difficult to combine all patients in one analysis and that we could perform the linear mixed models only in subgroups. Originally, the study was sufficiently powered to analyze mean symptom change of anxiety and depression. Due to the fact that finally multiple instruments had to be used to cover the age range, this was not possible. However, a revised power calculation for the chi-square analyses with the reliable change index indicated that we had enough power with the total of 70 patients in the RCT. Another limitation was the relatively small sample size. Therefore, our results should be interpreted with caution. We recommend screening for anxiety and/or depressive symptoms in youth with IBD, as these symptoms can affect the disease course (Alexakis et al., 2017; Van Tilburg et al., 2017) and HRQOL (Engelmann et al., 2015). Subclinical symptoms may develop into more severe psychological disorders, which even have a greater impact (Beesdo, Knappe, & Pine, 2009; Copeland, Shanahan, Costello, & Angold, 2009). CBT may be more effective in patients with more severe psychological symptoms or more IBD disease activity. This, however, should be examined in studies with a different design (i.e., not with standard medical care as the comparison condition). Based on our clinical experience, we consider PASCET-PI as suited also for patients with more severe IBD symptoms but with great flexibility in delivery (over the phone or in the hospital when patients are hospitalized). Yet, future research is needed to find out how the PASCET-PI or CBT can be best delivered to those patients, which patients with IBD benefit most from psychological treatment, but also how the long-term course of disease activity is associated to the long-term course of anxiety/depression.

In conclusion, in our RCT, all patients improved in their symptoms of anxiety and depression and their HRQOL over time (3 months). At the immediate posttreatment assessment, we found no additional effect for a disease-specific CBT on improving subclinical anxiety and depressive symptoms or HRQOL in adolescents and young adults with IBD, when compared with CAU. We hypothesize that the awareness the study elicited and the possible (unintended educational) support provided may have had a strong positive effect on all patients. CBT could be beneficial for patients with more severe psychological symptoms or IBD patients with clinical disease activity.