2. Materials and Methods
This secondary exploratory analysis was based on the RCT, which found that HT was non-inferior to CBT regarding the percentage symptom improvement after six months in patients with a depressive episode (Fuhr et al., 2017, 2021). The RCT followed the CONSORT and GCP guidelines and was conducted in accordance with the Declaration of Helsinki. Ethical approval was obtained by the Ethics Committee of the University Hospital Tuebingen (061/2015BO2, on 11 March 2015). The RCT was pre-registered (NCT02375308, ClinicalTrials.gov).
2.1. Participants
Patients with a current mild-to-moderate depressive episode confirmed with a diagnostic interview based on the Diagnostic and Statistical Manual of Mental Disorders—Fifth Edition (DSM-5, APA, 2013) were included in the RCT. In cases of concomitant ADM, a stable course of medication for the last three months or longer was necessary for inclusion. Patients with a lifetime diagnosis of bipolar disorder or schizophrenia, a diagnosis of chronic major depression, actual high suicidality, or other severe primary psychological disorders (active alcohol or drug dependence, actual post-traumatic stress disorder, or anorexia nervosa), or patients receiving outpatient psychotherapy within the last twelve months were excluded from the trial.
Overall, 152 patients were randomized 1:1 with blocks of 40 to 16–20 individual sessions of either HT or CBT. The treatment allocation was concealed from patients and therapists until just before the first treatment session. Most of the patients were female (n = 100, 65.8%), and 114 (75.0%) had a high-school degree. Mean age was M = 39.5 (SD = 14.6). Fifty-six (36.8%) of the patients were using concomitant ADM (dummy-coded yes/no), with mostly SSRIs being prescribed (41.1%). For the sample characteristics per intervention condition, see Table 1. Regarding the CONSORT flow and the different types of ADM, see Fuhr et al. (2021).
2.2. Interventions
Patients received individual outpatient psychotherapy in 16–20 sessions during a six-month period with a therapist who was specifically licensed and trained in the respective intervention. The CBT manual was based on the German treatment manual for CBT (Hautzinger, 2013). CBT included modules on psychoeducation, behavioral activation, cognitive restructuring and meta-cognitive strategies but also interpersonal competence and training problem-solving skills using different tasks, as well as homework. In terms of relaxation techniques, only progressive muscle relaxation was allowed in CBT. The HT manual was based on a collection of different modules published in German (Wilhelm-Goessling et al., 2020). The HT treatment was based on the psychodynamic model of the ego-structure by Rudolf (2013). Hypnotic elements should be more directive with patients with a lower ego-structure, while patients with a higher ego-structure might benefit from open techniques and suggestions. HT included modules on resource activation, processing and integrating positive and negative experiences from the past, and the development of positive solution imagery. Hypnotic inductions were conducted individually by therapists with their patients. Patients were encouraged to practice self-hypnosis outside the sessions using self-hypnosis skills that were developed with the therapist. Alternatively, the hypnotic trances were recorded by each patient to practice at home. Details regarding the training of therapists and manual adherence in the RCT have been described previously (Fuhr et al., 2021). All therapists had at least three years of experience with the respective treatment method and were trained in one (CBT) or two (HT) days. Supervisions were held monthly with an expert in the field and a treatment manual. To allow for the tailoring of both treatments, all therapists were allowed to select the sequence of treatment modules according to the patient’s needs rather than following a rigid protocol. To assure manual adherence by the therapists, four external raters were involved in assessing the treatment fidelity. Results are published in Fuhr et al. (2021). There was a certain “overlap” in the CBT and HT treatment, for example, regarding the current symptoms and topics, individual backgrounds of patients, and aspects of motivational interviewing. CBT covered psychoeducation, behavioral activation, cognitive techniques, concrete plans, agenda setting, homework, and SORC analysis. HT techniques included resource activation, for example, via a safe place or imagining a resourceful animal, using direct or indirect trance inductions, association or dissociation techniques, seeding, time progression and regression, posthypnotic suggestions, externalizing symptoms, using ideomotor signals, and utilizing the specific symbols and wording of the patient. The treatment fidelity items are available in German upon request from the authors.
2.3. Outcome Measure
The Montgomery–Åsberg Depression Rating Scale (MADRS, Montgomery & Asberg, 1979) was used. Depressive symptoms were assessed with the clinician-rated MADRS at baseline and after treatment. The MADRS shows a high clinician interrater reliability (Montgomery & Asberg, 1979) as well as external validity with other depression and a high sensitivity to change (Carmody et al., 2006; Davidson et al., 1986). Raters were blinded regarding the intervention of the patient. The MADRS score after treatment was used as the outcome variable for all analyses.
2.4. Sociodemographic Variables
Variables such as gender, age, and educational level were assessed via self-report at baseline.
2.5. Depression-Related Variables
The duration of the current depressive episode, the lifetime history of previous episodes and the presence of comorbid mental health conditions (yes/no) was assessed at baseline in a clinician-based interview (First et al., 1997).
2.6. Psychological Variables
The global functional level was measured with the 12-item version of the WHO Disability Assessment Schedule (WHODAS, WHO, 2012), which has good external validity and internal consistency within Cronbach’s α = 0.83–0.92 (Federici et al., 2017; Axelsson et al., 2017):
Quality of life concerning physical health, psychological health, social relationships, and environment was assessed with 26 items with the WHO Quality of Life (WHOQOL-BREF, The WHOQOL Group, 1998). The WHOQOL has acceptable internal consistencies (Cronbach’s α = 0.82–0.95) and a good construct validity (Bonomi et al., 2000).
The behavioral activation levels of patients were measured with the Behavioral Activation for Depression Scale (BADS, Kanter et al., 2007) with 25 items, reporting the total scale but also the four subscales of activation, avoidance/rumination, work/school impairment, and social impairment. The BADS has good internal consistency and construct validity (Kanter et al., 2009).
Self-efficacy was assessed with the ten-item General Self-Efficacy Scale (GSES, Schwarzer & Jerusalem, 1995), which shows good external validity among different cultures (Luszczynska et al., 2005).
Hypnotic susceptibility was assessed with the Harvard Group Scale of Hypnotic Susceptibility (HGSHS, Shor & Orne, 1962) in an individual session for each patient with an audio-tape-recorded German version (Bongartz, 1985) followed by ten self-reported questions.
Despite random allocation to the treatment, we asked patients before randomization to rate their expectations regarding the efficacy of both treatments separately with a visual analogous scale (0–100).
Characteristics of the total sample and for each treatment condition separately are displayed in Table 1.
2.7. Statistical Analysis
Statistical analyses were performed with IBM SPSS Statistics Versions 27 and 28.
The analyses were conducted with all randomized patients (N = 152) based on an intention-to-treat (ITT) analysis with the post-treatment depression severity score.
Missing data in the primary outcome MADRS occurred in 18 patients, either due to a declination to participate after randomization (n = 1), a discontinuation of the study treatment before 16 sessions (n = 10), or not being available for the outcome assessment (n = 7). After ensuring that the missing data of the post-treatment MADRS scores were at random with the Little MCAR test, we replaced missing data with the multiple imputations method based on the linear regression imputation algorithm automatically generated by SPSS. The pooled descriptive data of the five imputations are reported and were used for the subsequent analyses. Per-protocol sensitivity analysis regarding the primary outcome can be found in Fuhr et al. (2021).
Missing data also occurred in some sociodemographic, depression-related, and psychological variables at baseline. Four patients had up to two missing items in the GSES and one patient had one missing item in the WHODAS in the paper–pencil questionnaires. Those were replaced by the median to calculate the sum scores. If more than a 20% rate of missing data occurred, the number of participants was reduced for subsequent analyses.
To select variables for the regression analyses, bivariate Pearson’s or point-biserial correlations were first calculated between sociodemographic, depression-related, and psychological variables with the outcome MADRS after treatment. Correlation analyses were conducted in the total sample, as well as separately for HT and CBT patients. After assuring normal distribution, linearity, homoscedasticity, and checking for multicollinearity (r < 0.7), hierarchical and stepwise regression models were tested. In a first block, the symptom severity at baseline (assessed with MADRS) was entered as a variable. In a second block, the variables identified as potential predictors based on the regression analyses (p > 0.05) were included in a stepwise manner. Regression analyses were also calculated with the MADRS symptoms at baseline as an additional variable (not included as an extra block) to compare the relevance of the effects of symptom severity at baseline on the outcome.
To select potential moderators, variables which showed a significant correlation with the outcome in either HT or CBT were included in a moderator analysis with the PROCESS macro for SPSS based on Hayes (2018)’s method. The factor intervention (dummy-coded as HT = 0, CBT = 1) served as the independent variable, and the selected variables were included as the moderator variables. Moderator analyses were run separately for each variable.
CC BY 4.0. "Predictors and Moderators of Post-Treatment Depression Severity After Hypnotherapy and Cognitive Behavioral Therapy in an RCT for Mild-to-Moderate Depression" by Brockmann AL, Pfeiffer L, Batra A et al., Behavioral sciences (Basel, Switzerland) (2026). https://doi.org/10.3390/bs16091584. Licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Peer-reviewed and open access; full text from PubMed Central (PMC13603658). Headings and layout adapted for reading; figures and tables are in the original.