Affective disorders are one of the three leading non-fatal health burdens across the globe (1). They are not only common and costly (e.g.; due to premature retirement), but also associated with considerable morbidity and mortality. As effective treatments are available (2) it is primarily the implementation of effective as well as efficient treatment methods for these disorders which poses a major challenge for service provision worldwide. Accordingly, there are continuing efforts to disseminate efficacious treatments into the diverse segments of health services. Furthermore, considering patient preferences it is important to make different types of efficacious treatments available to establish a more precise matching between individual patients and individual psychotherapies (3).

Psychotherapy has been found to be at least as efficacious as pharmacotherapy in mild to moderate depression (4) and many psychotherapeutic interventions could demonstrate their long-term efficacy (3). Two types of psychotherapy are exceptionally well-researched and have proven their efficacy based on many randomized and controlled studies: cognitive-behavioral therapy (CBT) and behavioral activation (BA). Despite their demonstrated and comparable efficacy (5) one advantage of BA might be that it is easier to understand (for patients) and, hence, to implement (into service structures): Cognitive-behavioral therapy for depression includes different components focusing both on behavioral as well as cognitive change, which at least some patients may find complex and time-consuming (6). Particularly its cognitive components can be more difficult to understand for marginalized populations (e.g., due to language, cognitive or economical barriers). On the other hand, there is strong evidence that a simple behavioral approach, behavioral activation (BA), is easy to administer (e.g., 7) and efficacious within many diverse patient populations (e.g., 8–10). Dimidjan et al. (11) showed in an RCT a significant amount of extreme non-responders in the CBT study arm but not in the BA. Furthermore, Jahoda et al. (12) conducted a RCT with intellectual disabilities who showed a substantial decline in depressive symptoms after BA treatment. The main focus of BA lies on activity monitoring and activity planning (13), which both aim at directly counteracting depressogenic factors such as lack of reinforcement, extinction of instrumental behavior, and excesses of aversive control. Recent research has shown that brief activity planning instructions alone can have relevant effects for emotion regulation and prevention (14). Given the clear focus and the relative simplicity of BA, we assume that it can be a viable alternative to CBT in many settings, especially when large groups of patients need to be treated in relatively condensed ways, as is the case in inpatient psychosomatic rehabilitative treatment (15).

The positive effects of CBT and BA mentioned above seem to be observable across different modalities such as individual vs. group treatment (16, 17). Group psychotherapy in particular constitutes a competing alternative to individual psychotherapy as it makes psychotherapy accessible for many patients and is clearly more economical (18). Furthermore, it is the most viable form of psychotherapy in inpatient settings, and in countries such as the UK, the USA, and Germany, group psychotherapy is the most prevalent modality of inpatient psychological care (19–21).However, as countries such as the USA are following the trend of outsourcing psychological treatment away from psychiatric wards, there is less data on inpatient than on outpatient treatment (22). This is a shortcoming of high clinical relevance, as it has been shown that inpatients do not represent the same target population as outpatients as they demonstrate more severe symptomatology, are often chronically ill, and diagnosed with higher number of comorbidities (22).Hence, the effects observed in outpatient settings may or may not be valid in inpatient contexts as well. The need of independent research on that subject, especially regarding group therapy, is obvious.

Another research requirement is the more thorough study not just of patients’ symptom reduction but also of their improvement in quality of life, functioning and participation within society (23). In order to do so, clinical research as well as clinical decision-making regarding mental health disorders must consider level and pattern of functioning (24). As a system for providing data on disability-related functioning and non-fatal health outcomes, the World Health Organization (WHO) proposed the International Classification of Functioning, Disability and Health (ICF, 25). ICF acts complementary to ICD (see ICF, 2001) and will be consistently applied in this study. The terms ICF, functioning, and disability-related functioning and non-fatal health outcomes will be used interchangeably.

Summarizing, the aim of the present study was to compare BA to cognitive-behavioral therapy, embedded in a group inpatient treatment of depressed patients with a high number of comorbidities. Firstly, we expect patients in both treatment modalities to reduce depressive symptoms with BA exhibiting a significantly larger reduction. Furthermore, we expect patients with lower education level to show steeper decline than those with higher one. Secondly, we expect both groups BA and CBT to increase in activation, especially patients with lower education level in the BA group. Thirdly, participation/impairment: We expect both groups BA and CBT to show a decrease in social and work impairment.

Given the context of a multidimensional inpatient treatment for chronic mental disorders, which focuses on life and work participation, we also examine the changes in terms of the ICF (25).

The study was conducted in the naturalistic setting of two psychosomatic rehabilitation clinics but based on the admission date of the individual patient, quasi-randomization of patients to BA vs. CBT could be established.

In the present quasi randomized-controlled study, the effects of behavioral activation in a group setting versus cognitive-behavioral group therapy were compared as nested treatment components within a psychosomatic rehabilitation treatment in Germany. Given the large sample and the robust statistical analyses, the study significantly contributes to the rather scarce pool of literature on inpatient group psychotherapy. In line with Folke, Hursti (9), it shows that inpatient psychotherapy is beneficial for the treatment of depression, which is of relevant knowledge as it refers to a clientele for which treatment success was not deemed promising in the outpatient setting (e.g., multi-morbid, belongs to marginalized social groups, or is chronically ill).

Our results showed large within-group effect sizes in the decline of depressive symptoms in both treatment groups of about d = 1.24 at follow-up. Furthermore, there were highly favorable response and remission rates in CBT and BA (53.7% and 54.2%, respectively). These effects are among the largest when compared to previous reports on depression treatment in rehabilitation (50) and these results clearly demonstrate the effectiveness of depression treatment within the context of psychosomatic rehabilitation. Compared to outpatient treatment, inpatient treatment is of higher intensity, but also more time-condensed in terms of the duration of stay, which is limited to 5 weeks and can be prolonged to up to 7 weeks. Our results demonstrate that such short-term inpatient psychotherapy programs can be highly effective.

With regards to the comparison of BA vs. CBT, no differences were found between both treatments, neither in the change of depressive symptoms nor in the change of any other main outcomes. For example, activity levels increased throughout the stay, independent of the form of group treatment applied. These findings are in line with a recent meta-analysis, in which BA compared to an active control group yielded only marginal effect sizes in reduction of depression symptoms as well as increase in activation (51). In line with previous research (52) both treatment modalities showed significant improvement in measures of functional impairment ICF, which, however, did not significantly differ between treatments. This finding does not surprise, as psychosomatic rehabilitations in Germany are packaged treatments that are fine-tuned towards individual assessment and rehabilitation of functional impairments. Differences between BA and CBT regarding ICF were highly improbable, and we did not expect them. Furthermore, patients’ level of education did not moderate the outcome variables. Contrary to our hypothesis, patients with a low education level did not benefit more than other patients from the BA component. However, patients with low education showed higher avoidance rates, and social as well as work-related impairments. Interestingly, these findings did not correlate with depression scores as we did not find education level to significantly predict either of the depression measures. Contrary to previous research (53–55), patients with lower education status did not appear to be more depressed. However, psychosomatic rehabilitation clinics not only offer inpatient psychotherapeutic treatment, but they also conduct psychosocial assessment. The results of psychosocial assessment can then lead to measures for further outpatient occupational rehabilitation or even early retirement. Since persons with lower education level (LEL) tend to work in occupations with higher exposure to risks and tend to discharge from work life earlier (54), the above-mentioned results could mirror the fact that for persons with LEL, psychosomatic inpatient psychotherapeutic treatment is as important as psychosocial assessment and occupational therapy.

Overall, the findings suggest that using BA in the context of psychosomatic rehabilitation does not have any add-on effects on treatment outcome. Neither is it inferior to CBT in the present sample. Hence, the evidence of the present study leaves it open which of the two approaches should be preferred in treatment. As both have the same average effects, therapists as well as patients can freely decide which of the two approaches is more convenient and convincing for them. Even if these “weaker” reasons for choosing a treatment approach may not moderate the treatment outcome, they should not be neglected as they are central for what is perceived as a “quality treatment” from the patients’ perspective (56). For many patients, it is of importance how complex, feasible, cost-extensive, comprehensible, fast etc. a treatment is, independent of the outcome (56). Against the backdrop of the advances in personalized medicine, research on psychotherapy has started examining potential of patient-centered treatment. Making a larger variety of treatment modalities available, that can be then tailored towards individual patient’s needs, could lead to better outcomes in the future (57–60). Such development will mirror multi-optional treatment in other medical fields, where patients can e.g., express their wish for or against adjunctive pharmacotherapy. Accommodating individuals’ preferences for psychotherapy type could further catalyze the treatment due to a better fit between patient and intervention(s).

At this point, we would like to elaborate on the differences between inpatient psychosomatic rehabilitation and outpatient psychotherapy in Germany. Public insurance companies in Germany reimburse outpatient psychotherapeutic treatment (behavioral therapy), independent of the clinical diagnosis, with up to 60 sessions per case. However, in the practice, the median therapy length is 15,9 months (SD=0.64) and the median number of sessions is 40 (SD= 2 sessions) sessions (1). Since the therapeutic sessions occur on average 1x week, interventions can be precisely tailored towards the demands of the person´s natural environment. On the other hand, inpatient psychosomatic rehabilitation lasts 5 weeks, occasionally up to 7 weeks. This is a very condensed and intense therapeutic treatment package. In particular, it can be very effective for patients that would not be able to make a progress when treated in outpatient setting e.g. due to the severity of the symptoms (2). The challenge than is the transfer and application of newly acquired competencies and skills in the broader contexts of everyday life (e.g. at home, at work etc., 2).

In this respect, it must be emphasized that, behavioral activation is an especially simple and parsimonious method, that is easier to understand for cognitively impaired patients (61). Moreover, it does not require sophisticated therapeutic skills that take years of training (for the therapists of this study applying BA was new). The present study demonstrates that it can be easily integrated into an established and complex treatment program without any negative consequences for the overall effectiveness.

Although specifically targeted, the inpatient setting of this study limits the interpretation of the results. All patients participated in an extensive inpatient treatment program of approximately 25 hrs/week (e.g., occupational therapy, sports, etc.), of which group therapy lasted only 180 min/week. Hence, we could not dismantle the effects of the group therapy modality by itself. Our results should rather be interpreted as a comparison of two treatment packages of which one includes BA and the other CBT. Future studies should direct their attention towards this aspect. Also, the pre-registered sample size of N= 500 could not be achieved due to restrictions of the COVID-19 pandemic. Since we did not conceptualize this study as an equivalence trial, we did not set a margin to which we would compare the difference in treatment effects, hence we cannot conclude whether these two methods (BA vs. CBT) are equivalent. However, visual inspection of the mean scores showed no substantial differences between the treatment methods, and CIs overlap to a great extent, which is in line with previous research (7). Due to the design of the study, we cannot rule out other factors that might have played a role, such as other treatments (e.g. physical exercise 62), medication, and social environment of the hospital (63). As group therapy is only one aspect of the treatment package within inpatient care, a substantially larger sample or a completely different design (e.g., one arm of the study without any group treatment) would be necessary to detect a possibly small effect. Given the great overlap of the CIs, future researcher should also target the comparison between BA and CBT in terms of an equivalence study.

A further limitation is the number of videotaped sessions of the hospital 1 (8%) that could be used for the adherence check. We chose to follow the well-researched procedure suggested by Dennhag, Gibbons (48). Due to its strictness (e.g., only group sessions led by one therapist, etc.), a large number of group sessions had to be excluded. Considering the context in which this procedure was developed, outpatient addiction treatment in one-to-one setting, it is understandable that such rules have to be followed. However, in inpatient setting of psychosomatic rehabilitation, substitutes lead sessions that would be canceled in outpatient setting. Hence, groups are bound to experience more than one therapist. On the other hand, we can speculate that due to close proximity of team members, and regular meetings such as supervisions, head physician’s weekly rounds etc., adherence can be monitored and maintained better than in an outpatient setting despite sessions being substituted. Since there are no studies focusing on the number of sessions per therapist to check adherence of inpatient group therapies, we decided to choose a rather conservative rule suggested above. However, this is a research gap of clinical relevance that future studies should aim at.

Despite our great effort in contacting patients multiple times by multiple means (e.g. by phone, E-mail, mail), we could not prevent high dropout rates at follow up. This is in line with previous research conducted in inpatient settings. Particularly in psychosomatic rehabilitation, response rates have been found to be very low (50). Nevertheless, this reduces the chance to detect differences between the treatments. As the allocation was quasi-randomized, however, neither the researchers nor the patients could influence drop-out rates in any systematic way. Also, limited ethnical and cultural variability of the sample prevents this study from generalizing the results to other ethnical and cultural contexts. Additionally, use of self-reports for testing the main criteria poses a limitation of this study as there are typical risks of biases (e.g., response bias) inherent in the self-report approach. Nevertheless, it has economical and practical advantages for standard use in the inpatient setting and the expert rating measure (QIDS-C) that was conducted at follow up (T4) mirrored the results on the self-report level. Having to interview patients multiple times throughout the study might have additionally reduced the number of participants willing to partake.

Summarizing, this large data set demonstrates behavioral activation to provide a good alternative to cognitive-behavioral therapy in groups especially in fields with a chronically ill, multi-morbid clientele.

The original contributions presented in the study are included in the article/ Supplementary Materials . Further inquiries can be directed to the corresponding author.

The studies involving humans were approved by the local ethics committee of the Technische Universität Dresden (Germany; EK 327082018). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

UM and TS collected and participated in data preparation. TS overseen and conducted the manual adherence measurement. UM wrote the code and analyzed the data. VK and JH, as senior authors, together with UM conceptualized the study and wrote the manuscript. All authors extensively reviewed the manuscript, supplementary materials, and all tables and graphs. All authors contributed to the article and approved the submitted version.