Introduction
The successful results of individual cognitive behavioural therapy (CBT) in the treatment of moderate to severe depressions have been well-documented over the years (e.g. Rush et al. Reference Rush, Beck, Kovacs and Hollon1977; Beck et al. Reference Beck, Rush, Shaw and Emery1979; Rush & Giles, Reference Rush, Giles DE and Rush1982; Dobson, Reference Dobson1989; Freeman et al. Reference Freeman, Simon, Beutler and Arkowitz1989; Kragh-Sørensen et al. Reference Kragh-Sørensen, Gjerris and Bolwig1991; Paykel, Reference Paykel1992; Haaga & Beck, Reference Haaga, Beck and Paykel1992; Beck, Reference Beck1995). Even though there is less documentation of CBT group interventions, there are also studies that indicate promising results. Therefore it is quite surprising to find that the number of published studies about group CBT for depression as such is still so sparse. In fact, no prior studies of 12-session group CBT for adults who were inpatients and receiving antidepressant medication could be found. Although having different aims and subjects, Furlong & Oei (Reference Furlong and Oei2002), Lockwood et al. (Reference Lockwood, Page and Conroy-Hiller2004) and Shaffer (Reference Shaffer1981) were some of the most comparable studies regarding severity of depressive complaints, number of sessions and treatment outcome measures. Furlong & Oei (Reference Furlong and Oei2002) had a 12-session group CBT programme showing significant effect sizes [Beck Depression Inventory-II (BDI-II) pre-treatment mean: 20.33, s.d. = 7.62; BDI-II post-treatment mean: 10.72, s.d. = 9.46, p<0.01]. However, the treatment was offered to an outpatient population, with only some patients receiving antidepressant medication, which can be a confounder, and their main focus was on trying to clarify the role of cognitive change in depressive symptoms. Likewise Lockwood et al. (Reference Lockwood, Page and Conroy-Hiller2004) had 12-week programme with follow-up evaluation at 3, 6 and 12 months, showing significant decreases in BDI-II score (BDI-II pre-treatment mean: 29.0, s.d. = 5.9; BDI-II post-treatment mean: 6.2, s.d. = 4.9, p<0.05), but they also included adolescents. Shaffer (Reference Shaffer1981) had a CBT group programme of 10 weekly sessions and significant reduction in depressive symptoms was found between the pre- and post-treatment BDI-II ratings (t = 2.91, df = 7, p<0.02), but patients receiving antidepressant medication were excluded and the main aim of the study was to compare CBT group treatment with individual treatment as well as CBT with interpersonal group therapy rather than examining the effects of group CBT per se. A broad literature search also revealed that of the few published studies, several concerned rather specific segments such as group CBT of elderly outpatients (Steuer & Hammen, Reference Steuer and Hammen1983), of youngsters with abuse (Curry et al. Reference Curry, Wells, Lochman, Craighead and Nagy2001), underprivileged groups (Satterfield, Reference Satterfield1998), and postnatal depressions in rural Australia (Craig et al. Reference Craig, Judd and Hodgins2005), which must present a challenge for generalizability. The sample provided in this study is representative of age and comorbidity for the inpatient population with major depression in Copenhagen.
In 2002 a group of psychologists and psychiatrists working in different psychiatric departments in the Copenhagen region began to work on a pilot project with the goal of establishing a short term, manualized CBT group intervention for depression in psychiatry, using the manual developed for the project (Bramsen et al. Reference Bramsen, Due Madsen, Holton and Øibakken2006). The manual builds upon Fennell's approach to the treatment of depression (Fennell, Reference Fennell, Hawton, Salkovskis and Kirk1999, Reference Fennell, Gelder, Andreasen and Lopez-Ibot2000; Fennell et al. Reference Fennell, Westbrook, Bennett-Levy, Bennett-Levy, Butler, Fennell, Hackmann, Mueller and Westbrook2004), as well as the behavioural work of Lewinsohn and his team (Lewinsohn, Reference Lewinsohn, Friedman and Katz1974; MacPhillamy & Lewinsohn, Reference MacPhillamy and Lewinsohn1976, Reference MacPhillamy and Lewinsohn1982; Lewinsohn & Amenson, Reference Lewinsohn and Amenson1978; Lewinsohn & Talkington, Reference Lewinsohn and Talkington1979; Lewinsohn et al. Reference Lewinsohn, Youngren, Grosscup and Depue1979, Reference Lewinsohn, Sullivan, Grosscup and Rush1982, Reference Lewinsohn and Talkington1983/Reference Lewinsohn, Melmelstein, Alexander and MacPhillamy1985; Grosscup & Lewinsohn, Reference Grosscup and Lewinsohn1980).
In April 2004 a Specialized Treatment Unit for Depression and Anxiety opened at the Psychiatric Centre Copenhagen, at Bispebjerg Hospital, where the author is the clinical psychologist. In this setting it was natural and relevant to establish CBT treatment groups for inpatients and to join the project.
The quality of treatment was ensured by workshops and group supervision by the authors of the manual and a workshop with Melanie Fennell.
A short article on the preliminary results of the first four groups was published in the monthly bulletin of the Danish Psychologists’ Association in 2006 (Nielsen, Reference Nielsen2006). The manual has now been published as a book (Due Madsen et al. Reference Due Madsen, Bramsen, Holton and Øibakken2008).
Aims
The aim of this study is to evaluate treatment outcome (measured in mean decrease in BDI-II scores using paired-samples t tests) and relapse rates (measured in total number of readmissions for participants from termination of the 12-session treatment programme and up to 1 year after termination of treatment, compared to the expected readmission rate seen in a representative registry study). Given that the subjects in this study at the time of referral were an inpatient population, they were likely to have been experiencing moderate to severe depression and moderate to severe cognitive impairment, so a 12-session programme might prove to be too brief. On the other hand, with the help of the structure of the ward and the staff between sessions, patients were expected to be able learn and apply at least some of the CBT elements to help them cope with their symptoms and prevent relapse.
Method
Target group and inclusion criteria
The group treatment was offered primarily to inpatients admitted to the ward (45/48, while three were referred from outpatient units) between 2004 and 2010 with a depressive episode or recurrent depressive episodes of at least 2 weeks and up to 2 years and a severity level corresponding to a Hamilton Depression Rating Scale score (17 items) of 18–25 (moderate to severe). The Hamilton score was used for selection purposes only, to measure the severity of depression before the beginning of treatment and exclude those cases that were too severe, whereas BDI-II was used as an integrated part of the homework, so patients could monitor and report their symptoms, as well as an outcome measure. Post-treatment BDI-II scores and readmission rates were also used to justify the allocation of resources to the CBT group programme over the 5-year period. Major depression must be the primary problem, which in principle excluded patients diagnosed with bipolar disorder, severe personality disorders like BPD, severe physical illness, organic brain disorders and substance abuse. Participation also required good Danish-language skills, both spoken and written.
Recruitment
Out of the 62 referrals received during the 5 years in which group therapy took place, 48 participants (11 men, 37 women) aged between 20 and 67 years (mean age 37.5 years) were included in one of 10 groups, each comprising 4–6 participants. Of these, 46 finished the 12-session weekly programme. Most participants in the groups were inpatients in the ward at the time of referral (only five were referred from outpatient units and had never been inpatients, and only three of these were included). For most (n = 38, 79.1%) it was their first psychiatric admission, but others (n = 37, 77%) had had at least two depressive episodes previously which either went untreated or were treated with medication by their GP. Most of the participants (n = 38, 79.1%) had comorbid DSM-III-R-II (APA, 1980) personality disorders, primarily Cluster C (evasive, dependent, obsessive or unspecified). Participants were referred primarily at the ward's weekly treatment meeting and had been diagnosed with major depression according to ICD-10 criteria, and Hamilton-rated to meet the inclusion criteria of a Hamilton score between 18 and 25. The majority (n = 41, 85.41%) were receiving antidepressant medication (SSRIs, SNRIs or TCAs). Patients went through an initial interview of about 1 hour with one of the group therapists, in which a preliminary case formulation was made, outlining depressive symptoms, their duration and depth (using the BDI-II) and possible vulnerability and triggering factors, inspired in the stress-vulnerability model. The participants were presented with the treatment rationale, their motivation was assessed, and personal goals for treatment aligned. As is common practice in psychotherapy assessment, patients that did not seem capable of reflecting at a fair level in psychological terms during the initial assessment and based on reports of their participation in other therapeutic groups on the ward, did not speak fluent Danish and whose anxiety symptoms or personality disorder were the major problems, were excluded. It took a lengthy period to recruit sufficient patients to the study, so most participants actually had been discharged before the group sessions commenced and attended as outpatients.
Group therapists
There were two therapists in the first four groups: an experienced clinical psychologist with a CBT postgraduate 2-year course and with extensive experience with individual CBT and a co-therapist who was a nurse with experience within psychodynamic group therapy applied to milder psychiatric problems but no prerequisites regarding CBT. They both had supervision from one of the developers of the manual. Due to structural changes, the clinical psychologist had the last six groups on her own.
Treatment programme
The treatment programme consisted of 12 sessions, all structured alike, and two booster follow-up sessions: at 3 and 6 months following termination of treatment. It was built on the principle that the balance between the use of behavioural and cognitive techniques depends on the depth of depression (as seen in Mørch et al. Reference Mørch, Rosenberg and Elsass1995). Thus the first 3–4 sessions focused on psychoeducation and behavioural techniques. It was possible to use several sessions on the latter, if the depth of the single participant's depression required this. The emphasis of the following sessions was increasingly on cognitive interventions. Homework reflected the described principle and started with activity monitoring. Participants learned to notice changes in mood related to activities and discover which activities strained or improved their mood. Following this, distraction techniques were taught and participants started planning activities based on their discoveries in order to break the vicious cycle of depression. As a consequence of depression and the depressed person's often perfectionist demands, it was sometimes necessary to graduate plans and assignments in order to make them more realistic and manageable. From session 4, focus was on registering and modifying negative automatic thoughts using cognitive as well and behavioural interventions. In session 9 the work with depressogenic core beliefs, assumptions and rules of conduct was introduced and it continued for the next two sessions. In the last two sessions the focus was on relapse prevention, summarizing what had been learned, and on evaluation. A qualitative evaluation sheet developed for the treatment programme was given as homework at session 11. Participants were asked what their expectations had been before starting therapy, how useful therapy had been and which elements had been the most useful, what the outcome was regarding mood, their understanding of their current depression-related problems, and their chances of preventing relapse. Finally they were asked what they had missed during the course of group therapy and whether they had any suggestions to improve the programme. Participants were encouraged to carry on using the method. The group met again for two final follow-up sessions after 3 and 6 months following termination of treatment. Their main objective was to retain what had been learned and if needed to review some of the techniques.
The group therapists did their best to adapt the programme to the state of the patients, their personality, needs and resources, which resulted in the need to sometimes deviate from the manual's structure, while maintaining its overall content. The therapists were responsible for keeping a therapeutic approach, as a manualized programme could otherwise make the sessions seem inflexible and theoretical. Even though the sessions were based on each participant's experience of homework, the therapists encouraged the group's commitment to asking questions, coming up with constructive suggestions and the exchanging of personal experience, so that each individual also could benefit from the advantages of being in a group of like-minded participants, although in some respects with very different experiences of depression. The cohesion of the group was enhanced furthermore by a short coffee break without the presence of the therapists.
Deviations from Fennel's method
Already during the first group sessions the therapists became aware of the fact that it would be difficult to create a consensual way of registering pleasure and mastery (Fennell, Reference Fennell, Hawton, Salkovskis and Kirk1999) in a group, because the concept is often perceived differently. Therefore, they chose to follow the manual's recommendation and primarily focus on mood changes and changes in participants’ energy levels. This is the main deviation from Fennel's methodological approach. While conceptual confusion was probably avoided, regrettably the study was deprived of the possibility to nuance experience which lies in pleasure-mastery registration.
Results
Statistical analyses
The data from the 48 patients has been analysed using paired-samples t tests for pre- and post-treatment BDI-II scores. ANOVA and regression were used to compare pre- and post-treatment means for the one-therapist and two-therapist groups, and to examine whether a history of previous depressive episodes, age and gender differences and ongoing education predicted outcome. Results were adjusted for pre-treatment means on BDI-II (shown as ‘adjusted for BDI-II pre’).
Baseline data
Table 1 shows the demographics for the 10-group sample. Table 2 shows that there seems to be a significant effect of the CBT group intervention for depression across the 10 groups.
Table 1. Demographics and baseline characteristics of participants
Quantitative findings.
Table 2. Paired-samples t tests for pre- and post-treatment
* p<0.0001.
Groups 1–4 (n = 22) showed a 9.2 (s.d. = 14.2) decrease while groups 5–10 showed an 11.2 (s.d. = 10.7) decrease in BDI-II; ANOVA adjusted for BDI-II pre (p = 0.49, F = 0.48, df = 1); regression coefficient [−2.41, 95% confidence interval (CI)−9.46 to 4.64]. The study thus suggests that the difference in means between the one-therapist and two-therapist groups was non-significant.
The increase in experience and skills by the therapists with group CBT does not seem to affect participant outcome scores in BDI-II significantly [measured for every new participant, sequentially: mean decrease in BDI-II adjusted for BDI-II pre is 0.15 (95% CI −0.10 to 0.40), greater than the previous participant (p = 0.23)]. Alternatively, by group for every new group, sequentially, mean decrease in BDI-II adjusted for BDI-II pre is 0.62 (95% CI −0.65 to 1.89), greater than the previous group (p = 0.33). This could mean that having extensive experience with individual CBT for depression might be a greater advantage than experience of applying the method to a group setting.
There seems to be no gender difference in outcome [mean BDI-II decrease in women: 9.5, (s.d. = 10.8); mean BDI-II decrease in men: 12.8 (s.d. = 16.5). ANOVA adjusted for BDI-II pre (p = 0.30, F = 1.09, df = 1); regression coefficient (−4.20, 95% CI −12.3 to 3.91)].
Furthermore, there appears to be no significant difference in outcome measured by mean decrease in BDI-II between age groups no matter how they are sampled. For example, age 20–29 years (n = 13; 8.3, s.d. = 11.4), age 30–39 years (n = 20; 12.4, s.d. = 10.1), age ≥40 years (n = 15; 9.2, s.d. = 16.2). Adjusted for BDI-II pre (p = 0.50, df = 2, regression coefficient compared to ages 20–29 years; ages 30–39 years (−4.90, 95% CI −13.50 to 3.70), ages ≥40 years (−1.93, 95% CI −11.37 to 7.51). Alternatively ages 0–49 years (n = 38; 9.2, s.d. = 12.1), ages ≥50 years (n = 10; 15.6, s.d. = 12.4). Adjusted for BDI-II pre [ANOVA: p = 0.13, df = 1; regression coefficient −6.84 (95% CI −15.80 to 2.11)]. In other words, patients of all ages seem to have benefited from treatment.
The hypothesis that participants that had had more than two previous depressive episodes and longstanding personality impairment would benefit less than first-episode participants was not supported by the data. Quantitative measures thus revealed no significant difference in outcome (measured by decrease in BDI-II score) between the two groups. First episode: mean decrease in BDI-II (12.7, s.d. = 10.9); more than two previous depressive episodes and longstanding personality issues: mean decrease in BDI-II (9.7, s.d. = 12.7); ANOVA adjusted for BDI-II pre (p = 0.12, F = 2.48, df = 1); regression coefficient (6.69, 95% CI −1.89 to 15.3).
According to hospital admission records only three of the group members had been briefly readmitted with depressive relapse approximately 1 year following termination of the 12-session treatment programme. The rates of readmission in our sample were compared with the 20-month readmission data (with the greater risk of readmission being within the first 8 weeks upon discharge) published in a previous register-based study on Danish patients with unipolar depression, with a rate of 25% among 17434 patients (Kessing et al. Reference Kessing, Andersen and Bowlvig1998). The readmission rate was significantly lower in our sample (p = 0.003).
The sample at hand is fairly small, making the results tentative. However, the results so far and patient feedback indicate that most group participants seem to have benefited from the 12-session treatment, indicated by mean BDI-II decrease by session 12 (Table 2). There were 33 participants present at the 3-month follow-up and 26 at the 6-month follow-up. Unfortunately there are no quantitative measures after 3 and 6 months, as not enough participants handed in their BDI-II at follow-up, and there were no means to seek follow-up data by post. So the above impression relies solely on the face value clinical impression and participants’ orally self-reported state (maintained improvement, mood and level of functioning) at the follow-up sessions.
Qualitative findings
During session 11, all participants were given the aforementioned qualitative questionnaire to assess their benefits from the CBT group programme, which they handed in at session 12. The items they emphasized can be divided into two groups: first, non-specific group dynamics, such as learning from peers, the supportive atmosphere and the importance of group size for their sense of well-being and willingness to self-disclose about own experiences. Second, the factors related to the specific method with elements such as psychoeducation, activity planning, distraction techniques, restructuring of negative automatic thoughts, focusing on changing their inappropriate behaviour and modifying inappropriate core beliefs, assumptions and rules of conduct. Some pointed out that they had become better at communicating with their relatives about their illness and needs. Many said that the treatment had been too short, and some of the keenest could have wished for an earlier introduction to the cognitive interventions in order to have more time to work with schemata, as well as for monthly ‘booster sessions’ after the end of the 12-session programme. Some would have needed regular individual motivational interviewing as a means of supporting the completion of homework. Several mentioned that they had missed an element of assertion and self-esteem training.
Qualitatively the majority indicated feeling ‘much better’ or ‘decisively better’ (only four, or 8.3%, described their mental state as ‘unchanged’) and to have benefited from the group process as well as understanding their problems ‘much better’ or ‘decisively better’ at the end of the 12 sessions. Yet many still had significant depressive complaints which might be explained by the fact that the majority (80.9%) of participants indicated having had longstanding personality impairment and thus were still subjectively suffering and naturally did not experience having concluded their psychotherapeutic work by the end of the treatment programme, despite objective improvement (in terms of decrease in BDI-II).
Clinically it can be noted that younger participants (ages 20–49, n = 37), students and participants with few or brief episodes of depression and mild personality impairment, although not having benefited the most (in terms of reduction in BDI-II scores), often showed the most flexibility in thinking regarding themselves, others, the treatment rationale, etc. They came well-prepared to most sessions, completed homework without significant problems and no longer felt depressed at the end of treatment. Participants with more prolonged depressive complaints, severe personality impairment, excessive anxiety and comorbid somatic problems had more difficulty in completing homework on their own and reported still feeling depressed at the end of treatment, which primarily addresses the depressive symptoms. It is well known and documented that personality impairment benefits most from the work focused on modifying inappropriate schemata and dysfunctional behaviour (Young, 1990/Reference Young1999; Young et al. Reference Young, Klosko and Weishaar2003) which was barely addressed in the group sessions.
Discussion
The outcome results show a significant decrease in BDI-II score despite participants’ severe symptoms and pronounced cognitive and motivational impairment throughout the programme. The results with short term, manualized group CBT for depression described in this study thus leave the clear impression, supported by outcome data, that the method can be applied successfully in a group setting across genders and age groups, and even patients with moderate to severe problems seem to benefit from short-term, manualized group CBT. As the study was conducted in a naturalistic setting rather than with a highly selected patient sample, it has immediate implications for clinical practice. However, it is crucial to realize that expectations regarding effect must be modest, the greater the symptom severity and personality impairment. This corresponds to the published research results, for example, Ball et al. (Reference Ball, Kearney, Wilhelm, Dewhurst-Savellis and Barton2000). At the same time, in a manual-based group approach, while drawing on the benefits of group treatment, such as a sense of fellowship and the sharing of relevant experiences, one risks losing some of the benefits of an individual cognitive behavioural approach. This is because it is not possible to tailor therapy to the individual needs to the same degree, which could reduce the effectiveness of the treatment – a tendency which is implied in the scarce research on the subject (e.g. Rush & Watkins, Reference Rush and Watkins1981; Wierzbicki, Reference Wierzbicki1987).
Although no data on ease of homework completion was collected, students were observed to have fewer difficulties in doing their homework between sessions. Surprisingly this did not prove to have a significant impact on outcome, as measured by mean decrease in BDI-II score by session 12 [not studying currently (8.8, s.d. = 13.2), currently studying (14.6, s.d. = 8.5). ANOVA adjusted for BDI-II pre (p = 0.23, df = 1); regression (−4.77, 95% CI −12.63 to 3.09)]. But it might play a role following termination of treatment, as one would expect that it would seem easier to carry on using the method for patients trained in doing homework. Unfortunately there is no data available to test this assumption.
This study aimed to find out whether a short-term, manualized CBT group intervention could be used for patients in an inpatient setting by measuring treatment outcome and relapse rates following a 12-session CBT group programme. However, due to structural issues (mainly very short admissions) and recruitment problems, most participants had already been discharged before the group sessions commenced and participated as outpatients. Thus it was not possible to study this CBT group intervention used for inpatients as such, nor to draw upon the aid of the ward staff and ward structure for a greater impact of interventions, e.g. in doing homework and training skills. More studies on the central factors for successful CBT group treatment for depression in inpatient populations are required.
There were no extra research funds or staff assigned, which limited the collection of more extensive demographic data, using other measures besides BDI-II scores to evaluate outcome (e.g. level of functioning such as the Global Assessment of Functioning scale and other symptom scales), collecting follow-up data and other relapse measures that would be needed in order to establish the effect of treatment over time. For research purposes it would have been useful to use Hamilton scores at baseline as well as an outcome result after the 12-session programme and at the follow-ups but unfortunately there were no resources to do this.
While conceptual confusion may have been avoided by not registering pleasure and mastery, as suggested by the manual, on the other hand the programme missed the opportunity to nuance experience that lies in pleasure-mastery registering. Moreover, it may make it more difficult to compare with other studies, as pleasure-mastery registering is a standard method, whereas only registering mood alone is not.
Based on the results in this study, expanding treatment significantly could prove to be necessary, as it has been shown that most patients hospitalized with depression actually have more long-term problems and comorbid personality impairment. Although it is probably far from sufficient, one should consider extending the group programme to at least 14–15 sessions followed by a number of booster sessions, which can be encompassed by the resources presently available to the ward. In a healthcare system characterized by on-going budget cuts, it is ultimately a matter of resources, but the allocation of funds for more long-term group CBT with a focus on working with basic assumptions and inappropriate strategies following this 12-session programme primarily directed at the treatment of actual depressive states, could be considered in the future.
Main points
A short-term, manualized CBT group intervention can be used for patients with severe depressive symptoms, even though patients are admitted with severe affective and cognitive complaints. Data on outcome following termination of treatment was rather difficult to collect without the funding of extra staff. A 12-session programme may have to be extended for people with recurring depressions and more complex and longstanding personality impairment.
Declaration of Interest
None.
Acknowledgements
The author thanks Carsten Rygaard Hjorthøj, PhD, MSc, senior researcher at the research group of the Department of Psychiatry of Copenhagen for assistance with the statistical analyses.
Learning objectives
-
(1) A short term 12-session manualized CBT group intervention can be applied to patients with severe depression.
-
(2) More moderate outcome results can be expected when severity of symptoms for admission is taken into consideration.
-
(3) Basic CBT should be modified to meet the needs of this more severe group with longstanding problems.
Comments
No Comments have been published for this article.