Introduction
Clinical guidelines for depression in the UK recommend the use of Cognitive Behavioural Therapy (NICE, 2004, 2009). Given therapist capacity problems, this study aimed to compare outcome, costs and patient engagement for Group and Individual CBT in clinical settings using a prospective observational design. Comparative studies of group and individual CBT tend to show equal levels of effectiveness (Lockwood, Page and Conroy-Hiller, Reference Lockwood, Page and Conroy-Hiller2004), but there is less evidence on the impact of patient preference or satisfaction. Group interventions are likely to be less costly to provide than individual treatments, which is an important consideration given increasing budgetary constraints.
Method
Participants were recruited from five psychology services provided by a large mental health Trust in South East London. Therapists (n = 37) were asked to identify patients for the study who were over 17 years old, had a primary diagnosis of depression, and who were due to start treatment between January and November 2006. Potential participants (n = 130) were then contacted by the researcher to obtain ethical consent and complete baseline measures. The therapists allocated patients to Individual (I-CBT) or Group CBT (G-CBT). Commonly, I-CBT lasted 6–18 sessions and G-CBT comprised 10–12 weekly sessions with 8–12 attendees. A standardized treatments manual was not issued, so treatments reflected usual practice in each service. Research assessments were collected prior to the intervention (baseline), post-treatment and at a 3-month follow-up using the Beck Depression Inventory II (BDI-II and primary outcome measure; Beck, Steer and Brown, Reference Beck, Steer and Brown1996), the Centre for Epidemiological Studies Depression Scale (CES-D; Radloff, Reference Radloff1977), the Clinical Outcomes in Routine Evaluation-Outcome Measure (CORE-OM; Evans et al., Reference Evans, Mellor-Clark, Margison, Barkham, Audin, Connell and McGrath2000), the Client Satisfaction Questionnaire (CSQ-8; Nguyen, Attkisson and Stegner, Reference Nguyen, Attkisson and Stegner1983), and a Treatment Preference Scale scoring ‘strength of preference’ on a 0–10 scale.
The power calculation showed 26 patients in each arm were required to detect a large effect size in outcomes (Cohen's d = 0.8). Comparisons of the G-CBT and I-CBT groups used linear mixed modeling, including assessment of bias due to attrition (34%). Random intercept models were fitted to assess the effect of the two therapies on the outcome measures at post-treatment and follow-up. These models included baseline values, propensity scores based on all possible variables affecting treatment allocation, main effects of groups, post-treatment time points, and the interaction between time points and groups. If the time by group interaction was not significant, this term was removed from the model and only the main effects of time and groups tested.
Results
At baseline, 70% of all participants stated a preference for I-CBT, 10% for G-CBT and 20% stated no strong preference. Forty-four participants (53%) received G-CBT and 49 (47%) received I-CBT. Clinicians were found to have allocated patients to G-CBT who were older, had a longer history of depression with more previous treatments and hospital admissions, were less likely to be working, and had less education. Baseline outcome scores indicated severe depression and clinical distress for both groups (Table 1) and there were no significant differences between the groups.
Outcome by therapy groups and time point
At post-treatment, BDI scores decreased significantly in both groups (I-CBT: t = 5.80, df = 33, p <0.001; G-CBT: t = 4.80, df = 31, p <0.001) and then increased slightly 3 months later (Table 1). The final column of Table 1 shows the results of fitting random intercept models to the outcome measures. There were no significant group effects or changes over time on the BDI or CES-D or CORE-OM. These findings reflect those from earlier studies with no significant differences found between I-CBT and G-CBT.
Patients were equally satisfied with treatment whether they received I-CBT or G-CBT. Notably, 53% of those receiving G-CBT and who stated a strong preference for I-CBT had changed their treatment preference at the post-treatment assessment. Patients were equally likely to drop-out of either treatment (34% attrition) and neither treatment preference nor satisfaction had an impact on attrition rates.
The costs of providing I-CBT and G-CBT were estimated to include staff time, non-staff costs, organizational overheads and capital at 2006–07 rates. There was considerable variation in treatment duration in these routine settings but mean costs between baseline and post-treatment assessments were lower for G-CBT (n = 40) than for I-CBT (n = 45); £246.33 (SD 108) vs £375.32 (SD 216) (p = .001, bootstrapped CI 58.92, 199.07). Costs estimated for all other services used (including any CBT sessions) were similar before and after treatment. As the outcome findings indicate no differences between I-CBT and G-CBT in reducing symptoms, it could be argued that the equivalence of outcome at a lower cost implies a cost-effectiveness advantage to G-CBT.
Discussion
No differences were found in depression symptoms and costs after I-CBT and G-CBT treatment. However, a naturalistic design is not ideal; we cannot be sure that the outcome and cost findings did not result from the research format, or from unmeasured treatment or patient characteristics. Neither was the initial assessment of depression standardized, nor other potentially confounding diagnoses recorded. However, research has an important role to play in evaluating feasibility, acceptability and effectiveness of ways of working in those environments that are not in controlled conditions but where most patients will be treated.
Our results imply that by implementing G-CBT depression, symptoms can be reduced without sacrificing satisfaction or take-up, and it can be provided at a lower cost to psychological therapy services. This study also suggests that a mix of both treatment modes is feasible in routine practice, although G-CBT requires more preparation and organization and, indeed, so may patients. Since this study was conducted, the latest Guidance for Depression (NICE, 2009) recommends that G-CBT should be offered to people with persistent sub-threshold depressive symptoms or mild to moderate depression without chronic physical health problems. While a larger RCT is needed, perhaps using standardized treatment modes, this research suggests that Group CBT could help implement the NICE Clinical Guidelines in routine practice. What remains unclear is when and for what kinds of patients G-CBT and I-CBT may be most appropriate when CBT is offered routinely to patients in mental health services.
Acknowledgements
This project was funded by the South London and Maudsley NHS Foundation Trust. June Brown is supported by the NIHR Specialist Biomedical Research Centre for Mental Health at the South London and Maudsley NHS Foundation Trust and Institute of Psychiatry, King's College London.
Comments
No Comments have been published for this article.