Behaviour therapy for depression

In the first paper, Ekers et al. (pp. 611–623) report findings from a meta-analysis of 17 randomized controlled trials (RCT), involving 1109 subjects, of behaviour therapy (BT) for depression. They found that in terms of symptom levels post-treatment, BTs were more effective than control interventions, brief psychotherapy, and supportive therapy, and equal to cognitive behaviour therapy (CBT). The authors conclude that BT is an effective treatment for depression.

Commentaries

There are two commentaries on papers in this issue. In the first, McLoughlin (pp. 625–627) discusses Schlaepfer et al.'s report on a trial of vagus nerve stimulation (VNS) for treatment-resistant depression. He notes the discrepancy between findings from a previous RCT that showed no difference between real and sham VNS at 10 weeks and open studies that have suggested there are effects of VNS on depression over a longer period. McLoughlin concludes that it remains important to investigate the efficacy of VNS and other similar techniques, but notes the methodological challenges posed, including establishing suitable controls.

Clark et al. (pp. 629–634) comment on Stiles et al.'s report comparing the effectiveness of three forms of therapy (CBT, person-centred therapy, and psychodynamic therapy) in routine practice. Clark et al. applaud the work of Stiles and colleagues in promoting the use of routine outcome measures in clinical practice. However, they argue that a number of methodological problems with the present study raise doubts about the validity of Stiles et al.'s conclusion that each therapy is equally effective. These include the large number of missing cases (62% of the total), the inability to control for the natural course of depression and use of antidepressants, and failure to check whether the stated intervention was faithfully delivered.

Depression and suicide

Five papers examine various aspects of depression and suicide. In the first, Spek et al. (pp. 635–639) present data from a RCT of internet-based CBT, group CBT and a waiting- list control for subthreshold depression in people aged over 50 (n=301). They found that internet-based CBT, when compared with the waiting-list controls, was more effective in reducing symptoms at 1-year follow-up. Both internet-based and group CBT were similarly effective. The authors conclude that those aged over 50 with subthreshold depressive symptoms can benefit from internet-based CBT.

Tozzi et al. (pp. 641–649) investigated the relationship between a family history of depression and age of onset in a sample of 1022 cases with recurrent major depressive disorder. They found that a positive family history was significantly associated with an earlier age of onset. This association was strongest for onset before the age of 20 and non-significant for onset over the age of 50. Those with a positive family history were more likely to experience feelings of guilt, symptoms of anxiety, and functional impairment.

Schlaepfer et al. (pp. 651–661) examined the efficacy and safety of VNS for depression in an open, uncontrolled multi-centre trial involving 74 patients with treatment-resistant depression. After 3 months of VNS, response rates [i.e. a 50% or greater reduction in the Hamilton Depression Rating Scale (HAMD) score] were 37% and remission rates (i.e. a HAMD score of less than 10) were 17%. After 1 year, response rates increased to 53% and remission rates to 33%. The authors conclude that VNS was effective in reducing symptom severity and that efficacy increased over time.

Duberstein et al. (pp. 663–671) investigated the association between personality traits and first lifetime onset of depression in a cohort of 397 adults aged 70 years who were followed for 15 years. During the follow-up there were 59 new cases of depression. The authors found that higher Neuroticism scores, but not Extroversion scores, were associated with increased risk of a first lifetime episode. The authors conclude that these findings may have implications for the identification and treatment of depression in older adults.

Ajdacic-Gross et al. (pp. 673–676) examined the relationship between suicide and time following bereavement in data extracted from Swiss mortality statistics for the period 1987–2005. Rates were highest in the first week following bereavement at 94/100 000 for men and 207/100 000 for women. In the first months after bereavement, rates declined rapidly and then gradually. For most groups, rates did not reach baseline suicide levels until 1 year after bereavement. The authors conclude that widowed persons are at increased risk of suicide in the weeks and months following bereavement.

Intervention trials

Five further papers report findings from intervention trials. Stiles et al. (pp. 677–688) investigated the effectiveness of CBT, person-centred therapy (PCT), and psychodynamic therapy (PDT) in a sample of 5613 subjects who completed the Clinical Outcomes in Routine Evaluation – Outcome Measure (CORE-OM), drawn from 32 NHS primary-care services during a 3-year period. Subjects in all treatment groups showed significant improvements at the end of treatment, with no differences between the therapies.

Knekt et al. (pp. 689–703) report findings from a RCT of long-term and short-term psychodynamic psychotherapy and short-term solution-focused therapy for symptoms of depression and anxiety in a sample of 326 out-patients followed for 3 years. The authors found that, compared with long-term therapy, short-term psychotherapy was better in the first year, no different in the second year, and less effective in the third year in reducing symptoms. There were no differences between short-term psychotherapy and solution-focused therapy.

Thompson et al. (pp. 705–715) present data from a cluster RCT of an intervention to reduce polypharmacy compared with guidelines alone, involving 19 adult in-patient units in the UK. They found that, in those prescribed antipsychotics, the odds of polypharmacy were significantly lower in the intervention than in the control groups (OR 0.43). There was, however, notable variation between units in polypharmacy rates, and the effect size of the intervention was relatively modest. The authors conclude that more research is needed to elicit the factors that promote changes in prescribing behaviour.

Kitcheman et al. (pp. 717–723) report findings from a pragmatic RCT of the impact of sending an encouraging letter on attendance at psychiatric outpatient clinics in a sample of 764 subjects of working age given their first appointment. The authors found that those who were prompted were significantly less likely to fail to attend (OR 0.75). The authors conclude that prior encouragement can improve out-patient attendance.

Jackson et al. (pp. 725–735) present data from a RCT of CBT versus Befriending for first-episode psychosis in a sample of 62 subjects followed for 1 year. The authors found that both CBT and Befriending groups improved over time. CBT outperformed Befriending by improving functioning at mid-treatment, but it did not improve symptoms. Beyond mid-treatment, Befriending caught up with CBT and there were no further differences evident. The authors conclude that there is some evidence that CBT for early psychosis promotes better early recovery in functioning.

Psychosis

The final three papers examine aspects of psychosis. Gonzalez-Blanch et al. (pp. 737–746) investigated pre-treatment predictors of cognitive deficits in a sample of 131 patients with a first-episode of psychosis. The authors found that premorbid functioning was associated with five of six cognitive domains considered. More specifically, difficulties with social adjustment were associated with poorer executive functioning, motor dexterity, and sustained attention. Academic functioning was associated with verbal comprehension, verbal learning and memory. No further predictors of cognitive deficits were identified.

Baas et al. (pp. 747–754) examined social judgement, using tasks involving assessments of the trustworthiness of faces, in 18 patients with schizophrenia, 24 unaffected first-degree relatives, and 28 controls. They found that patients, compared with controls, were more likely to rate faces as more trustworthy. Relatives showed similar biases, but to a lesser extent. The authors note that these findings parallel studies of patients with abnormalities in brain areas involved in social cognition.

Sitzer et al. (pp. 755–763) investigated predictors of social skill performance in a sample of 194 middle-aged and older out-patients with schizophrenia and 60 controls. The authors found that patients performed worse than controls. Better performance was associated with a number of factors, including fewer positive and negative symptoms, fewer social contacts, and abstraction ability. These latter two factors were the strongest predictors of social skill performance when demographic and clinical factors were controlled.