Psychological intervention poststroke: ready for action?

See related article, pages 3073–3078. Stroke is distressing and frequently affects mood and behavior.1 The negative impact of mood and adjustment disorders on physical recovery and reintegration is recognized2–4 and makes it even more imperative that early effective interventions are established. However, there has been a failure to respond to this distress at an early stage after stroke, perhaps due to staff viewing distress as an inevitable psychological reaction to a debilitating event or due to a lack of clear guidance around prevention and treatment of mood problems. Later on in the recovery process when this distress has become established as a mood disorder, there has been a failure to react, and care remains suboptimal. The lack of robust evidence to guide treatment is a major contributing factor. Thus far, the results of antidepressant studies in those with established depression have been disappointing with shortterm gains not being translated into long-term benefits. Similarly, psychological interventions, when tested in groups of community-based patients with depression, have mostly proved ineffective.5 However, there is emerging evidence that some psychological interventions, when delivered early after stroke, have the potential to both prevent and treat depression, at least in the short term, and may improve survival.6–8 Further evidence has been needed to inform later treatment and understand the longer-term effects of interventions; Mitchell et al9 provide a thought-provoking article to prompt further debate. In this issue of Stroke, Mitchell et al9 examine the effectiveness of what they describe as a brief psychosocial–behavioral intervention, as an adjunct to antidepressant therapy, in reducing poststroke depression. In this randomized, controlled trial, at 4 months poststroke, patients were allocated to either 8 weeks (9 sessions) of psychosocial–behavioral intervention or usual care. For both groups, antidepressant therapy with selective serotonin reuptake inhibitors was the community standard treatment for those with established depression. Only 60% of patients in both groups were receiving selective serotonin reuptake inhibitors at the start of the study, although this increased to nearly 80% within the study period. The antidepressant component of treatment was a component of usual care rather than an explanation of an explicit component being evaluated within the study. There are some measurement issues raised by this research. The use of the Hamilton Rating Scale for Depression10 at baseline and outcome is unusual due to its inclusion of symptoms that are often more a reflection of the physical health problems associated with the stroke itself. We assume the authors have used it here because they considered it to be an interval scale, which allowed them to compare “change” scores and “percentage change” (primary outcome) between groups. The results of this analysis suggest promising benefits for their intervention. More convincing evidence of the effectiveness of the intervention, which is perhaps more clinically relevant, is seen in the difference in the proportion of patients who would be classified as not depressed at each outcome assessment time point. The a priori power calculation to inform the sample size could perhaps have used this difference in the proportions depressed rather than percentage reduction in scores, in what many may consider to be an ordinal scale. The research also raises an issue for further research. Over 1000 previous inpatients were considered for this study with 289 agreeing to be screened of whom 139 became ineligible for the study. Ineligibility was a result of Geriatric Depression Scale (GDS)11 scores being 10 (the cutoff for severe depression), although 49 were excluded for other reasons. This gave the study a modest sample size of 101 patients who screened positive on the Geriatric Depression Scale and who also then met the diagnostic criteria for depression when interviewed. Although tight criteria were appropriate for this study, because all those screening positive were diagnosed with depression, reducing the Geriatric Depression Scale cutoff (eg, to 5 to 10 for mild to moderate depression) for screening may be useful in future work to allow more potentially suitable patients to be correctly included when the diagnostic interview and criteria are applied. What Mitchell et al9 do demonstrate convincingly is the benefits to be gained in activity and participation should remission of depression be achieved, even if it cannot currently be definitely attributed as a treatment effect. Although this study adds to the debate, further research is required to clearly demonstrate the effectiveness of psychological interventions, antidepressant treatments, and their additive effects.