Is Add-on Psycho-education Effective in the Treatment of Depression?

摘要

To the Editor: I read with interest the study by Kumar and Gupta,¹ published in March 2015 issue, on the effectiveness of psycho-educational intervention in patients with depression and their caregivers that found significant differences in outcome measures. Psycho-educational interventions as add-on to pharmacotherapy were reported to significantly reduce depressive symptoms as well as improve global functioning and psychological wellbeing. I have some reservations about these findings based on the facts in the study. First, as mentioned by the authors, the study was carried out over a period of 3 months, from April to June 2012. All patients were assessed at baseline, 4, 8, and 12 weeks on the rating instruments. This would only have been possible if all 80 patients mentioned were recruited on the same day, as the last follow-up was after 3 months. This appears highly improbable. Second, the authors mentioned that “a total of 80 eligible subjects were recruited and randomised alternately into 2 groups”; such type of alternate allocation is not a true randomisation procedure.² In the subsequent paragraph they mentioned that another group of newly diagnosed subjects served as controls. This again raises concern about the process of randomisation. Third, it appears that all patients were receiving medications (details not mentioned) as per the discretion of the treating clinician, but the exclusion criteria for patients included “partially treated or current treatment for depression”. This seems counterintuitive. Fourth, the psycho-educational intervention included only 4 sessions (duration of each session was not specified) given at baseline, 2, 4, and 8 weeks to 1 group of patients and caregivers. The authors found a significant reduction in depression scale score from 4 weeks onwards in the intervention group, i.e. when the patients had had only 2 sessions of psycho-education. At 12 weeks the difference in the Hamilton Depression Rating Scale scores between the groups was significant at p 0.001, with a strong effect size (Cohen’s d = 1.3).³ Furthermore, there were no corrections made for multiple testing at various time points. A more appropriate statistical test in this situation would be repeated measures analysis of variance⁴ that could have controlled for the baseline differences in the scores too.