Commentary on ‘Evaluating Treatments for the Mood Disorders: Time for the Evidence to Get Real’

Clinical academics in Australasia take great pride in the ground-breaking efforts of Gavin Andrews and colleagues, who in association with the Royal Australian and New Zealand College of Psychiatrists, first argued the merits of an evidence-based approach to clinical psychiatry in the early 1980s [1,2]. In this, and a series of other closely related articles [3,4], Parker now asserts that there are major limitations to such research methods and that the evidence base for current recommendations for antidepressant treatments [5,6] is severely flawed. At first glance, health professionals, administrators, consumers and their families might rightly be distressed by this view, especially when it is expressed by such an eminent clinician. Additionally, wide media reporting [7] of such views reinforces negative beliefs about the lack of seriousness of depressive disorders [8], and the dangers of presenting for treatment [9]. It is in that wider professional and social context that Parker’s argument now requires closer scrutiny. To be frank, Parker’s critique of randomised controlled trials (RCTs) lacks substance. For very good reasons RCTs (and more importantly, the pooled results of such trials) are widely accepted as the highest level of evidence for determining the efficacy (as distinct from clinical effectiveness) of any new treatment [10]. Specifically, with regard to mood disorders (as with many other psychiatric and other chronic medical conditions), where the clinical course is unpredictable, the potential short or long-term benefits of providing any intervention need to be clearly established. Importantly, RCTs are not necessarily the appropriate tool for answering a variety of other questions concerning clinical or cost effectiveness, or population health impacts [10]. For example, the recent use of other statistical techniques has supported the population benefit (reduction in suicide rates) in Australia among those accessing current treatments for depression [11]. The problems we face are less the limitations of the available research tools available but rather the unwillingness of both clinicians and researchers to employ the right tools for the right jobs! New drug, psychological or lifestyle interventions are constantly proposed for treatment of a wide variety of depressive disorders [12]. Given the prevalence of depression and anxiety in the wider community, potentially, such interventions may be applied to a very large number of persons or to a wide variety of subpopulations. Before promoting any interventions to the wider public, we need to apply the most robust evaluation methods at our disposal. A large variety of study designs other than RCTs may be employed during the development of any new treatment and in relation to any of the clinical or neurobiological subpopulations described. However, the ultimate test of efficacy remains the RCT. Before making wider public health recommendations, typically we utilize the pooling of data from among the best constructed of these trials [5]. To do otherwise would simply set back the scientific basis of clinical psychiatry and further undermine our public credibility. Parker’s main concern appears to be that clinicians are not already aware that various subpopulations of people with depression run different clinical courses and may respond differentially to current interventions. In the early 20th century, Kraepelin’s classic lectures on clinical Ian B.