To the Editors, In their erudite commentary, Ballard and Zarate 1 advocate that, to be effective, suicide prevention strategies should be considerate of the multifactorial nature of this complex behaviour. Undoubtedly, prediction and prevention of suicide behaviour are very challenging tasks given its complex biological underpinnings, the uneven distribution of its risk over a lifetime trajectory, and the very different magnitude of risk associated with each determinant, either proximal or distal. However, we are less optimistic about the efficacy of the potential primary and secondary preventive options indicated by the authors, namely the Good Behaviour Game (GBG) and the glutamatergic modulator ketamine, on the basis of some phenomenological, epidemiological and pharmacological considerations. Suicide behaviour manifests according to a specific trajectory of severity (from suicide ideation to acts and eventually death by suicide). However, this trajectory might not necessarily be characterized by a temporal and hierarchical consequentiality, since suicide ideation is an almost necessary factor for suicide but rarely sufficient. Indeed, epidemiological data show that only a small proportion of patients who manifest suicide thoughts go on to commit an attempt, 2 and risk factors might be distinctly associated with suicide ideation or suicide acts. 3Of interest, the National Comorbidity Survey data showed that the risk of attempt was even higher among patients who presented suicide ideation without plan (odds ratio [OR] 2.5) than among those who had planned attempts (OR 1.5). 3 This evidence put clinicians in a quandary as to whether ketamine, a promising and plausibly effective treatment for depression and therefore possibly useful in controlling suicide ideation, might be successful in patients at high risk for suicidal acts, such as those with a personal history of previous attempts. Moreover, ketamine seems to be effective for short periods of time; thus it is hard to delineate its precise role in a therapeutic strategy for the treatment of suicide ideation. We also think that the authors should be careful in considering diverse pharmacological treatments (from antidepressants to mood stabilizers) deemed to be equally effective in controlling suicidality. In mood disorders, lithium is the only drug that significantly reduces suicidal risk (both ideation and acts), 4 probably even in the absence of an effective mood stabilization. 5 Finally, and possibly more importantly, we should all be mindful that suicide behaviour remains a very rare event. Thus, primary preventive strategies, such as GBG, although effective, would need such a large-scale and extensive implementation that their feasibility would inevitably be impacted. In summary, we agree that suicide behaviour should be considered in the light of its complexity. At the same time, we reckon that the search for complex answers to a complex problem should not lead to parsimonious solutions, such as the prevention of mood altered episodes, especially of dep...