Dr Geaney echoes arguments put forward for many years by Jerome Wakefield and already extensively addressed in the literature (e.g. Kendler, ^{Reference Kendler1} Murphy & Woolfolk ^{Reference Murphy and Woolfolk2} ). Indeed, the evolutionary approach may help us to understand how and why depression has developed in the human species as a response to major losses. However, the relevance of that approach to ordinary clinical practice (i.e. in helping the clinician to discern whether the depressive state of a given individual is a mental health problem deserving clinical attention) is, at the present state of knowledge, very doubtful, and the risk of ‘over-romanticizing the suffering associated with major depression’ ^{Reference Kendler1} is very high.

Wakefield himself has documented that as many as 95% of depressive episodes seen in the community are triggered by an adverse life event, according to the affected person’s report. ^{Reference Wakefield, Schmitz, First and Horwitz3} This is not surprising, because many people with depression try to find a meaning in their current state, ascribing it to a recent event. Whether there is really a causal relationship between the event and the depressive state, what is the direction of that relationship, and whether the depressive state is ‘proportionate’ or not to the event is very difficult or even impossible to establish reliably in the vast majority of cases. This was already acknowledged by Sir Aubrey Lewis many decades ago, when he tried to apply a set of criteria aimed to distinguish ‘contextual’ from ‘non-contextual’ depression and had to conclude that almost all cases he had encountered were ‘examples of the interaction of organism and environment, i.e. personality and situation; it was impossible to say which of the factors was decidedly preponderant’. ^{Reference Lewis4}

Actually, whether there is something like a ‘normal’ or ‘proportionate’ response to a given life event is highly debatable. Even when exposed to the most extreme life event, the majority of people will not develop a depressive state. Which ‘standard’ are we going to apply when deciding whether a given depressive response is proportionate or not to a given life event? Are we aware that there are mental health professionals who do believe that every psychopathological manifestation can be ‘explained’ by the individual’s environmental circumstances? Would we feel comfortable in basing the diagnosis of depression on that subjective judgement?

In addition to having poor reliability, the proposed ‘contextual’ exclusion criterion does not seem, at the current state of knowledge, to have a significant clinical utility (the main element which is being taken into account in the revision of ICD-10 and DSM-IV). Currently available evidence suggests that the response of a depressive state to pharmacological treatment does not depend on whether that state was or was not preceded by an adverse life event. ^{Reference Anderson, Nutt and Deakin5} Furthermore, interpersonal psychotherapy is based on the assumption that depression is often understandably related to a disturbing life event, such as a loss, a role dispute or a role transition, and that ‘if the patient can solve the life problem, depressive symptoms should resolve as well’. ^{Reference Markowitz and Weissman6} Should we conclude that all cases in which interpersonal psychotherapy is effective are not ‘true’ cases of depression?

Finally, that the ‘proportionality’ criterion enables us to ‘distinguish normal from abnormal responses of the brain’, or to identify those depressive states in which there is ‘a failure of some internal mechanism to perform a function for which it was biologically designed (i.e. naturally selected)’, ^{Reference Wakefield7} is at the moment an interesting theoretical assumption with no empirical basis. In Arthur Kleinman’s words, ‘the data on this allegedly universal biology of loss are simply not there’. ^{Reference Kleinman8}

Dr Shepherd’s comment clarifies that the mere diagnosis of depression is not sufficient to guide decisions concerning management. The assessment of the severity of the clinical picture and the characterisation of the individual case on biological, psychological and social domains are essential for that purpose. Both DSM-IV and ICD-10 identify different degrees of severity of depression based on the number of symptoms and the entity of functional impairment. However, the conceptualisation of functional impairment in both systems is too vague and depends too much on the subjective judgement of both the patient and the clinician. In the new edition of the two diagnostic systems, it will be necessary to anchor the assessment of functional impairment to clear and objective variables.

Furthermore, as Dr Shepherd implies, the functional status of a person with depression depends not only on the depressive state per se, but also on the multiple physical and mental conditions which commonly co-occur, so that a global assessment of functioning may be more relevant for management purposes. Ideally, that assessment should be as comprehensive as Dr Shepherd suggests, but the search for comprehensiveness will have to be balanced with the need to ensure feasibility (as well as reliability) in ordinary clinical practice.