My daughter and I have a running joke, which goes something like this:

My daughter may be a relatively late convert, but after nearly 20 years in old age psychiatry I think the message does at last seem to be getting through that depression in old age is common and disabling, can be life-threatening and is eminently treatable. The appearance of this large (A4) and handsomely bound volume is a good sign. The editors are among the USA's most eminent ‘melanchologists’, although neither specialises in old age psychiatry. The remaining 41 contributors all also work in the USA and represent a comfortable majority of the most eminent US academics in the area. The 29 chapters are organised into five broad themes: epidemiology, symptoms and diagnosis, psychobiology, treatment and comorbidity.

The editors state that the book should be useful ‘to the clinician who strives to understand the multiple dimensions of aging and the complexity of late-life depression and who aspires to practice evidence-based interventions’. It is therefore unsurprising that each chapter is extensively referenced, and that the emphasis throughout is on summarising the available scientific literature. Also unsurprising but perhaps more disappointing is the overwhelmingly biological orientation, with only a single eight-page chapter on psychotherapy. This is written by Chip Reynolds and his colleagues from the Pittsburgh group and consists mainly of a very lucid summary of their own (albeit pivotal) trials.

I decided to ‘road-test’ the book on what I thought would be one of its strengths. I was reviewing the (unpublished) clinical trial data on a new antidepressant against placebo in older people and needed to compare it with the placebo-controlled data on currently marketed antidepressants. The relevant chapter (written by the editors themselves) made the excellent point that there was a dearth of evidence relating to the ‘old old’. Better still, it had an easily found table entitled ‘placebo-controlled trials’ that seemed just what I wanted. Disappointingly, however, both the table and the text mentioned only four of the 18 studies collated in a recent meta-analysis (Reference Taylor and DoraiswamyTaylor & Doraiswamy, 2004).

My more general concern about the book's claimed orientation to the clinician is that most of the chapters fail to make the crucial move from literature summary to clinically relevant synthesis. There are plenty of statistics but no clinical vignettes and not even any clear ‘best practice’ recommendations. I hope there will be a new edition, and that the excellent authorial team that Roose and Sackeim have brought together will extend their collaboration to address the needs of clinical decision makers. Meanwhile, I welcome the present edition as a useful source of review material for the budding or established academic.