Depression in Primary Care: Evidence and Practice Edited by Simon Gilbody and Peter Bower Cambridge University Press, 2010, £29.99 pb, 172 pp. ISBN: 9780521870504

If you don’t really follow what a forest plot is, and the difference between cost-effectiveness and cost offset is still a mystery to you, this book is excellent. It will help you to understand the science behind evidence-based medicine by taking you through a set of worked examples. The book is based on the extensive work both authors have done over the past few years in systematically reviewing ways of improving quality of care for people with depression in primary care. It manages to be academically rigorous, relevant to developments in health policy and readable into the bargain, even for someone like me with acknowledged limitations in numeracy. The authors are not afraid to remind the reader of the basics as well as stretching their brains. If you are one of their many fans, as I am, you will be pleased to find so much of what they have produced distilled into a single volume which has a narrative arc, from theory through evidence synthesis and on to implementation. I particularly found the section on the evidence for cognitive-behavioural therapy in primary care as opposed to specialist care settings very interesting; the outcomes are nowhere near as impressive, but never mentioned when more investment is being justified.

What is fascinating, however, is that the authors are honest enough to admit that, despite the ‘systematic’ nature of the review process, there are still potentially areas of disagreement. The discussion of how decisions about ‘strength of recommendations’ are made still has a smack of alchemy about it. I would like to have seen more about what can be learnt from synthesis of qualitative data, and the problems of integrating patient experience into conventional systematic reviews - something which is only briefly touched upon, but that is a small quibble. However, at the end of the book I found myself asking why we always seem to have to show a ‘cost offset’ for an intervention for mental health in primary care? Why do we have to show, for example, that treating people with diabetes who have depression saves money somewhere else in the healthcare system? Well we do, but we would not have to make that justification for chest pain, would we?