This book is not one that can be easily pigeon-holed. It is not a textbook of evidence, nor a theoretical model, nor even a practitioner's guide to mental health services. After puzzling about it for some time I think it is really a map. People who do not have maps tend to get lost unless they already know the terrain, and Graham Thornicroft and Michele Tansella have discerned that very few have such prior knowledge. One reason for this is that the names on the map are like the street names in former East European capital cities: they change every few years. Who reading this will recognise all the following: AIMHS standards, CLSCs, HOMOs and MINI? (They are all systems — from Australia, Canada, the USA and the UK respectively — that give guidance or standards for mental health services to a geographically defined population.)

The map (matrix) that the authors plot is a comprehensive one in two dimensions, space and time, with three levels of scale for each: country, catchment area and individual (patient) in the geographical dimension, and the input, process and outcome phases of care for the temporal dimension. The subsequent plot of nine cells is not perhaps a creation of true genius, which I might describe as the comprehensive elicitation of a universal truth, but it grows with the telling. Indeed David Goldberg, in a challenging foreword that illustrates how lopsided are our present mental health care systems, comments rightly that the book “fizzles with excitement”, and there are certainly many passages that evoke the OCNIAMS phenomenon (of course, now it all makes sense), which is around level three on the nine-cell path to genius.

The main value of the matrix model is that it allows direct comparison between services, resources and care for all mental health ‘events’, illustrated beautifully by Leon Eisenberg in his account, in the preface, of a 1986 homicide tragedy involving a patient with schizophrenia on a ferry boat in New York. He had been seen 4 days earlier in an emergency room and the psychiatric house officer concerned was unable to find a bed because the patient was uninsured and without funds. The subsequent inquiry focused on the failures at patient level; they did not address the wider issues of service, including national and state policy, which was largely responsible for the absence of beds.

Time after time this book illustrates the value of an epidemiological approach to mental health care. Thornicroft & Tansella have four aims: to unify the description of mental health services, to give a sequence of order to complex events, to aid understanding and to enable priorities to be set for improvements in services. Like good scientists they subject their model to ‘field testing’ by five colleagues in other countries: Lesage (Canada), Munk Jørgenson (Denmark), Rosen (Australia), Tomov (Bulgaria) and Warner (USA). They perform this task admirably and illustrate how the model can be used to compare the nine cells of the matrix in different countries. It is interesting to note that psychiatric service provision to a population shows virtually the same variation in Bulgaria and the USA, that in Denmark, resources and service provision for psychiatry are almost entirely controlled at county level, to the extent that the central Ministry of Health is emasculated and is referred to as the Ministry of Attitudes, and that liaison between community mental health teams and primary care in Quebec is almost non-existent.

This book deserves to be a success because it is articulating the voice of an emerging part of psychiatry with which I am proud to be associated — public mental health. The authors quote Geoffrey Rose, who complained that psychiatrists “seem generally unaware of the existence and importance of mental health attributes of whole populations, their concern being only with sick individuals”, and although he wrote this in 1993 the comment is still largely true today. I hope that practitioners and planners of mental health and social services will see the benefits of this model in understanding how resources should be allocated and distributed and will determine whether their own particular services are up to scratch in all elements of the model.

Perhaps the authors will be able to produce a second edition before long in which they could indulge in the fashionable (but still useful) practice of league tables, whereby they can examine services from countries throughout the world and place them in divisions of merit. After reading this book I think that Australia would be very near the top of the league, that the UK would probably be in the middle of the premier division and the USA would be challenging for promotion from the second.