Fact-finding missions in the medical jungle

Evidence-based medicine is all the rage at the moment. Almost all trainee physicians at undergraduate and postgraduate levels as well as nurse trainees at both undergraduate and postgraduate levels and others in professions allied to medicine are being encouraged, indeed almost required, to practise according to evidence-based medicine.

This does not simply mean practice according to protocols and guidelines set up by experts based on supposed evidence, but that the next generation of trainees should be equipped to search for relevant evidence themselves and be able to appraise critically the articles that turn up in these searches. To this end, a number of books have been published in recent years such as Trisha Greenhalgh's How to Read a Paper and Evidence-Based Health Care Workbook .

This volume pitches itself as an evidence-based medicine textbook for those who are nervous about mathematics and statistics. However, it comes much closer to being a conventional statistics book than, for example, the Greenhalgh volumes mentioned. It is a useful book in that it contains descriptions of most of the tests and most of the terms and procedures that are in use and as such might function well as a reference volume.

The book comes with a CD that contains multiple-choice questions and answers, as well as published journal articles that are used to illustrate many of the key points. But neither the chapters nor the material on the CD engages this reader readily.

In terms of equipping clinicians to appraise critically the scientific literature, this volume falls short of others in the field. It does not specifically alert the reader to the tricks commonly used by pharmaceutical companies and others, such as the use of categorical as opposed to continuous variables, or the reporting of treatment successes on a raft of secondary outcome measures while neglecting to report failures on the primary outcome measure.

This book tries to do too much and so leaves out the critical touch that might engage the reader.

Almost all commonly used statistical tests are described without pointing out that there are fundamental differences between approaches that adopt confidence intervals and ones that use t-tests. The commonly accepted hierarchies of evidence and sources of bias are outlined, but the reader is given no hint as to where in the hierarchy of evidence to slot ghost-written articles, for instance.

In other words, this book is not a guerrilla guide to evidence in medicine; it assumes that everyone is playing pretty well by the rules.

While none of the other textbooks in this area is a guerrilla guide, a number of them look better crafted to help those who need the next best thing, if only for the purposes of getting through professional exams.

David Healy is professor of psychiatry, Cardiff University.