Skip navigation

Book Reviews



Practising Evidence-based Medicine . Sharon Strauss, Douglas Badenoch, Scott Richardson, William Rosenberg, David Sackett. Radcliffe Medical Press, Oxford, 1998. 80pp (email medical@radpress.win-uk.net ).

This A4-format loose-leaf book is a seven session course for clinicians who want to learn more about the practical issues around incorporating critical appraisal of the best available evidence with their individual expertise. It complements other publications for the NHS Centre for EBM in Oxford. Because Bandolier is interested in diagnosis and test use, we turned to the example of an elderly woman with pneumonia and a low haemoglobin, and the use of a ferritin test in helping make a diagnosis.

We are provided with a step-by-step approach in which the clinical problem is outlined, with searching for evidence. The key papers are readily available in most libraries to be read with the example and worked through. There follows the "experts" answers. It was straightforward, sensible, and illuminating. It would make any of us more comfortable about tackling another such problem. The only thing we found strange was why, as it says in the book, it should take 10 days to obtain a ferritin result, for a test whose analysis time is measured in minutes or hours.

This is one of a series, and many people will benefit from these great tutorials for which there are learner and tutor manuals. The prices (for the complete, self-contained course) are: learners' and tutors' manuals at £25 each or a set of 10 learners with 1 tutor's manual for £200.

Statistical Issues in Drug Development . Stephen Senn. John Wiley & Sons, Chichester, 1997. 423pp, ISBN 0-471-97488-9, £50.

Now this book is not for the general reader, but it is a great book. Bandolier is beginning to worry about getting enthused by statistics books of late, but perhaps that marks a sea-change in the way they are written. Stephen Senn has a wry sense of humour, and a terrific turn of phrase that makes the book interesting - "Statisticians know that words are important to statistics, yet surely their importance is not fully recognised" ; a quote from Kruskal in the Introduction which suggest that Senn, at least, has a real insight into the use and importance of words. Yes, there are lots of funny symbols in parts, but there is much wisdom without too much mining.

The first part of the book is given over to chapters of direct importance to clinical trials and drug development, and are of interest to any of us involved in designing, conducting, or interpreting clinical trial data. The brief history of medical statistics is also illuminating and interesting. For instance, Thomas Bayes' (of Bayes theorem) key work was published after his death by Richard Price. One of Bandolier 's editors grew up in the same village in which Price was born - what are the odds of that happening?

The second part of the book is concerned with debatable and controversial topics in drug development. This includes topics like intention to treat, one and two sided tests, multi-centre trials and meta-analysis, and even includes a section of pharmacoeconomics. " Pharmaco-economist: one who asks, not only if the treatment for dysentery was effective, but also after the price of toilet paper ." Senn scatters his book with fun quotes and jokes, and those of you who lecture often could do worse than get the book for these alone. They are very witty.

" Ideally, nobody should study statistics who hasn't studied it already ". Senn's words, and wise ones. This book is one that many people involved in trials and use of evidence would profit by reading. It may not quite be something for the beaches of Spain, but it is very, very close to it. One final quote: " Pharmaco-vigilance: a game of hunt the thimble, in which you are not sure if it is a thimble you are looking for, you don't want to find it anyway and the only time anyone shouts `warm' is when you have already burnt your fingers ".

An evidence-based resource for pain relief . Henry McQuay and Andrew Moore. Oxford University Press, Oxford, 1998. 264pp, ISBN 0-19-262718-X. £65.

This A4-format book is written by two of Bandolier 's editors, so it would be disingenuous to say more than what it contains. It draws together considerable work that has been going on for some years by researchers in Oxford and around the world in systematic review methods and results in pain relief. The question of which analgesic is the best in which particular pain context is one that has troubled researchers and clinicians for years, and still does.

The book is in three parts. The methodology - involving issues of validity and quality in pain trials, plus combining data - will particularly interest people doing trials and interpreting them. The sections on acute pain and chronic pain draw together a series of systematic reviews, with tables of all quality trials. Introductions and conclusions place the knowledge gained from the systematic reviews into clinical perspective.

Systematic Reviews. Synthesis of best evidence for health care decisions . Cynthia Mulrow and Deborah Cook. American College of Physicians, Philadelphia, 1998. 115pp. ISBN 0-943126-66-5. (no price available).

This is a collection of series of essays on systematic reviews published through 1997 in Annals of Internal Medicine for a general physician audience. They cover a range of issues, such as the synthesis of best evidence for clinical decisions, using systematic reviews in education and guidelines, and how consumers and policymakers can use systematic reviews.

The language is straightforward, and largely free of jargon. There are some interesting thoughts in these essays, and they contains some rules about how systematic reviews should be conducted and used.

Sophisticates will find some points with which to argue, but the book is a good introduction to systematic reviews - no, it is more than an introduction, perhaps an improvers primer, because most of us have by now been introduced to some of the basic concepts. The book (which is available through BMJ Publications) is useful for the departmental library shelf, and will be frequently consulted.




previous or next story in this issue