In preparing this third edition of An Introduction to Medical Statistics, I have taken the opportunity to correct a number of mistakes and typographical errors, and to change some of the examples and add a few more. I have extended the treatment of several topics and introduced some new ones, previously omitted through lack of space or energy, or because they were then rarely seen in the medical literature. In one case, number needed to treat, the concept had not even been invented when the second edition was written. Other new topics include consent in clinical trials, design and analysis of cluster- randomized trials, ecological studies, conditional probability, repeated testing, random effects models, intraclass correlation, and conditional odds ratios. Thanks to the wonders of computerized typesetting, I have managed to extend the contents of the book with a very small increase in the number of pages. This book is for medical students, doctors, medical researchers, nurses, members of professions allied to medicine, and all others concerned with medical data. The range of statistical methods used in the medical and health care literature, and hence described in this book, continues to grow, but the time available in the undergraduate curriculum does not. Some of the topics covered here are beyond the needs of many students, so I have indicated by an asterisk sections which would not usually be included in first courses. These are intended for postgraduate students and medical researchers.
