Evidence-Based Diagnosis (Practical Guides to Biostatistics and Epidemiology)

This book should be on every clinical researcher's bookshelf. By the time I reached page 40 of the library copy, I knew I had to have my old to underline, make notes and insert flags. Already I had found one of the clearest explanations of problems with assessing agreement that elude so many researchers. What followed were more clear, easy-to-follow explanations of many complicated, difficult-to-understand Issues such as the meaning and use of p-values and confidence intervals. There is even a section on how to construct clinical decision rules. Within a week of reading this book, I found occasion to refer putative authors to it for the only detailed explanation I had seen of a statistical procedure (Bland-Altman plots) they had performed but did not undersand how to use. Other occasions for use arise every day.