Book Review of " How to Report Statistics in Medicine"
The book " How to Report Statistics in Medicine" is written by Tom Lang and Michelle Secic in 1997, published by the American College of Physicians (ACP). Just as Edward J. Huth wrote in his foreword, " ........... physician who knows nothing about statistical methods expect to find in reports of clinical trials of drugs statistical evidence ................." "Unfortunately, what passed before our eyes as statistical analysis and reporting does not always represent the proper use of statistical methods or the clear and adequate reporting of statistical findings..., and the review system is not always infallible in judging statistical evidence and how it is presented." "Up until now, authors have had available little published guidance in how to report most effectively their statistical data." Under all this circumstances, the book came into being, aiming to bring valuable specific and detailed help to authors who wish to make their papers as statistically convincing as possible. In fact, this book is also written for medical writers and editors, authors reporting basic or clinical research, clinicians, residents, and students in all areas of medicine and health science, including nursing and allied health professionals. The first author of the book, Tom Lang, is Manager of Medical Editing Services at The Cleveland Clinic Foundation, where he supervises the editing of scientific manuscripts for publication in peer-reviewed journals. The second author of the book, Michelle Secic, is the Senior Biostatistician in the Department of Biostatistics and Epidemiology at The Cleveland Clinic Foundation, where she assists researchers in designing, analyzing, and interpreting medical research studies.
Here are some of the strengths and weakness of the book:
1. There are basically 4 major parts in the book: annotated guidelines to for reporting statistical information, guide to statistical terms and tests, an uninitiated, reference list of the guidelines, and 4 useful appendices. The guidelines are presented in very clear and easy-to -find way, marked by different signs and type fonts in alphabetical order. To me, part two is especially useful. Every time I do not know the exact meaning of a specific term in statistics, I can go directly to this part and look at the explanation and get a better idea. For example, the term "intention-to-treat analysis" is right there on page 262.
2. In part 4 of the book, the appendix on " Rules for Presenting Numbers in Text" provides useful information, since numbers are most commonly used in reporting results.
3. Part 1 is the key part of the book, which contains many specific topics on reporting statistics. For example, after looking through Section 6 of Part 1, "Testing for Relationships: Reporting Associations and Correlation Analysis", I know the function of correlation matrix and how to interpret different correlation values, and then check whether the author has come to a positive conclusion or not. The same is true of the knowledge about 95% CI, with which I may help medical researchers come to a more convincing conclusion.
4. This book is very carefully proofread and so far I have not found any typing error.
5. One weakness of the book is, I think, that it would probably be better for the book to come up with some exercises or a separate workbook, providing any one who wants to learn from this book some chances of practice. Here what I mean by exercise is not how to calculate a specific statistical value, and I am referring to the exercises that help readers to think of the correct report of statistics and to what extent they
6. Another minor limitation of the book is that it is too heavy to take as a portable reference book. If it was printed on thinner paper, I would like it even better.