Field of medicine: Epidemiology and the clinical specialties.

Format: Softcover book.

Audience: Written for busy clinical practitioners, this book is also useful for medical students and others who desire a brief, general overview of the fundamental epidemiologic and statistical tools necessary for understanding and critically evaluating medical literature.

Purpose: To "demystify" statistics and "provide clinicians with the practical skills to evaluate studies in the medical literature or to conduct studies for publication."

Content: The 16 chapters of this book are grouped into four sections: "Principles of Estimation," "Principles of Hypothesis Testing," "Application to Evaluating the Literature," and "Application to Conducting a Study." The first chapter covers basic descriptive population statistics and explains sample estimates, bias, and confidence intervals. The next eight chapters cover validity and predictive value in screening tests, study and clinical trial design, validity and reliability issues, probability, and basic computational tools for test statistics and for determining statistical power and establishing significance. An excellent stand-alone summary chapter on univariate and multivariate statistics includes a discussion of life tables and Kaplan-Meier approaches. Another chapter dwells on the use of confidence intervals in relative risks, odds ratios, and other hypothesis-testing applications. Chapters 10 through 12 provide more depth by revisiting hypothesis testing using a variety of examples that the reader often encounters: randomized clinical trials, case–control studies, cross-sectional cohorts, and meta-analysis. The final section of the book provides a methodical, stepwise approach to research, from conceptualizing the study to writing the manuscript. The book concludes with an excellent appendix that provides sample research protocols and data forms and crucial reminders for determining sample size methods of randomization.

Highlights: The material is presented in a practical, concise format for easy reference and includes a systematic flow chart guide for choosing appropriate statistical tests and measures. A handy detachable and reproducible list of questions for evaluating the literature and sections on sample size determination, randomization, data collection format, and publishing are also included. Illustrations and brief hypothetical studies allow for some critical application of the principles presented. The spiral-bound handbook format provides numerous graphs, charts, tables, and formulas commonly used in statistical analysis. Chapters on hypothesis testing and principles of estimation are particularly well written.

Limitations: The focus of the book on busy clinicians sometimes results in oversimplification of concepts. Although the book would be a useful adjunct text for journal clubs or "short courses" in epidemiologic and statistical methods, it necessarily lacks depth and detail.

Context: For many years, Huff's classic How To Lie with Statistics (WW Norton, 1984) stood virtually alone in this subgenre. Books such as Riegelman and Hirsch's Studying a Study and Testing a Test (Little, Brown, 1996) and Statistical First Aid: Interpretation of Health Research Data (Blackwell Science, 1992) have a similar yet classic approach designed for an audience less familiar with research principles. These texts are an alternative to the more focused and comprehensive introductory treatment given by Lilienfeld and Lilienfeld's Foundations of Epidemiology (Oxford Univ Pr, 1980) or Altman's Practical Statistics for Medical Research (Chapman & Hall, 1991).

Reviewer: Lou Fintor, MA, MPH, Centers for Disease Control and Prevention, Morgantown, West Virginia.

Commentary: This book addresses the need for broader appreciation of epidemiologic and statistical methods as applied to clinical research. It briefly and painlessly arms busy clinicians with basic skills useful not only for critically evaluating the medical literature but also for becoming more actively involved in clinical research design and execution. Hence, it succeeds in engendering stronger interest in biostatistical methods and critical thinking. It offers hope for those who have all but given up on understanding the vagaries of statistics and makes it easier to become comfortable with statistical concepts.