Medical Statistics: For Beginners by Dr H. K. Ramakrishna is, just as its title implies, a guide for the beginning researcher, or the medical student or resident interested in readying a manuscript or starting off for the first time conducting a study, to understand basic statistics.
The book is divided into 10 chapters. The first 2 chapters briefly describe Dr Ramakrishna’s bibliography and why and how he undertook the challenge of writing a book for those less experienced in medical statistics. Chapters 3 and 4 present an overview of basic statistical terms and concepts. I especially appreciate the intuitively important note on the historically “arbitrary” selection of a P value of <.05 for significance testing, and it echoes well with the recent American Statistician Association’s statement on P values.1 As advised, a beginner should refrain from chasing a significant P value (eg, <.05) without also paying attention to reporting measures of effect size.
In chapter 5, several basic statistical tests of significance are discussed, including the t test and Mann-Whitney-Wilcoxon test for continuous variables, and the χ2 test and Fisher exact test for categorical variables. Chapter 6 introduces concepts of study effects (eg, risk ratios and odds ratios) and correlation, and describes a number of regression methods including linear, logistic, Poisson, and survival regression. Despite considerable effort, these regression methods are not well described and lack enough rigor. For an in-depth exploration of regression methods, I recommend Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models.2
Chapters 7 and 8 discuss to a great extent how to design a study or clinical trial, for publication or for use in a dissertation, and how to write articles for journals. The breadth of subject matter coverage, extensive guidance, and advice on practical issues are impressive. Chapter 9 provides a short introduction to evidence-based medicine. With the right amount of insightful commentary from Dr Ramakrishna on the subject, this chapter could prove to be helpful for readers to learn to think critically with regard to evidence-based medicine. The final chapter presents an enlightening example illustrating the general principles discussed in chapters 7 and 8 of conducting a research project, with specific instructions focused on developing a protocol, presenting results, and writing a full-length report for consideration for publication. In my view, this chapter would have been well received had it been placed immediately after chapters 7 and 8. Nonetheless, the last 4 chapters are outstanding in terms of their depth of coverage and clinical relevance, and I commend Dr Ramakrishna for a job well done.
The strength of this well-written book is its emphasis on concepts and practice. All motivating examples throughout the text are coupled with illustrative tables and graphs that are readily reproducible using Excel features and online calculators. The step-by-step pedagogical instructions and engaging in-depth discussions effectively aid readers in developing conceptual understanding and critical appraisal. Despite the strengths of this book, there are a few minor problems. For example, in chapter 4, the definitions of sensitivity and specificity are mistakenly assumed to be the same as positive predictive value and negative predictive value, respectively. In addition, some contents in the book need to be consolidated. For instance, the descriptions of correlation, analysis of variance, and rank test in chapter 6 belong in chapter 5. Furthermore, the concept of statistical power is not covered in any detail. As statistical power and sample size calculation play a very important role in modern clinical research, there is a real need for a section devoted to the extensive assessment of the key components of statistical power, including the precision or variability of an outcome measure, the size of clinically significant effect to detect, type I and II error rates, and type of tests.
Overall, I think this book is highly suitable for a medical beginner who needs a solid reference to learn about a wide variety of basic epidemiological and statistical knowledge in planning, conducting, and interpreting clinical studies. However, this might not be the book for those who are already familiar with the general principles of statistics or those who have taken introductory courses in statistics and experimental design.
Feng Dai, PhD, MS
Department of Biostatistics
Yale Center for Analytical Sciences
Yale School of Public Health
New Haven, Connecticut
1. Wasserstein RL, Lazar NA. The ASA’s statement on P-values: context, process, and purpose. Am Stat. 2016;70:129–133.
2. Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE. Regression Methods in Biostatistics: Linear, Logistic, Survival, and Repeated Measures Models. 2005.Heidelberg, Germany: Springer.