Practice of Medical Statistics
S. Kumaresan *
Institute of Community Medicine, Madurai Medical College, Madurai District, Tamil Nadu- 625020, India.
*Author to whom correspondence should be addressed.
Abstract
This book aims to develop practical knowledge of and competence in applying research methodology and medical statistical procedures among students, researchers, healthcare professionals, and academics. It provides a comprehensive understanding of descriptive statistics, including data collection, organisation, presentation, measures of central tendency and dispersion, and the skewness and kurtosis of frequency distributions.
The book discusses relationship-based statistical analyses using correlation and regression techniques. In inferential statistics, it explains various parametric, non-parametric, and chi-square tests through step-by-step procedures and numerous illustrative examples.
The section on experimental design presents detailed procedures and flowcharts for analysing treatment effects using one-way ANOVA, two-way ANOVA, completely randomised design (CRD), randomised block design (RBD), and Latin square design (LSD). In addition, 2² and 2³ factorial designs are introduced for medical research to estimate main and interaction effects. Blocking and confounding techniques in factorial experiments are also discussed in detail. The adequacy of simple and multiple regression models is evaluated using statistical software.
The book also covers a wide range of medical research methodologies and study designs. It discusses probability and non-probability sampling techniques with appropriate examples and explains methods for determining sample sizes for various types of qualitative and quantitative data.
Furthermore, the book examines sources of vital statistics data, methods for measuring fertility and mortality, population projection techniques, and the construction and application of life tables. The final chapter focuses on evaluating diagnostic tests, including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), likelihood ratios, and receiver operating characteristic (ROC) curve analysis, highlighting their importance in evidence-based medical research and clinical decision-making.
Keywords: Biostatistics, medical statistics, epidemiology, research methodology, sampling techniques, sample size determination, probability distributions, hypothesis testing, correlation and regression, analysis of variance, non-parametric tests, diagnostic test evaluation, demography, vital statistics, clinical research, study design, randomised controlled trials, public health research, data analysis