Biostatistical Methods: The Assessment of Relative Risks, 2nd Edition The Assessment of Relative Risks

Praise for the First Edition

". . . an excellent textbook . . . an indispensable reference for biostatisticians and epidemiologists."
--International Statistical Institute

A new edition of the definitive guide to classical and modern methods of biostatistics

Biostatistics consists of various quantitative techniques that are essential to the description and evaluation of relationships among biologic and medical phenomena. Biostatistical Methods: The Assessment of Relative Risks, Second Edition develops basic concepts and derives an expanded array of biostatistical methods through the application of both classical statistical tools and more modern likelihood-based theories. With its fluid and balanced presentation, the book guides readers through the important statistical methods for the assessment of absolute and relative risks in epidemiologic studies and clinical trials with categorical, count, and event-time data.

Presenting a broad scope of coverage and the latest research on the topic, the author begins with categorical data analysis methods for cross-sectional, prospective, and retrospective studies of binary, polychotomous, and ordinal data. Subsequent chapters present modern model-based approaches that include unconditional and conditional logistic regression; Poisson and negative binomial models for count data; and the analysis of event-time data including the Cox proportional hazards model and its generalizations. The book now includes an introduction to mixed models with fixed and random effects as well as expanded methods for evaluation of sample size and power. Additional new topics featured in this Second Edition include:
* Establishing equivalence and non-inferiority
* Methods for the analysis of polychotomous and ordinal data, including matched data and the Kappa agreement index
* Multinomial logistic for polychotomous data and proportional odds models for ordinal data
* Negative binomial models for count data as an alternative to the Poisson model
* GEE models for the analysis of longitudinal repeated measures and multivariate observations

Throughout the book, SAS is utilized to illustrate applications to numerous real-world examples and case studies. A related website features all the data used in examples and problem sets along with the author's SAS routines.

Biostatistical Methods, Second Edition is an excellent book for biostatistics courses at the graduate level. It is also an invaluable reference for biostatisticians, applied statisticians, and epidemiologists.

This book focuses on the comparison, contrast, and assessment of risks on the basis of clinical investigations. It develops basic concepts as well as deriving biostatistical methods through both the application of classical mathematical statistical tools and more modern likelihood-based theories. The first half of the book presents methods for the analysis of single and multiple 2x2 tables for cross-sectional, prospective, and retrospective (case-control) sampling, with and without matching using fixed and two-stage random effects models. The text then moves on to present a more modern likelihood- or model-based approach, which includes unconditional and conditional logistic regression; the analysis of count data and the Poisson regression model; the analysis of event time data, including the proportional hazards and multiplicative intensity models; and elements of categorical data analysis (expanded in this edition). SAS subroutines are both showcased in the text and embellished online by way of a dedicated author website. The book contains a technical, but accessible appendix that presents the core mathematical statistical theory used for the development of classical and modern statistical methods.

This book focuses on the comparison, contrast, and assessment of risks on the basis of clinical investigations. It develops basic concepts as well as deriving biostatistical methods through both the application of classical mathematical statistical tools and more modern likelihood-based theories. The first half of the book presents methods for the analysis of single and multiple 2x2 tables for cross-sectional, prospective, and retrospective (case-control) sampling, with and without matching using fixed and two-stage random effects models. The text then moves on to present a more modern likelihood- or model-based approach, which includes unconditional and conditional logistic regression; the analysis of count data and the Poisson regression model; the analysis of event time data, including the proportional hazards and multiplicative intensity models; and elements of categorical data analysis (expanded in this edition). SAS subroutines are both showcased in the text and embellished online by way of a dedicated author website. The book contains a technical, but accessible appendix that presents the core mathematical statistical theory used for the development of classical and modern statistical methods.