Article Text

Download PDFPDF

Statistical analysis of epidemiologic data, 3rd ed
  1. M Delgado-Rodriguez

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    S Selvin, Oxford University Press, £39.95, pp 492. ISBN 0-19-517280-9

    This considerably updated edition contains common procedures and methods that are difficult to find in a single textbook. All the procedures are interestingly illustrated with one or more examples.

    The first chapter is dedicated to measures of risk (rate, risk, incidence, and prevalence). A difference with other similar textbooks is the inclusion in the chapter of basic measures in survival analysis and its relations with other common risk measures in epidemiology. An introduction to smoothing, adjusting, and transforming is also offered. Chapter 2 considers variation and bias from a statistical point of view. It begins with the definition of a simple model to compare two groups, and an explanation of how selection bias, ecological bias and, confounding can be assessed. After that the model is extended, interaction contrasts are defined to end with the measurement of misclassification bias. The third chapter deals with statistical power and sample size, in which explanations for most common situations in epidemiology are offered. Chapter 4 (cohort data) is mainly dedicated to cohort effect and how to measure it with several examples. The analysis of spatial data is described in chapter 5, where details of the most common methods are given: nearest neighbour, transformed maps, spatial distribution about a point, and time/distance spatial analysis. Chapter 6 displays the analysis of contingency tables (2×k and 2×2×2) and it is mainly focused in the testing of homogeneity and independence of data. The methods of logistic regression are clearly shown in chapters 7 and 8, starting with the simplest case and ending with the problem of collinearity. The Poisson model is fully described in chapter 9 and appendix B. Chapter 10 deals with the analysis of matched data and the last chapters explain in a useful way survival analysis.

    For a full understanding of the text an intermediate level in biostatistics and epidemiology is required. The book is a useful tool for teachers, doctoral students, and professionals who need a more thorough understanding of common statistical procedures in epidemiology and public health.