Statistical Estimation of Epidemiological Risk

Statistical Estimation of Epidemiological Risk provides coverage of the most important epidemiological indices, and includes recent developments in the field. A useful reference source for biostatisticians and epidemiologists working in disease prevention, as the chapters are self-contained and feature numerous real examples. It has been written at a level suitable for public health professionals with a limited knowledge of statistics.

Other key features include:

• Provides comprehensive coverage of the key epidemiological indices.
• Includes coverage of various sampling methods, and pointers to where each should be used.
• Includes up-to-date references and recent developments in the field.
• Features many real examples, emphasising the practical nature of the book.
• Each chapter is self-contained, allowing the book to be used as a useful reference source.
• Includes exercises, enabling use as a course text.

About the author.

Preface.

1 Population Proportion or Prevalence.

1.1 Binomial sampling.

1.2 Cluster sampling.

1.3 Inverse sampling.

Exercises.

References.

2 Risk Difference.

2.1 Independent binomial sampling.

2.2 A series of independent binomial sampling procedures.

2.2.1 Summary interval estimators.

2.2.2 Test for the homogeneity of risk difference.

2.3 Independent cluster sampling.

2.4 Paired-sample data.

2.5 Independent negative binomial sampling (inverse sampling).

2.6 Independent poisson sampling.

2.7 Stratified poisson sampling.

Exercises.

References.

3 Relative Difference.

3.1 Independent binomial sampling.

3.2 A series of independent binomial sampling procedures.

3.2.1 Asymptotic interval estimators.

3.2.2 Test for the homogeneity of relative difference.

3.3 Independent cluster sampling.

3.4 Paired-sample data.

3.5 Independent inverse sampling.

Exercises.

References.

4 Relative Risk.

4.1 Independent binomial sampling.

4.2 A series of independent binomial sampling procedures.

4.2.1 Asymptotic interval estimators.

4.2.2 Test for the homogeneity of risk ratio.

4.3 Independent cluster sampling.

4.4 Paired-sample data.

4.5 Independent inverse sampling.

4.5.1 Uniformly minimum variance unbiased estimator of relative risk.

4.5.2 Interval estimators of relative risk.

4.6 Independent poisson sampling.

4.7 Stratified poisson sampling.

Exercises.

References.

5 Odds Ratio.

5.1 Independent binomial sampling.

5.1.1 Asymptotic interval estimators.

5.1.2 Exact confidence interval.

5.2 A series of independent binomial sampling procedures.

5.2.1 Asymptotic interval estimators.

5.2.2 Exact confidence interval.

5.2.3 Test for homogeneity of the odds ratio.

5.3 Independent cluster sampling.

5.4 One-to-one matched sampling.

5.5 Logistic modeling.

5.5.1 Estimation under multinomial or independent binomial sampling.

5.5.2 Estimation in the case of paired-sample data.

5.6 Independent inverse sampling.

5.7 Negative multinomial sampling for paired-sample data.

Exercises.

References.

6 Generalized Odds Ratio.

6.1 Independent multinomial sampling.

6.2 Data with repeated measurements (or under cluster sampling).

6.3 Paired-sample data.

6.4 Mixed negative multinomial and multinomial sampling.

Exercises.

References.

7 Attributable Risk.

7.1 Study designs with no confounders.

7.1.1 Cross-sectional sampling.

7.1.2 Case–control studies.

7.2 Study designs with confounders.

7.2.1 Cross-sectional sampling.

7.2.2 Case–control studies.

7.3 Case–control studies with matched pairs.

7.4 Multiple levels of exposure in case–control studies.

7.5 Logistic modeling in case–control studies.

7.5.1 Logistic model containing only the exposure variables of interest.

7.5.2 Logistic regression model containing both exposure and confounding variables.

7.6 Case–control studies under inverse sampling.

Exercises.

References.

8 Number Needed to Treat.

8.1 Independent binomial sampling.

8.2 A series of independent binomial sampling procedures.

8.3 Independent cluster sampling.

8.4 Paired-sample data.

Exercises.

References.

Appendix Maximum Likelihood Estimator and Large-Sample Theory.

A.1: The maximum likelihood estimator, Wald’s test, the score test, and the asymptotic likelihood ratio test.

A.2: The delta method and its applications.

References.

Answers to Selected Exercises.

Index.

"…a concise, organized, and well-written text that provides the derivations of statistical formulas underlying much epidemiological research and practice." (Journal of the American Statistical Association, December 2005)

"...presents a considerable amount of recent research, much of which is the author's own..." (Royal Statistical Society, Vol.168, No.1, January 2005)

"...systematically organised...an excellent reference..." (Short Book Review, Vol.24, No.3 December 2004)

"...this book is strongly recommended..." (Statistical Methods in Medical Research, Vol 14 2005)

KUNG-JONG LUI is a professor in the Department of Mathematics and Statistics at San Diego State University. Since he obtained his Ph.D. in biostatistics from UCLA in 1982, he has published more than 100 papers in peer-reviewed journals, including Biometrics, Statistics in Medicine, Biometrical Journal, Psychometrika, Communications in Statistics: Theory and Methods, Science, Proceedings of National Academy of Sciences, Controlled Clinical Trials, Journal of Official Statistics, IEEE Transactions on Reliability, Environmetrics, Test, Computational Statistics and Data Analysis, American Journal of Epidemiology, American Journal of Public Health, etc. He is a Fellow of the American Statistical Association, a life member of the International Chinese Statistical Association, and a member of the Western North American Region of the International Biometric Society. Statistical estimation of indices for measuring risk is a key topic in epidemiology. A good estimator, that is unbiased and efficient, can help investigators search for the possible causes of disease. Good use of statistical methods can enable public health administrators to confidently allocate their limited resources to the appropriate methods of prevention and treatment. Statistical Estimation of Epidemiological Risk presents the most-commonly used measures of risk, and adopts a practical approach using many real and numerical examples to support the methodology.

• Presents a practical overview of the key measures of epidemiological risk.
• Features coverage of various sampling methods, and pointers to where each should be used.
• Each measure discussed is supported by a number of real and numerical examples that highlight their practical use.
• Each chapter is self-contained, allowing the book to be used as a reference source.
• Includes an abundance of exercises, which give the reader a clearer understanding of the theory.
• Suitable for epidemiologists and public health professionals with a modest statistical background.

Statistical Estimation of Epidemiological Risk is both a useful practical reference for researchers from biostatistics and epidemiology, and an accessible textbook for graduate students studying epidemiological risk. The range of examples and exercises are of great practical benefit to public health professionals working in disease prevention and control, and set this book apart from others on the topic.