Epidemiologic Analysis A Case-Oriented Approach [Paperback]

Using real data from published sources, this engaging and lucid casebook shows how statistical tools can be used to analyze important epidemiologic issues. Its 18 cases address a variety of interesting research problems from Mendel's classic sweet pea experiments to recent studies of AIDS and exposure to electro-magnetic field radiation. Each includes a data set. The cases are described succinctly and the methods used to analyze them are then discussed in detail. A wide range of statistical and graphical tools are included, from simple mean values to nonparametric bivariate regression smoothing techniques. The level of discussion is sophisticated but mathematically simple, affording access to a broad audience interested in using collected data to study human health and disease. The author's focus on describing, interpreting and presenting results will set this book apart from other texts.

1. Measurement of trend (a 2 by c table)
2. Odds Ratio and Relative Risk
3. Randomized trial
4. Bias
Part I: Regression to the Mean
Part II: Misclassification
5. Goodness-of-fit
6. Analysis of a two-way table
7. Cluster analysis
8. Matched analysis (and case and two controls)
9. Multivariable linear regression
10. Linear logistic regression
11. Part I: Poisson regression (standardized mortality ratio)
Part III: Poisson regression (adjusted rates)
12. Log-linear model (Poisson model applied to tabular data)
13. survival analysis (Cox's proportional hazards model)
14. Multivariate comparison of two groups
15. Smoothing sequential data
16. Nonparametric regression analysis
Appendix