Handbook of Statistical Methods for Case-Control Studies

Ørnulf Borgan is Professor of Statistics, University of Oslo. His book with Andersen, Gill and Keiding on counting processes in survival analysis is a world classic.

This handbook provides an in-depth treatment of up-to-date and currently developing statistical methods for the design and analysis of case-control studies, with a primary focus on case-control studies in epidemiology. Authors will be encouraged to illustrate the statistical methods they describe by application to datasets that are either already publicly available or can be made so. A handbook website will be established that provides readers access to those datasets and software used to analyze them. Wherever feasible, authors will be asked to construct their examples using the R statistical language, though programs in SAS, Stata or other common languages will also be accepted.

Introduction. Introduction. Origins. Classical Case-Control Studies. Design issues in case-control studies. Basic concepts and methods of analysis. Matched samples. Beyond logistic regression. Small sample methods. Multiple case or control groups. Power and sample size. Causal inference. Misclassification and measurement error. Analysis of secondary phenotype under case-control design. Sampling from a Defined Cohort. Two and three (or multi) phase sampling designs. Calibration and estimation of sampling weights. Maximum likelihood. Re-use of case-control samples. Misspecification. Case-control studies with complex sampling. Cohort sampling for time to event data. Case-cohort designs and analyses. Design options and partial likelihood analyses of nested case-control data. Inverse probability weighting in nested case-control studies. Multiple imputation. Maximum likelihood. Self controlled case series. Genetic Epidemiology. Basic design and association analysis of population-based case-control studies. Analysis of gene-environment interactions. Screening methods for detecting genetic association and interactions under case-control design. Analysis of family-based case-control studies. Fitting mixed model to case-control genome-wide association studies. Analysis of secondary phenotype under case-control design.