Statistical Models for Longitudinal Studies of Health

Longitudinal studies of health outcomes and their risk factors are an increasingly important source of knowledge in epidemiology, public health and clinical medicine. However, many of the statistical procedures suited to the analysis of findings from these studies have emerged only recently. This volume addresses theoretical and practical issues, demonstrating techniques with empirical examples. The topics covered include the specification and estimation of time-related statistical models involving continuous and categorical variables, fixed and random effects, variable coefficients, unconditional and conditional (autoregressive) structures, missing data and errors attrition, non-sampling errors of measurement, latent variables, reciprocal effects, stochastic differentials, empirical Bayes, weighted least squares estimation and biologic processses. Methodological issues in the design of large longitudinal surveys are also discussed.