Multivariate Methods in Epidemiology
The basis for much of medical public health practice comes from epidemiological research. This text describes current statistical tools that are used to analyze the association between possible risk factors and the actual risk of disease. Beginning with a broad conceptual framework on the disease process, it describes commonly used techniques for analyzing proportions and disease rates. These are then extended to model fitting, and the common threads of logic that bind the two analytic strategies together are revealed. Each chapter provides a descriptive rationale for the method, a worked example using data from a published study, and an exercise that allows the reader to practice the technique. Each chapter also includes an appendix that provides further details on the theoretical underpinnings of the method. Among the topics covered are Mantel-Haenszel methods, rates, survival analysis, logistic regression, and generalized linear models. Methods for incorporating aspects of study design, such as matching, into the analysis are discussed, and guidance is given for determining the power or the sample size requirements of a study. This text will give readers a foundation in applied statistics and the concepts of model fitting to develop skills in the analysis of epidemiological data.
What people are saying - Write a review
We haven't found any reviews in the usual places.
Study Desisn and New Directions
Appendix 1 Theory on Models for Disease
Appendix 2 Theory on Analysis of Proportions
Appendix 3 Theory on Analysis of Rates
Appendix 4 Theory on Analysis of Time to Failure
Other editions - View all
adjusted alternative Ao(t approach association assumption binary calculations cancer case-control study cell censored Chapter chi-square distribution compared consider constant hazard controls covariates data in Table described difference effect Epidemiology equation example expected number follow-up frequencies hazard function hazard ratio Hence individual interaction likelihood ratio test linear logistic model linear model linear predictor linear trend log likelihood log odds ratio log-linear hazard log-linear model logistic regression matched matrix maximum likelihood estimates methods null hypothesis number of failures observed obtain overall pairs Parameter Estimates Pearson chi-square Poisson Poisson distribution population proportional hazards model regression model regression parameters regressor variables relationship relative risk represent response risk factor sample scaled deviance score statistic score test shown in Figure shown in Table specified standard error strata stratum subjects summary survival curve survival function tion variance Wald test Weibull