Disease mapping with WinBUGS and MLwiN / Andrew B. Lawson, William J. Browne, Carmen L. Vidal Rodeiro.

Format:

Book

Author/Creator:

Lawson, Andrew (Andrew B.)

Contributor:

Browne, William J., Ph.D.

Vidal Rodeiro, Carmen L.

Series:

Statistics in practice

Language:

English

Subjects (All):

Medical geography.

Medical geography--Maps--Data processing.

Epidemiology--Statistical methods.

Epidemiology.

Epidemiology--Data processing.

Public health surveillance.

Epidemiologic Methods.

Mathematical Computing.

Models, Statistical.

Software.

Medical Subjects:

Epidemiologic Methods.

Mathematical Computing.

Models, Statistical.

Software.

Physical Description:

xiii, 277 pages : illustrations, maps ; 24 cm.

Place of Publication:

Chichester, West Sussex, England ; Hoboken, NJ : J. Wiley, [2003]

Summary:

Disease mapping involves the analysis of geo-referenced disease incidence data and has many applications, for example within resource allocation, cluster alarm analysis, and ecological studies. There is a real need amongst public health workers for simpler and more efficient tools for the analysis of geo-referenced disease incidence data. Bayesian and multilevel methods provide the required efficiency, and with the emergence of software packages -- such as WinBUGS and MLwiN -- are now easy to implement in practice. Disease Mapping with WinBUGS and MLwiN provides a practical introduction to the use of software for disease mapping for researchers, practitioners and graduate students from statistics, public health and epidemiology who analyse disease incidence data.

Contents:

0.1 Standard notation for multilevel modelling xi

0.2 Spatial multiple-membership models and the MMMC notation xii

0.3 Standard notation for WinBUGS models xii

1 Disease mapping basics 1

1.1 Disease mapping and map reconstruction 2

1.2 Disease map restoration 3

2 Bayesian hierarchical modelling 17

2.1 Likelihood and posterior distributions 17

2.2 Hierarchical models 18

2.3 Posterior inference 19

2.4 Markov chain Monte Carlo methods 20

2.5 Metropolis and Metropolis-Hastings algorithms 21

2.6 Residuals and goodness of fit 26

3 Multilevel modelling 29

3.1 Continuous response models 30

3.2 Estimation procedures for multilevel models 35

3.3 Poisson response models 38

3.4 Incorporating spatial information 42

4 WinBUGS basics 45

4.2 Start using WinBUGS 47

4.3 Specification of the model 50

4.4 Model fitting 59

4.5 Scripts 64

4.6 Checking convergence 65

4.7 Spatial modelling: GeoBUGS 67

5 MLwiN basics 75

5.3 Fitting statistical models 84

5.4 MCMC estimation in MLwiN 94

5.5 Spatial modelling 104

6 Relative risk estimation 115

6.1 Relative risk estimation using WinBUGS 115

6.2 Spatial prediction 137

6.3 An analysis of the Ohio dataset using MLwiN 139

7 Focused clustering: the analysis of putative health hazards 155

7.2 Study design 156

7.3 Problems of inference 158

7.4 Modelling the hazard exposure risk 160

7.5 Models for count data 164

7.6 Bayesian models 166

7.7 Focused clustering in WinBUGS 167

7.8 Focused clustering in MLwiN 190

8 Ecological analysis 197

8.2 Statistical models 198

8.3 WinBUGS analyses of ecological datasets 199

8.4 MLwiN analyses of ecological datasets 219

9 Spatially-correlated survival analysis 235

9.1 Survival analysis in WinBUGS 235

9.2 Survival analysis in MLwiN 237

Appendix 1 WinBUGS code for focused clustering models 253

A.1 Falkirk example 253

A.2 Ohio example 256

Appendix 2 S-Plus function for conversion to GeoBUGS format 263.

Notes:

Includes bibliographical references (pages 267-273) and index.

ISBN:

0470856041

OCLC:

52410089

Online:

Publisher description