1. INTRODUCTION AND CAUSAL CONCEPTS 1.1 Introduction p2 1.2 A brief history of multiple causation concepts p2 1.3 A brief history of scientific inference p6 1.4 Key components of epidemiologic research p9 1.5 Seeking causes p10 1.6 Models of causation p11 1.7 Counterfactual concepts of causation for a single exposure p18 1.8 Experimental versus observational evidence of causation p22 1.9 Constructing a causal diagram p23 1.10 Causal criteria p25 2. SAMPLING 2.1 Introduction p35 2.2 Non-probability sampling p36 2.3 Probability sampling p39 2.4 Simple random sample p39 2.5 Systematic random sample p40 2.6 Stratified random sample p40 2.7 Cluster sampling p41 2.8 Multistage sampling p42 2.9 Targeted (risk-based) sampling p43 2.10 Analysis of survey data p44 2.11 Sample-size determination p48 2.12 Sampling to detect disease p55 3. QUESTIONNAIRE DESIGN 3.1 Introduction p62 3.2 Designing the question p64 3.3 Open question p65 3.4 Closed question p65 3.5 Wording the question p69 3.6 Structure of questionnaires p69 3.7 Pre-testing questionnaires p70 3.8 Validation p71 3.9 Response Rate p71 3.10 Data-coding and editing p72 4. MEASURES OF DISEASE FREQUENCY 4.1 Introduction p78 4.2 Count, proportion, odds and rate p78 4.3 Incidence p79 4.4 Calculating risk p80 4.5 Calculating incidence rates p81 4.6 Relationship between risk and rate p83 4.7 Prevalence p84 4.8 Mortality statistics p85 4.9 Other measures of disease frequency p85 4.10 Standard errors and confidence intervals p87 4.11 Standardisation of risks and rates p89 5. SCREENING AND DIAGNOSTIC TESTS 5.1 Introduction p96 5.2 Attributes of the test per se p96 5.3 The ability of a test to detect disease or health p104 5.4 Predictive values p107 5.5 Interpreting test results that are measured on a continuous scale p109 5.6 Using multiple tests p115 5.7 Evaluation of diagnostic tests p117 5.8 Evaluation when there is no gold standard p121 5.9 Other considerations in test evaluation p125 5.10 Sample size requirements p127 5.11 Herd-level testing p128 5.12 Use of pooled samples p130 6. MEASURES OF ASSOCIATION 6.1 Introduction p140 6.2 Measures of association p141 6.3 Measures of effect p144 6.4 Study design and measures of association p147 6.5 Hypothesis testing and confidence intervals p147 6.6 Multivariable estimation of measures of association p152 7. INTRODUCTION TO OBSERVATIONAL STUDIES 7.1 Introduction p156 7.2 A unified approach to study design p159 7.3 Descriptive studies p161 7.4 Observational studies p162 7.5 Cross-sectional studies p164 7.6 Estimating incidence from one or more cross-sectional studies p168 7.7 Inferential limitations of cross-sectional studies p169 7.8 Repeated cross-sectional versus cohort studies p170 7.9 Reporting of observational studies p171 8. COHORT STUDIES 8.1 Introduction p180 8.2 Selecting the study group p182 8.3 The exposure p186 8.4 Disease as exposure p190 8.5 Ensuring exposed and non-exposed groups are comparable p190 8.6 Follow-up period p191 8.7 Measuring the outcome p191 8.8 Analysis p192 8.9 Reporting of cohort studies p194 9. CASE-CONTROL STUDIES 9.1 Introduction p202 9.2 The study base p202 9.3 The case series p205 9.4 Principles of control selection p207 9.5 Selecting controls in risk-based designs p207 9.6 Selecting controls in rate-based designs p209 9.7 Other sources of controls p214 9.8 The number of controls per case p215 9.9 The number of control groups p215 9.10 Exposure and covariate assessment p216 9.11 Keeping the cases and controls comparable p216 9.12 Analysis of case-control data p217 9.13 Reporting guidelines for case-control studies p218 10. HYBRID STUDY DESIGNS 10.1 Introduction p224 10.2 Case-crossover studies p224 10.3 Case-case studies p228 10.4 Case-case-control studies p229 10.5 Case-series studies p231 10.6 Case-cohort studies p233 10.7 Case-only studies p235 10.8 Two-stage sampling designs p237 11. CONTROLLED STUDIES 11.1 Introduction p244 11.2 Background, objectives, and summary trial design p246 11.3 Participants: the study group p247 11.4 Specifying the intervention p250 11.5 Measuring the outcome p251 11.6 Sample size p252 11.7 Allocation of study subjects p254 11.8 Follow-up/compliance p258 11.9 Statistical methods and analysis p259 11.10 Conclusions p262 11.11 Clinical trial designs for prophylaxis of communicable organisms p262 11.12 Reporting of clinical trials p265 12. VALIDITY IN OBSERVATIONAL STUDIES 12.1 Introduction p276 12.2 Selection bias p277 12.3 Examples of selection bias p281 12.4 Reducing selection bias p287 12.5 Information bias p288 12.6 Bias from misclassification p290 12.7 Validation studies to correct misclassification p297 12.8 Measurement error p297 12.9 Errors in surrogate measures of exposure p299 12.10 The impact of information bias on sample size p299 13. CONFOUNDING: DETECTION AND CONTROL 13.1 Introduction p308 13.2 Control of confounding prior to data analysis p311 13.3 Matching on confounders p311 13.4 Detection of confounding p316 13.5 Analytic control of confounding p322 13.6 Multivariable modelling to control confounding p328 13.7 Other approaches to control confounding and estimate causal effects p328 13.8 Propensity scores for controlling confounding p335 13.9 External adjustment and sensitivity analysis for unmeasured confounders p340 13.10 Understanding causal relationships p342 13.11 Summary of effects of extraneous variables p351 14. LINEAR REGRESSION 14.1 Introduction p360 14.2 Regression analysis p360 14.3 Hypothesis testing and effect estimation p362 14.4 Nature of the X-variables p368 14.5 Detecting highly correlated (collinear) variables p374 14.6 Detecting and modelling interaction p376 14.7 Causal interpretation of a multivariable linear model p377 14.8 Evaluating the least squares model p379 14.9 Evaluating the major assumptions p385 14.10 Assessment of individual observations p390 14.11 Time-series data p396 15. MODEL-BUILDING STRATEGIES 15.1 Introduction p402 15.2 Steps in building a model p403 15.3 Building a causal model p403 15.4 Reducing the number of predictors p404 15.5 The problem of missing values p408 15.6 Effects of continuous predictors p411 15.7 Identifying interaction terms of interest p418 15.8 Building the model p418 15.9 Evaluate the reliability of the model p423 15.10 Presenting the results p424 16. LOGISTIC REGRESSION 16.1 Introduction p430 16.2 The logistic model p430 16.3 Odds and odds ratios p431 16.4 Fitting a logistic regression model p432 16.5 Assumptions in logistic regression p433 16.6 Likelihood ratio statistics p434 16.7 Wald tests p436 16.8 Interpretation of coefficients p436 16.9 Assessing interaction and confounding p439 16.10 Model-building p441 16.11 Generalised linear models p444 16.12 Evaluating logistic regression models p445 16.13 Sample size considerations p455 16.14 Exact logistic regression p456 16.15 Conditional logistic regression for matched studies p456 17. MODELLING ORDINAL AND MULTINOMIAL DATA 17.1 Introduction p462 17.2 Overview of models p462 17.3 Multinomial logistic regression p466 17.4 Modelling ordinal data p470 17.5 Proportional odds model (constrained cumulative logit model) p471 17.6 Adjacent-category model p475 17.7 Continuation-ratio model p476 18. MODELLING COUNT AND RATE DATA 18.1 Introduction p480 18.2 The Poisson distribution p481 18.3 Poisson regression model p482 18.4 Interpretation of coefficients p483 18.5 Evaluating Poisson regression models p485 18.6 Negative binomial regression p488 18.7 Problems with zero counts p496 19. MODELLING SURVIVAL DATA 19.1 Introduction p502 19.2 Non-parametric analyses p507 19.3 Actuarial life tables p507 19.4 Kaplan-Meier estimate of survivor function p510 19.5 Nelson-Aalen estimate of cumulative hazard p512 19.6 Statistical inference in non-parametric analyses p512 19.7 Survivor, failure and hazard functions p514 19.8 Semi-parametric analyses p519 19.9 Parametric models p536 19.10 Accelerated failure time models p541 19.11 Frailty models and clustering p545 19.12 Multiple outcome event data p551 19.13 Discrete-time survival analysis p552 19.14 Sample sizes for survival analyses p557 20. INTRODUCTION TO CLUSTERED DATA 20.1 Introduction p564 20.2 Clustering arising from the data structure p564 20.3 Effects of clustering p570 20.4 Simulation studies on the impact of clustering p574 20.5 Introduction to methods for dealing with clustering p576 21. MIXED MODELS FOR CONTINUOUS DATA 21.1 Introduction p588 21.2 Linear mixed model p588 21.3 Random slopes p594 21.4 Contextual effects p598 21.5 Statistical analysis of linear mixed models p601 22. MIXED MODELS FOR DISCRETE DATA 22.1 Introduction p616 22.2 Logistic regression with random effects p617 22.3 Poisson regression with random effects p621 22.4 Generalised linear mixed model p623 22.5 Statistical analysis of GLMMs p630 22.6 Summary remarks on analysis of discrete clustered data p639 23. REPEATED MEASURES DATA 23.1 Introduction p646 23.2 Univariate and multivariate approaches to repeated measures data p648 23.3 Linear mixed models with correlation structure p654 23.4 Mixed models for discrete repeated measures data p662 23.5 Generalised estimating equations p665 24. INTRODUCTION TO BAYESIAN ANALYSIS 24.1 Introduction p676 24.2 Bayesian analysis p676 24.3 Markov chain Monte Carlo (MCMC) estimation p680 24.4 Statistical analysis based on MCMC estimation p685 24.5 Extensions of Bayesian and MCMC Modelling p689 25. ANALYSIS OF SPATIAL DATA: INTRODUCTION AND VISUALISATION 25.1 Introduction p702 25.2 Spatial data p702 25.3 Spatial data analysis p705 25.4 Additional topics p711 26. ANALYSIS OF SPATIAL DATA 26.1 Introduction p718 26.2 Issues specific to statistical analysis of spatial data p718 26.3 Exploratory spatial analysis p720 26.4 Global spatial clustering p728 26.5 Localised spatial cluster detection p735 26.6 Space-time association p738 26.7 Modelling p742 27. CONCEPTS OF INFECTIOUS DISEASE EPIDEMIOLOGY 27.1 Introduction p754 27.2 Infection vs disease p756 27.3 Transmission p758 27.4 Mathematical modelling of infectious disease transmission p760 27.5 Methods of control of infectious disease p763 27.6 Estimating R0 and other parameters p766 27.7 Developing more complex models p771 27.8 Using models p773 27.9 Summary p775 28. SYSTEMATIC REVIEWS AND META-ANALYSIS 28.1 Introduction p780 28.2 Narrative reviews p780 28.3 Systematic Reviews p781 28.4 Meta-analysis – Introduction p785 28.5 Fixed- and random-effects models p786 28.6 Presentation of results p789 28.7 Heterogeneity p791 28.8 Publication bias p798 28.9 Influential studies p801 28.10 Outcome scales and data issues p801 28.11 Meta-analysis of observational studies p804 28.12 Meta-analysis of diagnostic tests p806 28.13 Use of meta-analysis p807 29. ECOLOGICAL AND GROUP-LEVEL STUDIES 29.1 Introduction p814 29.2 Rationale for group level studies p815 29.3 Types of ecologic variable p816 29.4 Issues related to modelling approaches in ecologic studies p817 29.5 The linear model in the context of ecologic studies p818 29.6 Issues related to inferences p819 29.7 Sources of ecologic bias p820 29.8 Analysis of ecologic data p825 29.9 Non-ecologic group-level studies p826 30. A STRUCTURED APPROACH TO DATA ANALYSIS 30.1 Introduction p834 30.2 Data-collection sheets p834 30.3 Data coding p835 30.4 Data entry p835 30.5 Keeping track of files p836 30.6 Keeping track of variables p836 30.7 Program mode versus interactive processing p837 30.8 Data-editing p838 30.9 Data verification p839 30.10 Data processing—outcome variable(s) p839 30.11 Data processing—predictor variables p840 30.12 Data processing—multilevel data p840 30.13 Unconditional associations p841 30.14 Keeping track of your analyses p841 31. DESCRIPTION OF DATASETS p843