| |
| |
Preface to Second Edition | |
| |
| |
Mixed Model Notations | |
| |
| |
| |
Introduction | |
| |
| |
| |
The Use of Mixed Models | |
| |
| |
| |
Introductory Example | |
| |
| |
| |
Simple model to assess the effects of treatment (Model A) | |
| |
| |
| |
A model taking patient effects into account (Model B) | |
| |
| |
| |
Random effects model (Model C) | |
| |
| |
| |
Estimation (or prediction) of random effects | |
| |
| |
| |
A Multi-Centre Hypertension Trial | |
| |
| |
| |
Modelling the data | |
| |
| |
| |
Including a baseline covariate (Model B) | |
| |
| |
| |
Modelling centre effects (Model C) | |
| |
| |
| |
Including centre-by-treatment interaction effects (Model D) | |
| |
| |
| |
Modelling centre and centre-treatment effects as random (Model E) | |
| |
| |
| |
Repeated Measures Data | |
| |
| |
| |
Covariance pattern models | |
| |
| |
| |
Random coefficients models | |
| |
| |
| |
More about Mixed Models | |
| |
| |
| |
What is a mixed model | |
| |
| |
| |
Why use mixed models | |
| |
| |
| |
Communicating results | |
| |
| |
| |
Mixed models in medicine | |
| |
| |
| |
Mixed models in perspective | |
| |
| |
| |
Some Useful Definitions | |
| |
| |
| |
Containment | |
| |
| |
| |
Balance | |
| |
| |
| |
Error strata | |
| |
| |
| |
Normal Mixed Models | |
| |
| |
| |
Model Definition | |
| |
| |
| |
The fixed effects model | |
| |
| |
| |
The mixed model | |
| |
| |
| |
The random effects model covariance structure | |
| |
| |
| |
The random coefficients model covariance structure | |
| |
| |
| |
The covariance pattern model covariance structure | |
| |
| |
| |
Model Fitting Methods | |
| |
| |
| |
The likelihood function and approaches to its maximisation | |
| |
| |
| |
Estimation of fixed effects | |
| |
| |
| |
Estimation (or prediction) of random effects and coefficients | |
| |
| |
| |
Estimation of variance parameters | |
| |
| |
| |
The Bayesian Approach | |
| |
| |
| |
Introduction | |
| |
| |
| |
Determining the posterior density | |
| |
| |
| |
Parameter estimation, probability intervals and p-values | |
| |
| |
| |
Specifying non-informative prior distributions | |
| |
| |
| |
Evaluating the posterior distribution | |
| |
| |
| |
Practical Application and Interpretation | |
| |
| |
| |
Negative variance components | |
| |
| |
| |
Accuracy of variance parameters | |
| |
| |
| |
Bias in fixed and random effects standard errors | |
| |
| |
| |
Significance testing | |
| |
| |
| |
Confidence intervals | |
| |
| |
| |
Model checking | |
| |
| |
| |
Missing data | |
| |
| |
| |
Example | |
| |
| |
| |
Analysis models | |
| |
| |
| |
Results | |
| |
| |
| |
Discussion of points from Section 2.4 | |
| |
| |
| |
Generalised Linear Mixed Models | |
| |
| |
| |
Generalised Linear Models | |
| |
| |
| |
Introduction | |
| |
| |
| |
Distributions | |
| |
| |
| |
The general form for exponential distributions | |
| |
| |
| |
The GLM definition | |
| |
| |
| |
Fitting the GLM | |
| |
| |
| |
Expressing individual distributions in the general exponential form | |
| |
| |
| |
Conditional logistic regression | |
| |
| |
| |
Generalised Linear Mixed Models | |
| |
| |
| |
The GLMM definition | |
| |
| |
| |
The likelihood and quasi-likelihood functions | |
| |
| |
| |
Fitting the GLMM | |
| |
| |
| |
Practical Application and Interpretation | |
| |
| |
| |
Specifying binary data | |
| |
| |
| |
Uniform effects categories | |
| |
| |
| |
Negative variance components | |
| |
| |
| |
Fixed and random effects estimates | |
| |
| |
| |
Accuracy of variance parameters and random effects shrinkage | |
| |
| |
| |
Bias in fixed and random effects standard errors | |
| |
| |
| |
The dispersion parameter | |
| |
| |
| |
Significance testing | |
| |
| |
| |
Confidence intervals | |
| |
| |
| |
Model checking | |
| |
| |
| |
Example | |
| |
| |
| |
Introduction and models fitted | |
| |
| |
| |
Results | |
| |
| |
| |
Discussion of points from Section 3.3 | |
| |
| |
| |
Mixed Models for Categorical Data | |
| |
| |
| |
Ordinal Logistic Regression (Fixed Effects Model) | |
| |
| |
| |
Mixed Ordinal Logistic Regression | |
| |
| |
| |
Definition of the mixed ordinal logistic regression model | |
| |
| |
| |
Residual variance matrix | |
| |
| |
| |
Alternative specification for random effects models | |
| |
| |
| |
Likelihood and quasi-likelihood functions | |
| |
| |
| |
Model fitting methods | |
| |
| |
| |
Mixed Models for Unordered Categorical Data | |
| |
| |
| |
The G matrix | |
| |
| |
| |
The R matrix | |
| |
| |
| |
Fitting the model | |
| |
| |
| |
Practical Application and Interpretation | |
| |
| |
| |
Expressing fixed and random effects results | |
| |
| |
| |
The proportional odds assumption | |
| |
| |
| |
Number of covariance parameters | |
| |
| |
| |
Choosing a covariance pattern | |
| |
| |
| |
Interpreting covariance parameters | |
| |
| |
| |
Checking model assumptions | |
| |
| |
| |
The dispersion parameter | |
| |
| |
| |
Other points | |
| |
| |
| |
Example | |
| |
| |
| |
Multi-Centre Trials and Meta-Analyses | |
| |
| |
| |
Introduction to Multi-Centre Trials | |
| |
| |
| |
What is a multi-centre trial? | |
| |
| |
| |
Why use mixed models to analyse multi-centre data? | |
| |
| |
| |
The Implications of using Different Analysis Models | |
| |
| |
| |
Centre and centre-treatment effects fixed | |
| |
| |
| |
Centre effects fixed, centre-treatment effects omitted | |
| |
| |
| |
Centre and centre treatment effects random | |
| |
| |
| |
Centre effects random, centre-treatment effects omitted | |
| |
| |
| |
Example: A Multi-Centre Trial | |
| |
| |
| |
Practical Application and Interpretation | |
| |
| |
| |
Plausibility of a centre-treatment interaction | |
| |
| |
| |
Generalisation | |
| |
| |
| |
Number of centres | |
| |
| |
| |
Centre size | |
| |
| |
| |
Negative variance components | |
| |
| |
| |
Balance | |
| |
| |
| |
Sample Size Estimation | |
| |
| |
| |
Normal data | |
| |
| |
| |
Non-normal data | |
| |
| |
| |
Meta-Analysis | |
| |
| |
| |
Example: Meta-analysis | |
| |
| |
| |
Analyses | |
| |
| |
| |
Results | |
| |
| |
| |
Treatment estimates in individual trials | |
| |
| |
| |
Repeated Measures Data | |
| |
| |
| |
Introduction | |
| |
| |
| |
Reasons for repeated measurements | |
| |
| |
| |
Analysis objectives | |
| |
| |
| |
Fixed effects approaches | |
| |
| |
| |
Mixed models approaches | |
| |
| |
| |
Covariance Pattern Models | |
| |
| |
| |
Covariance patterns | |
| |
| |
| |
Choice of covariance pattern | |
| |
| |
| |
Choice of fixed effects | |
| |
| |
| |
General points | |
| |
| |
| |
Example: Covariance Pattern Models for Normal Data | |
| |
| |
| |
Analysis models | |
| |
| |
| |
Selection of covariance pattern | |
| |
| |
| |
Assessing fixed effects | |
| |
| |
| |
Model checking | |
| |
| |
| |
Example: Covariance Pattern Models for Count Data | |
| |
| |
| |
Analysis models | |
| |
| |
| |
Analysis using a categorical mixed model | |
| |
| |
| |
Random Coefficients Models | |
| |
| |
| |
Introduction | |
| |
| |
| |
General points | |
| |
| |
| |
Comparisons with fixed effects approaches | |
| |
| |
| |
Examples of Random Coefficients Models | |
| |
| |
| |
A linear random coefficients model | |
| |
| |
| |
A polynomial random coefficients model | |
| |
| |
| |
Sample Size Estimation | |
| |
| |
| |
Normal data | |
| |
| |
| |
Non-normal data | |
| |
| |
| |
Categorical data | |
| |
| |
| |
Cross-Over Trials | |
| |
| |
| |
Introduction | |
| |
| |
| |
Advantages of Mixed Models in Cross-Over Trials | |
| |
| |
| |
The AB/BA Cross-Over Trial | |
| |
| |
| |
Example: AB/BA cross-over design | |
| |
| |
| |
Higher Order Complete Block Designs | |
| |
| |
| |
Inclusion of carry-over effects | |
| |
| |
| |
Example: four-period, four-treatment cross-over trial | |
| |
| |
| |
Incomplete Block Designs | |
| |
| |
| |
The three-treatment, two-period design (Koch's design) | |
| |
| |
| |
Example: two-period cross-over trial | |
| |
| |
| |
Optimal Designs | |
| |
| |
| |
Example: Balaam's design | |
| |
| |
| |
Covariance Pattern Models | |
| |
| |
| |
Structured by period | |
| |
| |
| |
Structured by treatment | |
| |
| |
| |
Example: four-way cross-over trial | |
| |
| |
| |
Analysis of Binary Data | |
| |
| |
| |
Analysis of Categorical Data | |
| |
| |
| |
Use of Results from Random Effects Models in Trial Design | |
| |
| |
| |
Example | |
| |
| |
| |
General Points | |
| |
| |
| |
Other Applications of Mixed Models | |
| |
| |
| |
Trials with Repeated Measurements within Visits | |
| |
| |
| |
Covariance pattern models | |
| |
| |
| |
Example | |
| |
| |
| |
Random coefficients models | |
| |
| |
| |
Example: random coefficients models | |
| |
| |
| |
Multi-Centre Trials with Repeated Measurements | |
| |
| |
| |
Example: multi-centre hypertension trial | |
| |
| |
| |
Covariance pattern models | |
| |
| |
| |
Multi-Centre Cross-Over Trials | |
| |
| |
| |
Hierarchical Multi-Centre Trials and Meta-Analysis | |
| |
| |
| |
Matched Case-Control Studies | |
| |
| |
| |
Example | |
| |
| |
| |
Analysis of a quantitative variable | |
| |
| |
| |
Check of model assumptions | |
| |
| |
| |
Analysis of binary variables | |
| |
| |
| |
Different Variances for Treatment Groups in a Simple Between-Patient Trial | |
| |
| |
| |
Example | |
| |
| |
| |
Estimating Variance Components in an Animal Physiology Trial | |
| |
| |
| |
Sample size estimation for a future experiment | |
| |
| |
| |
Inter- and Intra-Observer Variation in Foetal Scan Measurements | |
| |
| |
| |
Components of Variation and Mean Estimates in a Cardiology Experiment | |
| |
| |
| |
Cluster Sample Surveys | |
| |
| |
| |
Example: cluster sample survey | |
| |
| |
| |
Small Area Mortality Estimates | |
| |
| |
| |
Estimating Surgeon Performance | |
| |
| |
| |
Event History Analysis | |
| |
| |
| |
Example | |
| |
| |
| |
A Laboratory Study Using a Within-Subject 4 x 4 Factorial Design | |
| |
| |
| |
Bioequivalence Studies with Replicate Cross-Over Designs | |
| |
| |
| |
Example | |
| |
| |
| |
Cluster Randomised Trials | |
| |
| |
| |
Example: a trial to evaluate integrated care pathways for treatment of children with asthma in hospital | |
| |
| |
| |
Example: Edinburgh randomised trial of breast screening | |
| |
| |
| |
Software for Fitting Mixed Models | |
| |
| |
| |
Packages for Fitting Mixed Models | |
| |
| |
| |
Basic use of PROC Mixed | |
| |
| |
| |
Using SAS to Fit Mixed Models to Non-Normal Data | |
| |
| |
| |
PROC GLIMMIX | |
| |
| |
| |
PROC GENMOD | |
| |
| |
Glossary | |
| |
| |
References | |
| |
| |
Index | |