Epidemiology with R /
Carstensen, Bendix,
Epidemiology with R / Bendix Carstensen. - First edition. - UK : OUP Oxford, 2021. - xiv, 231 pages : illustrations ; 25 cm
Includes bibliographical references and index.
Epidemiology with R
Copyright
Contents
Preface
What this book is not
Acknowledgements
List of Figures
Introduction
What you should do
Code chunks
Graphs in this book
Practicing R
Chapter 1: Using R
1.1 Installing and using R
1.2 Documenting your code and results
1.2.1 R markdown
1.2.2 Sweave / knitr
1.2.3 Coding style in R
1.2.4 R lingo
1.3 Simple usage of R
1.3.1 Using R as a calculator
1.3.2 A functional language
Probability functions
Objects and functions
What makes R different: functions
1.3.3 Sequences
1.3.4 The births data
1.3.5 Referencing parts of a data frame
1.3.6 Summaries
1.3.7 Generating new variables
1.3.8 Logical variables
1.3.9 Turning a variable into a factor
Manipulating factor levels
Grouping values of a quantitative variable
1.3.10 Tables
Tables of means and other things
1.3.11 Reading data
1.3.12 Saving data
Saving the work space
Saving R objects in a file
1.3.13 The search path
Attaching a data frame
Using with
1.4 Graphics
1.4.1 ggplot2
1.4.2 Base graphics
1.4.3 Simple base graphs
Plot on the screen
Colours
Adding to a plot
Using indexing for plot elements
Interacting with a plot
Saving graphs for use in other documents
Same graph on multiple devices
The par command
1.5 Frequency data
1.5.1 Graphical overview
1.5.2 Ad hoc analyses of admissions
1.6 Tables and arrays for results
1.7 Dates in R
1.8 Numerical accuracy
1.8.1 Accuracy of matching variables
1.9 tidyverse and data.table
Chapter 2: Measures of disease occurrence
2.1 Prevalence
2.2 Mortality rate
2.3 Incidence rate
2.4 Standardized mortality ratio
2.5 Survival
2.5.1 Cumulative risk
2.5.2 Competing risks
2.5.3 Sojourn time
Chapter 3: Prevalence data—models, likelihood, and binomial regression
3.1 Likelihood
3.1.1 A single probability
3.1.2 Simple confidence interval
3.1.3 Confidence intervals in general
3.1.4 The normal distribution
3.1.5 Simple confidence intervals from models
3.1.6 Tests and p-values
3.2 Prevalence by age
3.3 Comparing different models for the same data
3.3.1 Likelihood-ratio test
3.3.2 Deviance
3.3.3 Deviance and goodness of fit
3.3.4 AIC and BIC
Chapter 4: Regression models
4.1 Types of models
4.2 Normal linear regression model
4.3 Simple linear regression
4.4 Multiple regression
4.4.1 Estimation in the normal linear regression model
4.4.2 R-squared
4.4.3 Multiple regression
4.4.4 Standardized variables
4.4.5 Predictions from the normal regression model
4.5 Model formulae in R
4.6 Regression models and generalized linear models
4.6.1 Categorical effects
4.6.2 Linear and categorical effects
4.6.3 ANOVA–ANCOVA
4.6.4 Categorical-linear interaction
Special interaction?
4.6.5 Categorical by categorical interaction
4.7 Collinearity and aliasing
4.8 Logarithmic transformations
4.8.1 Logarithms
4.8.2 Log transform of the response variable
4.8.3 Coefficient of variation
4.8.4 Log transform of an explanatory variable
4.8.5 Log transform of both the response and explanatory variables
Chapter 5: Analysis of follow-up data
5.1 Basic data structure
5.2 Probability model
5.2.1 Data
5.2.2 Likelihood for a rate
5.2.3 Estimates of rates and rate ratios
5.3 Representation of follow-up data
5.3.1 Lexis object for follow-up data
Scaling of Lexis diagrams
5.4 Splitting the follow-up time along a time-scale
5.5 Smooth age-effects for rates
5.5.1 Disaggregated data
5.5.2 Including sex in the model
5.6 SMR
5.6.1 Modelling the SMR
5.7 Time-dependent variables
5.7.1 Cutting time at a specific date
The precursor states
5.7.2 Modelling time-dependent variables
Survival?
5.7.3 Clinical measurements in cohort studies
Analysis using clinical measurements
Chapter 6: Parametrization and prediction of rates
6.1 Predictions and contrasts
6.2 Prediction of a single rate
6.3 Categorical variables
6.3.1 Groups and rate ratios
Comparing all groups
6.4 Modelling the effect of quantitative variables
6.4.1 Categorizing quantitative variables: don’t
6.4.2 Linear effect
Predicting the rates
6.4.3 Polynomial effects
6.4.4 Other types of non-linear effects
Natural splines
Penalized splines
6.5 Two quantitative predictors
6.5.1 Age and period
6.5.2 Age and cohort
6.5.3 Contours of joint effects
Image plot / heatmap
6.6 Quantitative interactions
6.6.1 Age–period interaction
Age-specific rates at different dates (periods)
Period-specific rates at different ages
6.6.2 Age and cohort interaction
6.6.3 Parametric interaction models
6.6.4 Varying coefficients models for interaction
6.6.5 Summary of quantitative interactions
Chapter 7: Case-control and case-cohort studies
7.1 Follow-up and case-control studies
7.1.1 Probabilities and odds in case-control studies
7.1.2 The sampling fractions
7.1.3 A simple example
7.2 Statistical model for the odds ratio
7.2.1 Analysis by logistic regression
7.3 Odds ratio and rate ratio
7.3.1 Incidence density sampling
7.4 Confounding and stratified sampling
7.4.1 Stratified sampling
7.5 Individually matched studies
7.5.1 An example
7.5.2 When conditional analysis is not needed
7.6 Nested case-control studies
7.6.1 Register-based case-control studies
7.7 Case-cohort studies
Chapter 8: Survival analysis
8.1 Introduction
8.2 Life table estimator of survival function
8.3 Kaplan--Meier estimator of survival
8.3.1 Survival in two groups
8.4 The Cox model
8.4.1 Mean survival or survival at mean?
8.5 The time-scale
8.6 Relation between Cox and Poisson models
8.6.1 Simple parametric mortality functions
Baseline mortality rate
Survival curves
8.6.2 Proportional hazards?
8.6.3 The Cox model as a Poisson model
8.7 Time-dependent covariates
8.8 Competing risks
8.9 Modelling cause specific rates
8.9.1 Limitations
8.10 The Fine--Gray approach to competing risks
8.11 Time-dependent variables and competing risks
Chapter 9: Do not group quantitative variables
9.1 Problems Caused by Categorizing Continuous Variables
References
Index
Epidemiology with R
This practical guide is designed for students and researchers with an existing knowledge of R who wish to learn how to apply it in an epidemiological context and exploit its versatility. It also serves as a broader introduction to the quantitative aspects of modern practical epidemiology. The standard tools used in epidemiology are described and the practical use of R for these is clearly explained and laid out. R code examples, many with output, are embedded throughout the text. The entire code is also available on the companion website so that readers can reproduce all the results and graphs featured in the book. Epidemiology with R is an advanced textbook suitable for senior undergraduate and graduate students, professional researchers, and practitioners in the fields of human and non-human epidemiology, public health, veterinary science, and biostatistics.
0198841337 9780198841333 9780198841326 0198841329
Epidemiology
R (Computer program language)
Epidemiology
Epidemiology
R (Computer program language)--Statistical methods.--Statistics.--Statistical methods.
614.407 / CAR
WA 950 / .C321e 2021
Epidemiology with R / Bendix Carstensen. - First edition. - UK : OUP Oxford, 2021. - xiv, 231 pages : illustrations ; 25 cm
Includes bibliographical references and index.
Epidemiology with R
Copyright
Contents
Preface
What this book is not
Acknowledgements
List of Figures
Introduction
What you should do
Code chunks
Graphs in this book
Practicing R
Chapter 1: Using R
1.1 Installing and using R
1.2 Documenting your code and results
1.2.1 R markdown
1.2.2 Sweave / knitr
1.2.3 Coding style in R
1.2.4 R lingo
1.3 Simple usage of R
1.3.1 Using R as a calculator
1.3.2 A functional language
Probability functions
Objects and functions
What makes R different: functions
1.3.3 Sequences
1.3.4 The births data
1.3.5 Referencing parts of a data frame
1.3.6 Summaries
1.3.7 Generating new variables
1.3.8 Logical variables
1.3.9 Turning a variable into a factor
Manipulating factor levels
Grouping values of a quantitative variable
1.3.10 Tables
Tables of means and other things
1.3.11 Reading data
1.3.12 Saving data
Saving the work space
Saving R objects in a file
1.3.13 The search path
Attaching a data frame
Using with
1.4 Graphics
1.4.1 ggplot2
1.4.2 Base graphics
1.4.3 Simple base graphs
Plot on the screen
Colours
Adding to a plot
Using indexing for plot elements
Interacting with a plot
Saving graphs for use in other documents
Same graph on multiple devices
The par command
1.5 Frequency data
1.5.1 Graphical overview
1.5.2 Ad hoc analyses of admissions
1.6 Tables and arrays for results
1.7 Dates in R
1.8 Numerical accuracy
1.8.1 Accuracy of matching variables
1.9 tidyverse and data.table
Chapter 2: Measures of disease occurrence
2.1 Prevalence
2.2 Mortality rate
2.3 Incidence rate
2.4 Standardized mortality ratio
2.5 Survival
2.5.1 Cumulative risk
2.5.2 Competing risks
2.5.3 Sojourn time
Chapter 3: Prevalence data—models, likelihood, and binomial regression
3.1 Likelihood
3.1.1 A single probability
3.1.2 Simple confidence interval
3.1.3 Confidence intervals in general
3.1.4 The normal distribution
3.1.5 Simple confidence intervals from models
3.1.6 Tests and p-values
3.2 Prevalence by age
3.3 Comparing different models for the same data
3.3.1 Likelihood-ratio test
3.3.2 Deviance
3.3.3 Deviance and goodness of fit
3.3.4 AIC and BIC
Chapter 4: Regression models
4.1 Types of models
4.2 Normal linear regression model
4.3 Simple linear regression
4.4 Multiple regression
4.4.1 Estimation in the normal linear regression model
4.4.2 R-squared
4.4.3 Multiple regression
4.4.4 Standardized variables
4.4.5 Predictions from the normal regression model
4.5 Model formulae in R
4.6 Regression models and generalized linear models
4.6.1 Categorical effects
4.6.2 Linear and categorical effects
4.6.3 ANOVA–ANCOVA
4.6.4 Categorical-linear interaction
Special interaction?
4.6.5 Categorical by categorical interaction
4.7 Collinearity and aliasing
4.8 Logarithmic transformations
4.8.1 Logarithms
4.8.2 Log transform of the response variable
4.8.3 Coefficient of variation
4.8.4 Log transform of an explanatory variable
4.8.5 Log transform of both the response and explanatory variables
Chapter 5: Analysis of follow-up data
5.1 Basic data structure
5.2 Probability model
5.2.1 Data
5.2.2 Likelihood for a rate
5.2.3 Estimates of rates and rate ratios
5.3 Representation of follow-up data
5.3.1 Lexis object for follow-up data
Scaling of Lexis diagrams
5.4 Splitting the follow-up time along a time-scale
5.5 Smooth age-effects for rates
5.5.1 Disaggregated data
5.5.2 Including sex in the model
5.6 SMR
5.6.1 Modelling the SMR
5.7 Time-dependent variables
5.7.1 Cutting time at a specific date
The precursor states
5.7.2 Modelling time-dependent variables
Survival?
5.7.3 Clinical measurements in cohort studies
Analysis using clinical measurements
Chapter 6: Parametrization and prediction of rates
6.1 Predictions and contrasts
6.2 Prediction of a single rate
6.3 Categorical variables
6.3.1 Groups and rate ratios
Comparing all groups
6.4 Modelling the effect of quantitative variables
6.4.1 Categorizing quantitative variables: don’t
6.4.2 Linear effect
Predicting the rates
6.4.3 Polynomial effects
6.4.4 Other types of non-linear effects
Natural splines
Penalized splines
6.5 Two quantitative predictors
6.5.1 Age and period
6.5.2 Age and cohort
6.5.3 Contours of joint effects
Image plot / heatmap
6.6 Quantitative interactions
6.6.1 Age–period interaction
Age-specific rates at different dates (periods)
Period-specific rates at different ages
6.6.2 Age and cohort interaction
6.6.3 Parametric interaction models
6.6.4 Varying coefficients models for interaction
6.6.5 Summary of quantitative interactions
Chapter 7: Case-control and case-cohort studies
7.1 Follow-up and case-control studies
7.1.1 Probabilities and odds in case-control studies
7.1.2 The sampling fractions
7.1.3 A simple example
7.2 Statistical model for the odds ratio
7.2.1 Analysis by logistic regression
7.3 Odds ratio and rate ratio
7.3.1 Incidence density sampling
7.4 Confounding and stratified sampling
7.4.1 Stratified sampling
7.5 Individually matched studies
7.5.1 An example
7.5.2 When conditional analysis is not needed
7.6 Nested case-control studies
7.6.1 Register-based case-control studies
7.7 Case-cohort studies
Chapter 8: Survival analysis
8.1 Introduction
8.2 Life table estimator of survival function
8.3 Kaplan--Meier estimator of survival
8.3.1 Survival in two groups
8.4 The Cox model
8.4.1 Mean survival or survival at mean?
8.5 The time-scale
8.6 Relation between Cox and Poisson models
8.6.1 Simple parametric mortality functions
Baseline mortality rate
Survival curves
8.6.2 Proportional hazards?
8.6.3 The Cox model as a Poisson model
8.7 Time-dependent covariates
8.8 Competing risks
8.9 Modelling cause specific rates
8.9.1 Limitations
8.10 The Fine--Gray approach to competing risks
8.11 Time-dependent variables and competing risks
Chapter 9: Do not group quantitative variables
9.1 Problems Caused by Categorizing Continuous Variables
References
Index
Epidemiology with R
This practical guide is designed for students and researchers with an existing knowledge of R who wish to learn how to apply it in an epidemiological context and exploit its versatility. It also serves as a broader introduction to the quantitative aspects of modern practical epidemiology. The standard tools used in epidemiology are described and the practical use of R for these is clearly explained and laid out. R code examples, many with output, are embedded throughout the text. The entire code is also available on the companion website so that readers can reproduce all the results and graphs featured in the book. Epidemiology with R is an advanced textbook suitable for senior undergraduate and graduate students, professional researchers, and practitioners in the fields of human and non-human epidemiology, public health, veterinary science, and biostatistics.
0198841337 9780198841333 9780198841326 0198841329
Epidemiology
R (Computer program language)
Epidemiology
Epidemiology
R (Computer program language)--Statistical methods.--Statistics.--Statistical methods.
614.407 / CAR
WA 950 / .C321e 2021