Data Analysis in Medicine and Health using R

Description

In medicine and health, data are analyzed to guide treatment plans, patient care and control and prevention policies. However, in doing so, researchers in medicine and health often lack the understanding of data and statistical concepts and the skills in programming. In addition, there is also an increasing demand for data analyses to be reproducible, along with more complex data that require cutting-edge analysis. This book provides readers with both the fundamental concepts of data and statistical analysis and modeling. It also has the skills to perform the analysis using the R programming language, which is the lingua franca for statisticians. The topics in the book are presented in a sequence to minimize the time to help readers understand the objectives of data and statistical analysis, learn the concepts of statistical modeling and acquire the skills to perform the analysis. The R codes and datasets used in the book will be made available on GitHub for easy access. The book will also be live on the website bookdown.org, a service provided by RStudio, PBC, to host books written using the bookdown package in the R programming language. Cover Half Title Series Page Title Page Copyright Page Dedication Contents Preface 1. R, RStudio and RStudio Cloud 1.1. Objectives 1.2. Introduction 1.3. RStudio IDE 1.4. RStudio Cloud 1.4.1. The RStudio Cloud registration 1.4.2. Register and log in 1.5. Point and Click Graphical User Interface (GUI) 1.6. RStudio Server 1.7. Installing R and RStudio on Your Local Machine 1.7.1. Installing R 1.7.2. Installing RStudio IDE 1.7.3. Checking R and RStudio installations 1.7.4. TinyTeX, MiKTeX or MacTeX (for Mac OS) and TeX live 1.8. Starting Your RStudio 1.8.1. Console tab 1.8.2. Files, plots, packages, help and viewer pane 1.8.3. Environment, history, connection and build pane 1.8.4. Source pane 1.9. Summary 2. R Scripts and R Packages 2.1. Objectives 2.2. Introduction 2.3. Open a New Script 2.3.1. Our first R script 2.3.2. Function, argument and parameters 2.3.3. If users require further help 2.4. Packages 2.4.1. Packages on CRAN 2.4.2. Checking availability of R package 2.4.3. Install an R package 2.5. Working Directory 2.5.1. Starting a new R job 2.5.2. Creating a new R project 2.5.3. Location for dataset 2.6. Upload Data to RStudio Cloud 2.7. More Resources on RStudio Cloud 2.8. Guidance and Help 2.9. Bookdown 2.10. Summary 3. RStudio Project 3.1. Objectives 3.2. Introduction 3.3. Dataset Repository on GitHub 3.4. RStudio Project on RStudio or Posit Cloud 3.5. RStudio Project on Local Machine 3.6. Summary 4. Data Visualization 4.1. Objectives 4.2. Introduction 4.3. History and Objectives of Data Visualization 4.4. Ingredients for Good Graphics 4.5. Graphics Packages in R 4.6. The ggplot2 Package 4.7. Preparation 4.7.1. Create a new RStudio project 4.7.2. Important questions before plotting graphs 4.8. Read Data 4.9. Load the Packages 4.10. Read the Dataset 4.11. Basic Plots 4.12. More Complex Plots 4.12.1. Adding another variable 4.12.2. Making subplots 4.12.3. Overlaying plots 4.12.4. Combining different plots 4.12.5. Statistical transformation 4.12.6. Customizing title 4.12.7. Choosing themes 4.12.8. Adjusting axes 4.13. Saving Plots 4.14. Summary 5. Data Wrangling 5.1. Objectives 5.2. Introduction 5.2.1. Definition of data wrangling 5.3. Data Wrangling with dplyr Package 5.3.1. dplyr package 5.3.2. Common data wrangling processes 5.3.3. Some dplyr functions 5.4. Preparation 5.4.1. Create a new project or set the working directory 5.4.2. Load the libraries 5.4.3. Datasets 5.5. Select Variables, Generate New Variable and Rename Variable 5.5.1. Select variables using dplyr::select() 5.5.2. Generate new variable using mutate() 5.5.3. Rename variable using rename() 5.6. Sorting Data and Selecting Observation 5.6.1. Sorting data using arrange() 5.6.2. Select observation using filter() 5.7. Group Data and Get Summary Statistics 5.7.1. Group data using group_by() 5.7.2. Summary statistic using summarize() 5.8. More Complicated dplyr Verbs 5.9. Data Transformation for Categorical Variables 5.9.1. forcats package 5.9.2. Conversion from numeric to factor variables 5.9.3. Recoding variables 5.9.4. Changing the level of categorical variable 5.10. Additional Resources 5.11. Summary 6. Exploratory Data Analysis 6.1. Objectives 6.2. Introduction 6.3. EDA Using ggplot2 Package 6.3.1. Usage of ggplot2 6.4. Preparation 6.4.1. Load the libraries 6.4.2. Read the dataset into R 6.5. EDA in Tables 6.6. EDA with Plots 6.6.1. One variable: Distribution of a categorical variable 6.6.2. One variable: Distribution of a numerical variable 6.6.3. Two variables: Plotting a numerical and a categorical variable 6.6.4. Three variables: Plotting a numerical and two categorical variables 6.6.5. Faceting the plots 6.6.6. Line plot 6.6.7. Plotting means and error bars 6.6.8. Scatterplot with fit line 6.7. Summary 7. Linear Regression 7.1. Objectives 7.2. Introduction 7.3. Linear Regression Models 7.4. Prepare R Environment for Analysis 7.4.1. Libraries 7.4.2. Dataset 7.5. Simple Linear Regression 7.5.1. About simple linear regression 7.5.2. Data exploration 7.5.3. Univariable analysis 7.5.4. Model fit assessment 7.5.5. Presentation and interpretation 7.6. Multiple Linear Regression 7.6.1. About multiple linear regression 7.6.2. Data exploration 7.6.3. Univariable analysis 7.6.4. Multivariable analysis 7.6.5. Interaction 7.6.6. Model fit assessment 7.6.7. Presentation and interpretation 7.7. Prediction 7.8. Summary 8. Binary Logistic Regression 8.1. Objectives 8.2. Introduction 8.3. Logistic Regression Model 8.4. Dataset 8.5. Logit and Logistic Models 8.6. Prepare Environment for Analysis 8.6.1. Creating a RStudio project 8.6.2. Loading libraries 8.7. Read Data 8.8. Explore Data 8.9. Estimate the Regression Parameters 8.10. Simple Binary Logistic Regression 8.11. Multiple Binary Logistic Regression 8.12. Convert the Log Odds to Odds Ratio 8.13. Making Inference 8.14. Models Comparison 8.15. Adding an Interaction Term 8.16. Prediction from Binary Logistic Regression 8.16.1. Predict the log odds 8.16.2. Predict the probabilities 8.17. Model Fitness 8.18. Presentation of Logistic Regression Model 8.19. Summary 9. Multinomial Logistic Regression 9.1. Objectives 9.2. Introduction 9.3. Examples of Multinomial Outcome Variables 9.4. Models for Multinomial Outcome Data 9.5. Estimation for Multinomial Logit Model 9.5.1. Log odds and odds ratios 9.5.2. Conditional probabilities 9.6. Prepare Environment 9.6.1. Load libraries 9.6.2. Dataset 9.6.3. Read data 9.6.4. Data wrangling 9.6.5. Create new categorical variable from fbs 9.6.6. Exploratory data analysis 9.6.7. Confirm the order of cat_fbs 9.7. Estimation 9.7.1. Single independent variable 9.7.2. Multiple independent variables 9.7.3. Model with interaction term between independent variables 9.8. Inferences 9.9. Interpretation 9.10. Prediction 9.11. Presentation of Multinomial Regression Model 9.12. Summary 10. Poisson Regression 10.1. Objectives 10.2. Introduction 10.3. Prepare R Environment for Analysis 10.3.1. Libraries 10.4. Poisson Regression for Count 10.4.1. About Poisson regression for count 10.4.2. Dataset 10.4.3. Data exploration 10.4.4. Univariable analysis 10.4.5. Multivariable analysis 10.4.6. Interaction 10.4.7. Model fit assessment 10.4.8. Presentation and interpretation 10.4.9. Prediction 10.5. Poisson Regression for Rate 10.5.1. About Poisson regression for rate 10.5.2. Dataset 10.5.3. Data exploration 10.5.4. Univariable analysis 10.5.5. Multivariable analysis 10.5.6. Interaction 10.5.7. Model fit assessment 10.5.8. Presentation and interpretation 10.6. Quasi-Poisson Regression for Overdispersed Data 10.7. Summary 11. Survival Analysis: Kaplan–Meier and Cox Proportional Hazard (PH) Regression 11.1. Objectives 11.2. Introduction 11.3. Types of Survival Analysis 11.4. Prepare Environment for Analysis 11.4.1. RStudio project 11.4.2. Packages 11.5. Data 11.6. Explore Data 11.7. Kaplan–Meier Survival Estimates 11.8. Plot the Survival Probability 11.9. Comparing Kaplan–Meier Estimates across Groups 11.9.1. Log-rank test 11.9.2. Peto-peto test 11.10. Semi-Parametric Models in Survival Analysis 11.10.1. Cox proportional hazards regression 11.10.2. Advantages of the Cox proportional hazards regression 11.11. Estimation from Cox Proportional Hazards Regression 11.11.1. Simple Cox PH regression 11.11.2. Multiple Cox PH regression 11.12. Adding Interaction in the Model 11.13. The Proportional Hazard Assumption 11.13.1. Risk constant over time 11.13.2. Test for PH assumption 11.13.3. Plots to assess PH assumption 11.14. Model Checking 11.14.1. Prediction from Cox PH model 11.14.2. Residuals from Cox PH model 11.14.3. Influential observations 11.15. Plot the Adjusted Survival 11.16. Presentation and Interpretation 11.17. Summary 12. Parametric Survival Analysis 12.1. Objectives 12.2. Introduction 12.2.1. Advantages of parametric survival analysis models 12.3. Parametric Survival Analysis Model 12.3.1. Proportional hazard parametric models 12.3.2. Accelerated failure time model (AFT) models 12.4. Analysis 12.4.1. Dataset 12.4.2. Set the environment 12.4.3. Read dataset 12.4.4. Data wrangling 12.4.5. Exploratory data analysis (EDA) 12.4.6. Exponential survival model 12.4.7. Weibull (accelerated failure time) 12.4.8. Weibull (proportional hazard) 12.4.9. Model adequacy for Weibull distribution 12.5. Summary 13. Introduction to Missing Data Analysis 13.1. Objectives 13.2. Introduction 13.3. Types of Missing Data 13.4. Preliminaries 13.4.1. Packages 13.4.2. Dataset 13.5. Exploring Missing Data 13.6. Handling Missing Data 13.6.1. Listwise deletion 13.6.2. Simple imputation 13.6.3. Single imputation 13.6.4. Multiple imputation 13.7. Presentation 13.8. Resources 13.9. Summary 14. Model Building and Variable Selection 14.1. Objectives 14.2. Introduction 14.3. Model Building 14.4. Variable Selection for Prediction 14.4.1. Backward elimination 14.4.2. Forward selection 14.4.3. Stepwise selection 14.4.4. All possible subset selection 14.5. Stopping Rule and Selection Criteria in Automatic Variable Selection 14.6. Problems with Automatic Variable Selections 14.7. Purposeful Variable Selection 14.8. Summary Bibliography Index