Biostatistics for Epidemiology and Public Health Using R

Bertram K. C. (Bert) Chan, PhD, is currently Consulting Biostatistician at the School of Medicine, Department of Preventive Medicine at Loma Linda University, USA where he is currently involved with the 10-Year NIH funded project Adventist Health Study-2. He is also a Forum Lecturer at the Loma Linda University School of Public Health's Department of Biostatistics and Epidemiology on introducing the use of R programming to the curriculum. His extensive professional experience as either a consulting or full-time engineer includes Foundry Networks, Hewlett-Packard (PC and Network Servers Divisions), Apple Computer, Lockheed-Martin, and the Australian and Canadian Atomic Energy Commission Research Establishments. Dr. Chan has published over 30 peer-reviewed publications and technical reports and holds three U.S. Patents. He also published a 10-volume series of mathematics textbooks for secondary schools in Hong Kong.

Contents
Preface


1. INTRODUCTION


1.1 Medicine, Preventive Medicine, Public Health, and Epidemiology


Medicine


Preventive Medicine and Public Health


Public Health and Epidemiology


Review Questions for Section 1.1


1.2 Personal Health and Public Health


Personal Health Versus Public Health


Review Questions for Section 1.2


1.3 Research and Measurements in EPDM and PH


EPDM: The Basic Science of PH


Main Epidemiologic Functions


The Cause of Diseases


Exposure Measurement in Epidemiology


Additional Issues


Review Questions for Section 1.3


1.4 BIOS and EPDM


Review Questions for Section 1.4


References


2. RESEARCH AND DESIGN IN EPIDEMIOLOGY AND PUBLIC HEALTH


Introduction


2.1 Causation and Association in Epidemiology and Public Health


The Bradford-Hill Criteria for Causation and Association in Epidemiology


Legal Interpretation Using Epidemiology


Disease Occurrence


Review Questions for Section 2.1


2.2 Causation and Inference in Epidemiology and Public Health


Rothman’s Diagrams for Sufficient Causation of Diseases


Causal Inferences


Using the Causal Criteria


Judging Scientific Evidence


Review Questions for Section 2.2


2.3 Biostatistical Basis of Inference


Modes of Inference


Levels of Measurement


Frequentist BIOS in EPDM


Confidence Intervals in Epidemiology and Public Health


Bayesian Credible Interval


Review Questions for Section 2.3


2.4 BIOS in EPDM and PH


Applications of BIOS


BIOS in EPDM and PH


Processing and Analyzing Basic Epidemiologic Data


Analyzing Epidemiologic Data


Using R


Evaluating a Single Measure of Occurrence


Poisson Count (Incidence) and Rate Data


Binomial Risk and Prevalence Data


Evaluating Two Measures of Occurrence—Comparison of Risk: Risk Ratio and Attributable Risk


Comparing Two Rate Estimates: Rate Ratio rr


Comparing Two Risk Estimates: Risk Ratio RR and Disease (Morbidity) Odds Ratio DOR


Comparing Two Odds Estimates From Case–Control: The Salk Polio Vaccine Epidemiologic Study


Review Questions for Section 2.4


Exercises for Chapter 2


Using Probability Theory


Disease Symptoms in Clinical Drug Trials


Risks and Odds in Epidemiology


Case–Control Epidemiologic Study


Mortality, Morbidity, and Fertility Rates


Incidence Rates in Case-Cohort Survival Analysis


Prevalence


Mortality Rates


Estimating Sample Sizes


References


Appendix


3. DATA ANALYSIS USING R PROGRAMMING


Introduction


3.1 Data and Data Processing


Data Coding


Data Capture


Data Editing


Imputations


Data Quality


Producing Results


Review Questions for Section 3.1


3.2 Beginning R


R and Biostatistics


A First Session Using R


The R Environment


Review Questions for Section 3.2


3.3 R as a Calculator


Mathematical Operations Using R


Assignment of Values in R and Computations Using Vectors and Matrices


Computations in Vectors and Simple Graphics


Use of Factors in R Programming


Simple Graphics


x as Vectors and Matrices in Biostatistics


Some Special Functions That Create Vectors


Arrays and Matrices


Use of the Dimension Function dim in R


Use of the Matrix Function matrix in R


Some Useful Functions Operating on Matrices in R


NA: “Not Available” for Missing Values in Datasets


Special Functions That Create Vectors


Review Questions for Section 3.3


Exercises for Section 3.3


3.4 Using R in Data Analysis in BIOS


Entering Data at the R Command Prompt


The Function list() and the Making of data.frame() in R


Review Questions for Section 3.4


Exercises for Section 3.4


3.5 Univariate, Bivariate, and Multivariate Data Analysis


Univariate Data Analysis


Bivariate and Multivariate Data Analysis


Multivariate Data Analysis


Analysis of Variance (ANOVA)


Review Questions for Section 3.5


Exercises for Section 3.5


References


Appendix: Documentation for the plot function


Generic X–Y Plotting


4. GRAPHICS USING R


Introduction


Choice of System


Packages


4.1 Base (or Traditional) Graphics


High-Level Functions


Low-Level Plotting Functions


Interacting with Graphics


Using Graphics Parameters


Parameters List for Graphics


Device Drivers


Review Questions for Section 4.1


Exercises for Section 4.1


4.2 Grid Graphics


The lattice Package: Trellis Graphics


The Grid Model for R Graphics


Grid Graphics Objects


Applications to Biostatistical and Epidemiologic Investigations


Review Questions for Section 4.2


Exercises for Section 4.2


References


5. PROBABILITY AND STATISTICS IN BIOSTATISTICS


Introduction


5.1 Theories of Probability


What Is Probability?


Basic Properties of Probability


Probability Computations Using R


Applications of Probability Theory to Health Sciences


Typical Summary Statistics in Biostatistics: Confidence Intervals, Significance Tests, and Goodness of Fit


Review Questions for Section 5.1


Exercises for Section 5.1


5.2 Typical Statistical Inference in Biostatistics: Bayesian Biostatistics


What Is Bayesian Biostatistics?


Bayes’s Theorem in Probability Theory


Bayesian Methodology and Survival Analysis (Time-to-Event) Models for Biostatistics in Epidemiology and Preventive Medicine


The Inverse Bayes Formula


Modeling in Biostatistics


Review Questions for Section 5.2


Exercises for Section 5.2


References


6. CASE–CONTROL STUDIES AND COHORT STUDIES IN EPIDEMIOLOGY


Introduction


6.1 Theory and Analysis of Case–Control Studies


Advantages and Limitations of Case–Control Studies


Analysis of Case–Control Studies


Review Questions for Section 6.1


Exercises for Section 6.1


6.2 Theory and Analysis of Cohort Studies


An Important Application of Cohort Studies


Clinical Trials


Randomized Controlled Trials


Cohort Studies for Diseases of Choice and Noncommunicable Diseases


Cohort Studies and the Lexis Diagram in the Biostatistics of Demography


Review Questions for Section 6.2


Exercises for Section 6.2


References


7. RANDOMIZED TRIALS, PHASE DEVELOPMENT, CONFOUNDING IN SURVIVAL ANALYSIS, AND LOGISTIC REGRESSIONS


7.1 Randomized Trials


Classifications of RTs by Study Design


Randomization


Biostatistical Analysis of Data from RTs


Biostatistics for RTs in the R Environment


Review Questions for Section 7.1


Exercises for Section 7.1


7.2 Phase Development


Phase 0 or Preclinical Phase


Phase I


Phase II


Phase III


Pharmacoepidemiology: A Branch of Epidemiology


Some Basic Tests in Epidemiologic Phase Development


Review Questions for Section 7.2


Exercises for Section 7.2


7.3 Confounding in Survival Analysis


Biostatistical Approaches for Controlling Confounding


Using Regression Modeling for Controlling Confounding


Confounding and Collinearity


Review Questions for Section 7.3


Exercises for Section 7.3


7.4 Logistic Regressions


Inappropriateness of the Simple Linear Regression When y Is a Categorical Dependent Variable


The Logistic Regression Model


The Logit


Logistic Regression Analysis


Generalized Linear Models in R


Review Questions for Section 7.4


Exercises for Section 7.4


References


Index