Modern Statistical Methods for Health Research (Emerging Topics in Statistics and Biostatistics)

Preface
	Part I: Health Data Analysis and Applications to EHR Data (Chaps. 1 –5)
	Part II: Clinical Trials, FDR, and Applications in Health Science (Chaps. 6 –10)
	Part III: Big Data Analytics and Its Applications (Chaps. 11 –15)
	Part IV: Survival Analysis and Functional Data Analysis (Chaps. 16 –18)
	Part V: Statistical Modeling in Genomic Studies (Chaps. 19 –21)
Contents
About the Editors
List of Contributors
List of Chapter Reviewers
Part I Health Data Analysis and Applications to EHR Data
	The Effective Sample Size of EHR-Derived Cohorts Under Biased Sampling
		1 Introduction
		2 Methods
			2.1 Notation and Definitions
			2.2 Bias and Mean-Squared Error of the Simple Random Sample and the EHR-Based Sample
			2.3 Effective Sample Size of the EHR-Derived Cohort
			2.4 Simulation Study Design
		3 Results
		4 Discussion
		References
	Non-Gaussian Models for Object Motion Analysis with Time-Lapse Fluorescence Microscopy Images
		1 Introduction
		2 Method
			2.1 Particle Tracking Framework
			2.2 Object Segmentation
			2.3 Observation and Dynamics Models
				Ellipsoid Model
				Voxel-Based Model
			2.4 Multiple Object Tracking Management
		3 Experiments and Results
			3.1 Validation with Artificial Data
			3.2 Bacteria Motility Analysis
			3.3 Tumor Spheroid Study
		4 Conclusions
		References
	Alternative Capture-Recapture Point and Interval Estimators Based on Two Surveillance Streams
		1 Introduction
		2 Methods
			2.1 The LP Conditions and Their Central Role
			2.2 Some Cautionary Notes on Alternatives to the LP Estimator
			2.3 Loglinear Models and a Perspective on the Use of Covariates
			2.4 A Rationale for Renewed Statistical Interest Under the LP Conditions
			2.5 Review of Classical Point Estimators in the Two-Capture Case
			2.6 A Class of Estimators Including  LP  and  Chap  as Special Cases
			2.7 A New Estimator Targeting Median Bias as a Criterion
			2.8 New Alternatives to the Chapman Estimator Aimed at Reduced Mean Bias
			2.9 Closed-Form Confidence Interval Estimation in the Two-Capture Case
			2.10 An Adapted Bayesian Credible Interval Approach
		3 Motivating Example Data and Results
		4 Simulation Studies
		5 Discussion
		References
	A Uniform Shrinkage Prior in Spatiotemporal Poisson Models for Count Data
		1 Introduction
		2 Derivation of a USP for the Variance Components in GLMM with Proper CAR and Its Properties
			2.1 Derivation of the USP
			2.2 Motivation of the Derived USP
			2.3 Analytical Properties of the Derived USP
		3 Application to the Leptospirosis Data
		4 Simulation Study
		5 Discussion
		References
	A Review of Multiply Robust Estimation with Missing Data
		1 Introduction
		2 Basic Setups
		3 Multiply Robust Estimation Procedure
			3.1 Calibration Approach
			3.2 Projection Approach
			3.3 Multiple Imputation Approach
		4 Simulation Study
		5 Real Application
		6 Discussion
		References
Part II Clinical Trials, FDR, and Applications in Health Science
	Approaches to Combining Phase II Proof-of-Conceptand Dose-Finding Trials
		1 Introduction
		2 Two Studies—Phase IIa (Proof-of-Concept) and Phase IIb (Dose-Finding)
		3 A Single Study with Combined Objectives (PoC and DF)
			3.1 Single Fixed Design
			3.2 Two-Stage Phase IIa/IIb Adaptive Designs
		4 Sample Size Comparison and Discussion
		5 Concluding Remarks
		References
	Designs of Early Phase Cancer Trials with Drug Combinations
		1 Introduction
		2 Designs for Phase I Clinical Trials
			2.1 Phase I Model-based Designs for Drug Combinations
				Model
				Prior and Posterior Distributions
				Trial Design
				Design Operating Characteristics
				Results
			2.2 Attributing Dose-Limiting Toxicities
				Model
				Trial Design
				Results
			2.3 Adding a Baseline Covariate
				Model
				Prior and Posterior Distributions
				Trial Design
				Results
		3 Designs for Phase I–II Clinical Trials
			3.1 Binary Endpoint
				Model
				Trial Design
				Results
			3.2 Survival Endpoint
				Introduction
				Model
				Trial Design
				Results
		4 Discussion
		References
	Controlling the False Discovery Rate of Grouped Hypotheses
		1 Introduction
		2 Modeling and Sequential Framework
			2.1 Notation and Models
			2.2 A General Framework for Grouped Multiple Testing Procedures
		3 Procedures for Group Multiple Testing
			3.1 Conditional Local FDR (CLfdr)
			3.2 Group-Weighted Benjamini–Hochberg (GBH)
			3.3 Weighting Fixed Cutoff (WFC)
			3.4 Structure-Adaptive Benjamini–Hochberg (SABHA)
			3.5 Independent Hypothesis Weighting (IHWc)
			3.6 Adaptive p-Value Thresholding (AdaPT)
			3.7 Linear and Nonlinear Rankings
		4 Simulation
			4.1 Results
		5 Application
		6 Conclusions and Discussions
		Appendix
			A.1 Two-Parameter AdaPT
			A.2 EM Steps
			A.3 Initialization
		References
	Classic Linear Mediation Analysis of Complex Survey Data Using Balanced Repeated Replication
		1 Introduction
		2 Technical Details
			2.1 Mediation Model
			2.2 Complex Surveys Using BRR
			2.3 Mediation Incorporating BRR
				Point Estimate
				Standard Error Estimate
				Significance Test
		3 SAS Macro and Illustration
			3.1 Components of %MediationBRR
			3.2 Application to PISA: A Single-Mediator Model
			3.3 Application to TUS-CPS: A Multi-Mediator Model
		4 Discussion
		Appendix
		References
	A Review of Bayesian Optimal Experimental Design on DifferentModels
		1 Introduction
			1.1 Pseudo-Bayesian Optimal Design
			1.2 Fully Bayesian Optimal Design
		2 Bayesian Designs for Linear Models
		3 Bayesian Designs for Generalized Linear Models
		4 Bayesian Designs for Nonlinear Models
			4.1 Bayesian Optimal Designs for PKPD Models
			4.2 Bayesian Optimal Designs for Biological and Chemical Models
		5 Conclusions
		References
Part III Big Data Analytics and Its Applications
	A Selective Review on Statistical Techniques for Big Data
		1 Introduction
		2 Randomized Numerical Linear Algebra
			2.1 Random Projection
			2.2 Nonuniform Random Sampling
		3 Information-Based Optimal Subdata Selection
		4 Informative Subsampling
			4.1 Optimal Subsampling
			4.2 Local Case–Control Subsampling
		5 Divide-and-Conquer and Updating Methods
			5.1 Divide-and-Conquer Methods
			5.2 Updating Methods
				Online Updating Methods
				Stochastic Gradient Descent
		6 Summary and Discussion
		References
	A Selective Overview of Recent Advances in Spectral Clustering and Their Applications
		1 Introduction
		2 Spectral Clustering
			2.1 The Similarity Matrix
			2.2 Unnormalized Spectral Clustering
			2.3 Normalized Spectral Clustering
			2.4 Equivalence to Weighted Kernel k-Means
			2.5 Selecting the Total Number of Clusters
				General Clustering-Independent Criteria
				Cluster Selection Criteria Specific to Spectral Clustering
		3 New Developments of Spectral Clustering
			3.1 Spectral Biclustering
			3.2 Multi-View Spectral Clustering
			3.3 High-Order Spectral Clustering
			3.4 Constrained Spectral Clustering
			3.5 Evolutionary Spectral Clustering
				PCQ
				PCM
				Determining the Weight Parameter α
			3.6 Incremental Spectral Clustering
			3.7 Sparse Spectral Clustering
		4 Discussion
		References
	A Review on Modern Computational Optimal Transport Methods with Applications in Biomedical Research
		1 Introduction
		2 Background of the Optimal Transport Problem
		3 Regularization-Based Optimal Transport Methods
			3.1 Computational Cost for OT Problems
			3.2 Sinkhorn Distance
			3.3 Sinkhorn Algorithms with the Nyström Method
		4 Projection-Based Optimal Transport Methods
			4.1 Random Projection OT Method
			4.2 Projection Pursuit OT Method
		5 Applications in Biomedical Research
			5.1 Identify Development Trajectories in Reprogramming
			5.2 Data Augmentation for Biomedical Data
		References
	Variable Selection Approaches in High-Dimensional Space
		1 Introduction
		2 Penalized Likelihood Approaches
			2.1 Penalty Functions
			2.2 Canonical Models in High Dimension
				Linear Regression Model
				Logistic Regression Model
				Proportional Hazards Model
			2.3 Algorithm and Implementation
				Penalized Weighted Least Squares
				Penalized Likelihoods
				Tuning Parameter Selection
		3 Feature Screening for Ultra-High-Dimensional Data
			3.1 Sure Independence Screening
				Correlation Ranking
				Maximum Marginal Likelihoods
			3.2 Iterative Sure Independence Screening
			3.3 Reduction of False Positive Rate
		4 Real Data Example
		5 High-Dimensional Inference
		6 Conclusion
		References
	Estimation Methods for Item Factor Analysis: An Overview
		1 Introduction
		2 IFA Models
			2.1 Modeling Framework
			2.2 Examples of IFA Models
			2.3 Exploratory and Confirmatory Analyses
		3 Estimation Methods
			3.1 Estimation Based on Joint Likelihood
			3.2 Estimation Based on Marginal Likelihood
			3.3 Limited-Information Estimation
			3.4 Spectral Method
		4 Computer Implementations
		5 Conclusions
		References
Part IV Survival Analysis and Functional Data Analysis
	Functional Data Modeling and Hypothesis Testing for Longitudinal Alzheimer Genome-Wide Association Studies
		1 Introduction
		2 Functional Modeling of Longitudinal Phenotype Data and Estimation Procedure
			2.1 Model Assumptions
			2.2 Estimation Under the Full Model
			2.3 Estimation Under the Reduced Model
		3 Nonparametric Test on Genotype Effects
			3.1 Generalized Quasi-Likelihood Ratio Test
				Asymptotic Distribution of GQLR Statistic Under the Null
				Power of the GQLR Test
			3.2 Functional F-Test
		4 Implementation Issues
			4.1 Bandwidth Selection
			4.2 Covariance Estimation
				Semiparametric Covariance Estimation
				Nonparametric Covariance Estimation
			4.3 Wild Bootstrap
		5 Simulations
			5.1 Gaussian Case
			5.2 Non-Gaussian Response
		6 Analysis of Longitudinal GWAS Data from ADNI
			6.1 Analysis of the Hippocampal Volume Data
			6.2 Analysis of the RAVLT Data
		7 Summary
		References
	Mixed-Effects Negative Binomial Regression with Interval Censoring: A Simulation Study and Application to Aridityand All-Cause Mortality Among Black South AfricansOver 1997–2013
		1 Background
		2 Methods
			2.1 Simulation
				Statistical Analysis of Simulated Data
				Simulation Summaries
			2.2 South African Data Analysis
				Mortality Data
				Standardized Precipitation Index
				Model Specification and Statistical Analysis
		3 Results
			3.1 Simulation Study
			3.2 Aridity and Mortality Among Black South Africans
		4 Discussion
		5 Conclusion
		Appendix 1: Simulation Results
		Appendix 2: SAS Code Examples
			Simulation Data Generation and Sample Models
			South African Data SAS Code
		References
	Online Updating of Nonparametric Survival Estimatorand Nonparametric Survival Test
		1 Introduction
		2 Notation and Preliminaries
			2.1 Nonparametric Survival Estimator
			2.2 Two-Group Nonparametric Test
		3 Online Updating
			3.1 Online Updating for Nonparametric Estimator
			3.2 Bias-Corrected Online Updating Estimators
			3.3 Online Updating for Nonparametric Test
		4 Simulation Study
			4.1 Simulation for Nonparametric Estimators
			4.2 Simulation for Nonparametric Test
				4.2.1 Under the Null Hypothesis
				4.2.2 Under the Alternative Hypothesis
		5 Real Data Application
			5.1 Survival Estimator
			5.2 Nonparametric Test
		6 Conclusion
		References
Part V Statistical Modeling in Genomic Studies
	Graphical Modeling of Multiple Biological Pathwaysin Genomic Studies
		1 Introduction
		2 Method
			2.1 MRF Modeling of Biological Pathways
				2.1.1 Undirected Graphs and Biological Pathways
				2.1.2 A Nearest Gibbs Measure
			2.2 Combine Multiple Pathways
			2.3 Likelihood Function
			2.4 Posterior Probability Under Bayesian Framework
			2.5 Monte Carlo Markov Chain (MCMC) Simulation
			2.6 Making Inference Based on the Marginal Posterior Probability
		3 Simulation Studies
		4 10-Node Network
		5 27-Node Network
		6 Lung Cancer Data
		7 Discussion
		References
	A Nested Clustering Method to Detect and Cluster Transgenerational DNA Methylation Sites via Beta Regressions
		1 Introduction
		2 Model
			2.1 Model Assumption
			2.2 Identify the Transmission Status
			2.3 Clustering the Transmitted CpG Sites
			2.4 The Likelihood Function and the Posterior Distribution
		3 Simulation Study
			3.1 Simulation Scenarios
			3.2 Results
			3.3 Further Assessment of the Method
		4 Real Data Analysis
		5 Summary and Discussion
		References
	Detecting Changepoint in Gene Expressions over Time: An Application to Childhood Obesity
		1 Introduction
		2 Background
		3 Proposed Method
		4 Power Comparisons
		5 Application to the Obesity Problem
		6 Discussion
		References
Index