- Module 1: Introduction to Clinical Trials I: Design
- Module 2: Introduction to Longitudinal Data Analysis
- Module 3 (Day One of Two): Modern Statistical Learning Methods for Observational Biomedical Data
- Module 4: Evaluation of Biomarkers and Risk Prediction Models
- Module 5: Introduction to Clinical Trials II: Conduct and Implementation
- Module 6: Case Studies in Longitudinal Data Analysis
- Module 3 (Day Two): Modern Statistical Learning Methods for Observational Biomedical Data
- Module 8: Leveraging Individual Subject Characteristics to Guide Treatment Decisions: Methodology for Precision Medicine
- Module 9: Introduction to the Design and Evaluation of Group Sequential Clinical Trials
- Module 10: Special Topics in the Design, Conduct and Analysis of Clinical Trials
- Module 11: Introduction to Survival Analysis
- Module 12: Propensity Scores: Methods, Models, and Adjustment
- Module 13: Survival Analysis in Clinical Trials
- Module 14: Developing Prognostic and Predictive Biomarkers with High Dimensional Data
- Module 15: Bayesian Methods for Clinical Research: Introduction
- Module 16: Global Sensitivity of Randomized Trials with Missing Data: Methods and Software (tentative title)
- Module 17: Survival Analysis in Observational Studies
- Module 18: Bayesian Methods for Clinical Research: Computational Methods and Applications
Session 1 – Monday, July 23, 8:30 a.m.-5 p.m.
Module 1: Introduction to Clinical Trials I: Design
This module provides an overview of the primary design issues that must be considered by the many disciplines that collaborate on a clinical trial. We will consider the clinical, scientific, and regulatory setting of clinical trials, describing the phased approach to “treatment discovery” in which the safety, efficacy, and effectiveness of candidate treatments are investigated.
This one-day module focuses on the issues surrounding disease area and identification of the target population, definition of the treatment, choice of clinical outcomes, and choice of comparators. Throughout, the instructors will emphasize pre-specification of these key elements in a well-defined study protocol.
Module 2: Introduction to Longitudinal Data Analysis
Instructor(s): Sitlani, Colleen
Longitudinal studies follow individuals over time and repeatedly measure health status. This module will provide an overview of statistical methods for the analysis of longitudinal data.
A brief introduction to pre-post data will be followed by an in-depth discussion of regression-based methods such as generalized linear mixed-effects models and generalized estimating equations. Relevant theoretical background will be provided. Examples from randomized controlled trials and observational studies will be used to illustrate key approaches and modeling strategies. Illustrative examples will be presented with R and STATA.
This course is targeted toward individuals with little or no prior exposure to statistical methods for longitudinal data analysis. Attendees interested in more experience applying these methods should also consider Module 6: Case Studies in Longitudinal Data Analysis.
Module 3 (Day One): Modern Statistical Learning Methods for Observational Biomedical Data
While clinical trials provide the highest level of evidence to compare clinical treatments or public health interventions, they are often not feasible due to ethical, logistic or economical constraints. Observational studies provide an opportunity to learn about the effect of interventions for which little or no trial data are available. These studies constitute a potentially rich and relatively cheap source of information. However, in such studies, treatment or intervention allocation may be strongly confounded by other important patient characteristics and much care is needed to disentangle observed relationships and infer causal effects.
In this course, we will provide an overview of modern statistical techniques for analyzing observational data. We will focus primarily on recent advances in the field of targeted learning, which uses of state-of-the-art machine learning tools to flexibly adjust for confounding while yielding valid statistical inference. In contrast, conventional techniques for confounding adjustment rely on restrictive statistical models and may therefore lead to severely biased inference. Use of the Super Learner framework, an implementation of model stacking, will be discussed as a particularly appealing means of performing flexible, pre-specified adjustment for confounding.
We will discuss methods for comparative effectiveness studies for single time-point interventions. We will also introduce the multi time-point extension of these methods and discuss strategies for dealing with missing data. Methods will be illustrated using data from recent observational studies and extracted from electronic medical records. Analyses will be illustrated in R but knowledge of R is not required for this module. In addition to lectures, the course will include in-class hands-on activities to allow students to familiarize themselves with the methods and tools.
Module 4: Evaluation of Biomarkers and Risk Prediction Models
Instructor(s): Kerr, Kathleen
This module covers methodology for evaluating biomarkers and risk prediction models, covering principles, concepts, metrics, and graphical tools.
We will discuss motivations for risk prediction in clinical medicine and public health, and clarify the concept of “personal” risk. Metrics and graphical tools will include ROC curves and AUC; calibration plots for risk prediction models; and net benefit and decision curves. The module will also discuss methods for comparing risk prediction models and, in particular, assessing the prediction increment of a new biomarker. We will also consider evaluating the utility of a single or composite biomarker for prognostic enrichment of a clinical trial.
There will be an opportunity for hands-on practice in R using relevant packages such as rms, rmda, and BioPET. However, the software component of the module is small and knowledge of R is not required for this module.
Session 2 – Tuesday, July 24, 8:30 a.m.-5 p.m.
Module 5: Introduction to Clinical Trials II: Conduct and Implementation
This module presents topics regarding the conduct and implementation of clinical trials that build upon the key concepts introduced in the Module 1: Introduction to Clinical Trials I: Design. Participants in this module are encouraged to take Module 1: Introduction to Clinical Trials I: Design, although, this is not required for participants who believe they have the necessary background.
This one-day module focuses on choice of randomization strategies, blinding, specification of secondary endpoints, handling of missing data, study conduct and monitoring, the role of an independent data monitoring committee, and plans for reporting of the result. Throughout the module, the instructors emphasize pre-specification of these elements in a well-defined study protocol.
Module 6: Case Studies in Longitudinal Data Analysis
Instructor(s): French, Ben
This course is targeted toward individuals who take Module 2: Introduction to Longitudinal Data Analysis or have background knowledge of longitudinal data analysis methods.
Longitudinal studies follow individuals over time and repeatedly measure health status. This module will focus on the practical application of statistical methods to the analysis of longitudinal data.
A series of case studies will be used to discuss analysis strategies, the application of appropriate analysis methods, and the interpretation of results. Case studies will arise from randomized controlled trials and observational studies.
Software implementation will be primarily in R and STATA; attendees should bring a laptop with relevant software installed.
Module 3 (Day Two): Modern Statistical Learning Methods for Observational Biomedical Data
Module 8: Leveraging Individual Subject Characteristics to Guide Treatment Decisions: Methodology for Precision Medicine
There is currently great interest in developing rules for recommending for or against treatment based on individual subject characteristics, especially in clinical settings where treatment is toxic, burdensome or costly, or where the condition being treated is thought to be heterogeneous. This course will focus on individual characteristics used to guide treatment decisions, sometimes called predictive, prescriptive, or treatment-selection biomarkers. These characteristics might be as simple as subject demographics, clinical characteristics or disease risk factors, or might be more complex, such as the results of gene expression technology or medical imaging. This course will describe preferred statistical approaches for discovering individual characteristics predicting treatment efficacy, and for developing and evaluating treatment rules. We will also discuss study design in this context. Students will receive training using R packages implementing the methodology.
Session 3 – Wednesday, July 25, 8:30 a.m.-5 p.m.
Module 9: Introduction to the Design and Evaluation of Group Sequential Clinical Trials
Increasingly, clinical trials are conducted using group sequential methods in order to address the ethical and efficiency requirements for performing experiments with human volunteers. The design, conduct, and analysis of a sequential clinical trial is necessarily more involved than that for a clinical trial in which the data would only be analyzed at the end of the study.
In this module, we provide an introduction to methodology for the design and evaluation of group sequential trials from both a frequentist and Bayesian perspective. Methodology will be presented using case studies and examples of clinical trials with continuous, binary, and survival endpoints.
The sequential methods will be illustrated using RCTdesign, an R module for the design, monitoring, and analysis of clinical trials. Emphasis will be placed on both the scientific and statistical impact of group sequential trial design.
Module 10: Special Topics in the Design, Conduct and Analysis of Clinical Trials
Instructor(s): Fleming, Thomas
This module has two key segments. In the morning, two topics will be discussed that have broad implications in the design and analysis of clinical trials: the role of biomarkers as replacement (i.e., surrogate) endpoints, and the interpretability of confirmatory vs. exploratory analyses. In the afternoon, attention will be given to a topic having broad implications in the conduct of clinical trials: the role of Data Monitoring Committees and current challenges in their implementation.
Module 11: Introduction to Survival Analysis
Censored time-to-event data, where not all subjects experience the event of interest, are common in clinical and epidemiologic research. Examples include randomized controlled trials of therapies for cancer and other chronic diseases, comparative effectiveness research, and epidemiologic cohort studies. This module provides an introduction to censored time-to-event data and classical survival data analysis methods used in biomedical studies.
In this module, we will provide examples of studies where survival analysis is used and where it should not be used. We will describe how incomplete data on time-to-event outcomes (censoring) occurs. We will introduce important functions, including the survival function, the hazard function, and the median survival time, and show how censored data can be used to estimate them and compare the time-to-event experience between groups.
The module will explain key concepts unique to survival analysis such as risk sets and informative censoring. It will introduce the Cox regression model, and show how to examine the proportional hazards assumption.
The course will focus on application and understanding the concepts with examples from the biomedical literature; mathematical details will be kept to a minimum.
Module 12: Propensity Scores: Methods, Models, and Adjustment
Instructor: Stephens, David
The propensity score is a key component of many causal inference procedures. After establishing the basic causal inference framework, we will outline the key methods of construction of propensity score functions, and study their core mathematical properties. We will detail the use of the propensity score in matching, inverse weighting and regression adjustments that allow the unconfounded effect of an exposure or treatment of interest to be estimated consistently.
Using the framework of semiparametric inference, we will contrast the statistical properties of estimators derived using each method. We will investigate issues of model selection for the propensity score, and demonstrate the utility of judicious choice of predictors that enter into the propensity function. This will be illustrated in standard problems and also in the case of high-dimensional predictors. Longitudinal data will also be studied in the causal setting.
Finally, we will develop the Bayesian framework for handling causal inference and investigate how propensity function construction and usage translates to the new setting.
All methods will be illustrated using examples from biostatistics, health research and econometrics. Computation will be performed in R.
Session 4 – Thursday, July 26, 8:30 a.m.-5 p.m.
Module 13: Survival Analysis in Clinical Trials
This module will cover advanced topics in survival analysis, with an emphasis on applications in randomized clinical trials (RCTs) with censored time-to-event outcomes. The module will:
- Review the logrank test and introduce testing procedures that weight group comparisons differently over the follow-up time interval;
- Introduce the restricted mean survival time and tests to compare it between groups;
- Review methods suitable for examining the adjusted association between an explanatory variable and a censored event-data outcome;
- Cover how information is accrued when there is group-sequential monitoring, and
- Cover power and sample size computations for an RCT with censored time-to-even outcomes.
The course will focus on application and understanding the concepts with examples from the literature; mathematical details will be kept to a minimum. Knowledge of material in the Module 11: Introduction to Survival Analysis will be assumed.
Module 14: Developing Prognostic and Predictive Biomarkers with High Dimensional Data
Modern medicine has graduated from broad spectrum treatments to molecularly targeted therapeutics. A major medical challenge is to develop companion diagnostics that provide guidance in directing use of these drugs to patients whose prognosis warrants them and whose disease is likely to be responsive to them. With continued advances in high throughput biotechnology, it is feasible to run whole genome assays on disease tissue whose features can be developed into prognostic and predictive biomarkers.
Specialized statistical methods are required for the development and evaluation of multivariate biomarkers using such assays because the number of candidate features is generally much greater than the number of patients. The goal of this module is to introduce ideas in high dimensional predictive modeling, to discuss model validation and testing, and to give hands-on experience with these tools to build predictive biomarkers on real, high-dimensional datasets.
Through the course, participants will become familiar with various methods in penalized regression, applications of cross-validation, as well as ideas in multiplicity and selection bias. Participants will gain experience applying these methods in R.
Module 15: Bayesian Methods for Clinical Research: Introduction
This module provides an introduction to Bayesian methods for biomedical research. Specifically, we discuss the Bayesian paradigm introducing the subjective interpretation of probability; Bayes Theorem; and prior, posterior and predictive distributions. We contrast the Bayesian and frequentist approaches using simple biomedical problems including diagnostic testing and design and monitoring of clinical trials, among others.
This module uses a number of R packages to illustrate the application of Bayesian methods to analyze independent data.
Pre-requisite: introductory course in statistics/biostatistics; linear regression; familiarity with R/RStudio.
Session 5 – Friday, July 27, 8:30 a.m.-5 p.m.
Module 16: Global Sensitivity Analysis of Randomized Trails with Missing Data: Methods and Software (tentative title)
Instructor(s): Scharfstein, Daniel
Description coming soon!
Module 17: Survival Analysis in Observational Studies
This module will cover advanced topics in survival analysis, with an emphasis on applications in studies relying on observational data, including epidemiologic cohort studies and comparative effectiveness studies. This module will:
- Describe Cox regression models suitable for examining adjusted associations and effect modification;
- Cover the important choice of the time scale for the analysis, and discuss how to analyze data on subjects who enter observation after time zero (left entry and left truncation);
- Cover methods for appropriate inferences when there are competing risks, including the Cox regression model for cause-specific hazard functions and the Fine-Gray model for hazards related to the cumulative incidence function;
- Discuss biases that can arise in observational data: confounding, immortal time bias and index event bias, and how to treat them in the analysis, and
- Cover issues related to time-dependence in the Cox regression model, including how to incorporate hazard ratios, exposures, and adjustment variables that depend on time using time-dependent coefficients, time dependent covariates, and time-dependent stratification.
The course will focus on application and understanding the concepts with examples from the biomedical literature; mathematical details will be kept to a minimum. Knowledge of material in the Module 11: Introduction to Survival Analysis will be assumed.
Module 18: Bayesian Methods for Clinical Research: Computational Methods and Applications
In this module, we discuss Bayesian hierarchical modeling and general computational methods for Bayesian estimation of hierarchical models. We apply hierarchical models to some biomedical problems including diagnostic testing in the absence of a gold standard, meta-analyses, and mixed treatment comparisons, among others.
We use INLA, JAGS and a number of R packages to illustrate the application of Bayesian methods to analyze dependent data.
Pre-requisites: Module 15: Bayesian Methods for Clinical Research: Introduction or other first course on Bayesian statistics; familiarity with R/RStudio.