Advanced-level topics in biostatistics offered by regular and visiting faculty. Prerequisite: permission of instructor. Offered: jointly with STAT 578; AWSpS.
Course Description Development of new biomarkers and medical tests has rapidly accelerated in recent years. Their rigorous evaluation is a high priority for research. Yet principles and techniques for the design and analysis of these studies are not widely known. Moreover, there are fundamental differences with standard statistical methods that exist for therapeutic and etiologic studies.
This course will present concepts and techniques necessary for evaluating biomarkers or other medical tests used for classification or prediction purposes. This includes applications in diagnosis, disease screening, prognosis and risk prediction. However, we will not cover in depth evaluation of biomarkers/tests for: use as surrogate variables in treatment clinical trials; measuring exposures in etiologic research; or selecting optimal treatment choice for individuals. The focus will be on rigorous evaluation of the classification/prediction performance of a proposed marker or marker combination. Although many concepts apply to biomarker discovery, that is not a focus of this course.
The following topics will be covered: (1) Quantifying performance; (2) Estimating and comparing performance; (3) Adjusting for covariates; (4) Evaluating factors affecting biomarker performance; (5) Study design strategies; (6) Sample size calculations; (7) Methods for evaluating risk prediction markers/models; (8) Special issues for event time outcomes; (9) Combining markers and assessing incremental value; and (10) Missing or mismeasured reference standards.
Student learning goals
Students should learn strategies for design and analysis of studies for evaluating medical tests and biomarkers for classification.
They should know how to implement methods in Stata and have sufficient knowledge to program methods in other languages.
Students will become aware of some common pitfalls and misconceptions in biomarker/test evaluation.
Students will be aware of gaps that exist in statistical methodology for this field.
General method of instruction
traditional: lectures with in class discussion.
Introductory level statistics: Biostat 517-518 preferred but Biostat 511-12-13 will suffice
Class assignments and grading
Quizes: 10 minute quiz each Tuesday on previous week’s material drop lowest score, no make ups Homework: weekly, due Thursday of the following week, graded credit/no-credit;Solutions will be provided Midterm: in class February 4 Final exam: in class March 13, take home analysis March 6-13
Grade: average quizzes(30%)+midterm(30%)+final(30%)+homework(10%)