Available through the International Epidemiology Summer
Muenster, Germany, co-directed by Dr. Tobias Kurth
Introduction to Principals and Methods of Epidemiolog
Five 4-hour lectures
The objective of this course is to develop the participants' ability to
critically evaluate the quality of published medical literature. This
will be accomplished by acquainting the participants with the basic
principles and methods of the design, conduct and interpretation of
epidemiologic studies. Particular emphasis will be placed on the
various epidemiologic strategies, such as descriptive studies
(case-reports and case series, ecologic studies and cross-sectional
surveys), observational analytic studies (case-control and cohort) and
randomized clinical trials. In addition, the course will include basic
biostatistical techniques for the presentation and analysis of data and
the understanding of statistical association as well as cause and
effect relationships. Actual examples will be provided from the
published medical literature.
An Introduction to Propensity Score Analyse
Four hours lecture
The objective of this course is to introduce propensity score analyses
based on clinical examples and theoretical considerations. Particular
emphasis will be placed on how to construct a propensity score model
and on how to utilize the propensity score to adjust for confounding.
Whether the propensity score differs compared to other methods in
confounding control will also be discussed. Participants will receive
theoretical and practical examples of propensity score analyses from
the literature. This course is particularly recommended for individuals
with interest in pharmacoepidemiology and the evaluation of drug safety
studies. Upon completion of the course, participants will have an
overview of what propensity score methods can and cannot accomplish.
Available through the Harvard School of Public Health
Dr. A. Walker
Lectures. Four 2.5-hour sessions each week.
Within the framework of formal epidemiologic analysis, this course covers inference about the effects of pharmaceuticals from case reports, case series, vital statistics and other registration schemes, cohort studies, and case-control studies. Decision-making with inadequate data is examined from the perspectives of manufacturers and of regulators. Students are graded on the basis of group projects. This course is intended primarily for students wishing to pursue a career in the pharmaceutical industry or in national regulatory bodies, but may have more general interest as an applied mid-level course with a heavy methodological emphasis.
Course Activities: Written and oral group projects, individual class presentations, class discussion.
Study Design in Epidemiologic Research
Dr. A. Walker
Lectures. Two 2-hour sessions each week.
Beginning with the randomized clinical trial as a paradigm, this course examines common problems in the design, analysis, and interpretation of observational studies. Cohort and case-control studies are the focus of the discussion, but not to the exclusion of other designs. Problems of exposure and disease definitions, time-dependent effects, confounding, and misclassification are considered in the light of data sources typically available. Relevant statistical methods are introduced but not developed in detail.
Course Activities: Review of published studies, class discussion.
Propensity Score Analysis: Theoretical & Practical Considerations
Drs. T. Kurth and J. Seeger
Five 2-hour lectures and four 2-hour computer lab sessions
This course introduces basic and advanced theory underlying propensity score analyses and provides practical insights into the conduct of studies employing the method. Course readings will include propensity score theory as well as applications. Lectures are complemented by computer lab sessions devoted to the mechanics of estimating and using the propensity score as a tool to control for confounding in observational research. Students should have knowledge in multivariable modeling approaches. A course project will involve the application of propensity scores to a data set.