Design and Conduct of Observational Epidemiological Studies
3 Credits
P8438
COURSE DESCRIPTION
As a basic science of public health, epidemiology is responsible for identifying causes of disease that can guide the development of rational public health policies. The accuracy of the information provided by epidemiologic studies is therefore of central concern. Epidemiologic methods are the tools we use to make valid causal arguments.
This course builds upon the methods introduced in the Core (or P6031 and P6400). The primary objective is to provide students with the basic tools necessary to conceptualize the design of, and interpret the results from, observational epidemiologic studies.
COURSE LEARNING OBJECTIVES
By the time you complete this course, you should be able to:
• Articulate the relationship between association and causation
• Apply causal concepts to the design and interpretation of epidemiologic studies
• Calculate and interpret basic measures of association
• Develop testable research hypotheses from a causal theory
• Recognize and explain the effects of non-exchangeability
• Distinguish among the sources of non-exchangeability
• Choose study designs appropriate for specific research questions
• Identify sources of, and methods to avoid, invalidity in epidemiologic research
• Relate these sources of invalidity to the definition of a cause
• Estimate the likely direction and magnitude of non-exchangeability in specific studies
• Test research hypotheses using stratification, standardization, and logistic regression
• Interpret logistic regression output to address causal questions
• Define all the terms presented in the weekly glossaries
• Critically evaluate the limitations of current epidemiologic methods
• Work efficiently and productively in a team setting.
ADVANCED PREPARATION
The prerequisites for this course are either the Quantitative Methods Core (P6031) or both Introduction to Biostatistics (P6103/4) and Principles of Epidemiology (P6400).
Students entering this course are assumed to be able to:
• Calculate basic measures of association between exposures and disease
• Interpret data in 2 by 2 tables
• Identify major epidemiologic study designs
• Define confounding, selection bias and information bias (aka measurement error).
COURSE REQUIREMENTS
Class Norms
A goal of this class is to work in teams to have open and robust discussions of the course material. Each team will discuss and develop team norms, which we will synthesize into class norms, to help create an environment where vigorous intellectual arguments can take place.
AI Policy
Academic integrity is a core value at Mailman. For this reason, the use of generative artificial intelligence (AI) sites, (for example, but not only, Chat GPT) to complete an assignment or exam is not permitted unless the course instructor has provided clear written instruction about the use of generative AI.
Use of generative AI to complete an assignment or exam without written instruction from the course instructor will be regarded as the same as receiving unauthorized assistance from another person and can be reported as an academic integrity violation.
Required Course Materials
Savitz, David A. and Wellenius, Gregory A. (2016) Interpreting Epidemiologic Evidence: Connecting Research to Applications. The textbook is available for purchase at the bookstore.
The complete text is also available online through Columbia’s library, at this link:
https://academic.oup.com/book/8266