20-24 August 2012 (Stellenbosch) Advanced Epidemiological Methods. 5-day Course presented by Dr. Matt Fox. Course completed

Posted on Sun, Aug 19 2012 23:00:00


Course completed.



Dr. Matthew Fox of the Department of Epidemiology and the Center for Global Health and Development at Boston University will be presenting an intensive 5-day course on advanced epidemiological methods at Stellenbosch University under the auspices of the South African DST/NRF Centre for Epidemiological Modeling and Analysis (SACEMA). The course will take place from 9 am to 4:30 pm daily at a venue in Stellenbosch (to be announced).  The course registration fee is R2 000.  Application should be made online at www.sacema.com. Enquiries may be directed to the SACEMA Research Manager, Ms Lynnemore Scheepers, at  scheepersl@sun.ac.za  (and copy to Matthew Fox <mfox@bu.edu>), or phone: 021 808 2589/2893.


Introductory and intermediate courses in epidemiological methods teach students the concepts needed to begin a career conducting valid epidemiological research; however these courses typically only briefly cover the causal models that should underlie the design of valid epidemiological studies. We will use these models as a jumping-off point to begin rethinking what we have already learned and to go further in our understanding of basic concepts of measures of effect, confounding, misclassification and selection bias. From there we will begin to question the implications of various sources of bias in our studies and we will work through novel methods and approaches for doing more than simply speculating about these biases. We will then finish by exploring the basic statistics used in epidemiological research and we will correct misunderstandings about what these statistics can tell us.

Throughout the course we will focus on the core concepts of validity and precision and will further develop our understanding of these central concepts. We will emphasize the development of skills that every doctoral level epidemiologist should have, skills that are both practical and marketable. Note that this course is not offered for any credit. It is a course designed to help doctoral level and advanced master’s students advance their skills.


A rough sketch of the session titles for each day have been provided below in the timetable but are subject to change.



August 20


August 21


August 22


August 23


August 24








9.00h – 11.30h


INTRODUCTION TO MODERN EPIDEMIOLOGY: Review of basic epidemiology and introduction to advanced epidemiologic concepts


THE POTENTIAL OUTCOMES MODEL: Confounded definitions of confounding


NOVEL APPROACHES TO DEALING WITH CONFOUNDING: Propensity Scores and Marginal Structural Models






A SHOW OF CONFIDENCE: Random Error II: P-values or confidence intervals?

11.30h – 13.30h


13.30h – 16.00h

THE SUFFICIENT CAUSES MODEL: Introduction to causal models and the benefits of basis of causal thinking.

STRUCTURAL APPROACHES TO BIAS: Directed Acyclic Graphs and the potential harms of statistical adjustment?

THREE CONCEPTS OF INTERACTION: What do really mean by interaction?

THE ABUSED P-VALUE: Random error I: what’s in a p-value?

STATISTICAL ALTERNATIVES: Introduction to Bayesian thinking




Students are expected to prepare fully for class by reading the material ahead of time. There will be two readings per day, listed below. The readings for this course are challenging. You will not understand everything you read and you will need to read the articles more than once before coming to class. You will struggle if you wait until the last minute or do not read the material. The class can only function if you are prepared and have challenged yourself to think about the material. Come to class with questions about what you do not understand.




Ioannidis JP. Why most published research findings are false. PLoS Med 2005;2(8):e124.


Rothman, KJ.  Causes.  Am J Epidemiol 1976;104:587–592


Greenland S and Robins JM.  Identifiability, exchangeability, and epidemiological confounding.  Int J Epidemiol 1986;15:412–418


Greenland S, Pearl J, and Robins JM.  Causal diagrams for epidemiologic research.  Epidemiology 1999;10:37–48.


Hernan MA, Brumback B, Robins JM. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology 2000;11(5):561-570.


Rothman KJ.  Modern Epidemiology, 1st Edition.  Chapter 15 - Interaction between Causes.  Little, Brown, and Company, Boston, MA:  1986.  pp 311–326.


Jurek AM, Greenland S, Maldonado G et al. Proper interpretation of non-differential misclassification effects: expectations vs observations. Int J Epidemiol 2005;34(3):680-687.


Greenland S.  Randomization, Statistics, and Causal Inference.  Epidemiology 1990;1:421–429.


Goodman, S N. Toward Evidence-Based Medical Statistics. 1: The P Value Fallacy. Annals of Internal Medicine 1999;130(12): 995-1004. Paper available at http://www.acponline.org.

Poole, C. Low P-Values or Narrow Confidence Intervals: Which Are More Durable? Epidemiology 2001;12(3): 291-294.