Bias Study

Introduction
to Study

Step 2: Introduction to the Study

The internship at the Epiville Department of Health has kept you very busy. After having completed investigative work on the two outbreaks (SARS and Susser Syndrome), you feel that in addition to learning "shoe leather epidemiology" you would like to gain a deeper understanding of some of the complexities of epidemiologic methods.

One problem in particular - how to make sure that you have drawn appropriate study samples - is worrying you. In the cohort and case-controls studies you completed over the last few weeks you learned how to properly choose exposed and unexposed groups and cases and controls. For instance, recall Question 3 from Step 4 of the cohort study where you were asked to define eligibility criteria for study participants. The correct answer was to choose those who have worked at the factory for at least two years and who were shown to be healthy at their initial or annual health check-ups (these people had to be employed at the factory from September 2000 to September 2002).

Dr. Zapp explained to you that if you decided to choose everyone at the time point when you began your study (September 2002), you would have introduced selection bias because "everyone" could have included persons who have been on the job for less than two years. She also said that because the minimal induction period is at least 6 months if you had included in your sample people who were employed for less than the induction period, the association between exposure and disease would have been diluted since these new employees would not have had the same opportunity as the long-term employees to develop the disease. This made sense to you, but you begin to worry that there must be many other possible methodologic errors just waiting to derail your march from wide-eyed intern to senior investigator.

You begin to lose sleep over this. Your friends suggest that you "get a life" but, instead, you decide, once again, to turn to Dr. Morissa Zapp. She tells you that you must be very careful in interpreting study results and pay particular attention to the choice of your comparison groups to prevent selection bias. Furthermore, she says that if you are not careful in your interviewing techniques then you can introduce information bias. Dr. Zapp recommends that you undertake careful study of the concept of bias by exploring two studies that were conducted by her friends, prominent scientists at the Mailman School of Public Health at Columbia University.

The first study which you will look at to learn about biases is "Artificial Sweetener Use and Bladder Cancer" by E.L. Wynder and Steven D. Stellman. This study was extremely influential in resolving the debate about the safety of saccharin. The second study, conducted by Dr. Samuel Shapiro and colleagues, looked at the relationship between estrogen and cancer of the endometrium in women, and is another classic study that resolved a different critical debate in the scientific community.

Learning
Objectives

Introduction
to Study