Bias

Introduction

>

Learning Objectives

>

Student Role

>

Design Questions

>

Data Analysis Questions

>

Discussion Questions

>

Print Module

Step 3: Design Questions

1.	Suppose that you were told that the clinical setting from which researchers got their cases and controls was a tertiary care facility, which is a specialized hospital equipped with diagnostic and treatment facilities to treat bladder cancer. (Please note that different tertiary care facilities may be specialized to treat different diseases and therefore people may travel from long distances to get to the facility. They may also be admitted to it simply because they happen to live nearby.) What type of bias might be introduced in the selection of controls and why?
		a.	Selection bias
		b.	Recall bias
		c.	Self-selection bias

Interactive exercise: Hypothetical selection of cases and controls for the hospital-based case-control study investigating the potential association between diet soda consumption and bladder cancer to show selection procedures and their effects.

[ Open Interactive Exercise in New Window ]

2.	Suppose that you are designing a case-control study similar to the one described in the interactive module. Which method of accruing cases and controls do you think is the most practical?
		a.	Situation 1: Cases are selected from the Epiville General Hospital. Because it is a tertiary care medical facility, they come from different areas throughout Epiville. Controls are selected from the outpatient clinic for tattoo removal.
		b.	Situation 2: Epiville Cases are selected from the Epiville General Hospital and since it is a tertiary care medical facility, they come from different geographical areas- both within and outside of Epiville. (As you may recall, Epiville General is a tertiary care facility and people from many areas come to see its specialists.) Controls are selected from the admissions office with diagnoses of cancer, other than the cancer under study, and not related to our exposure, or with diagnoses of non-cancer diseases. Controls are matched to cases on sex, age, ethnicity, and SES.
		c.	Situation 3: Cases come from the source population around the Epiville General Hospital. Controls are selected from the admissions office with diagnoses of cancer other than the cancer under study, or with diagnoses of non-cancer disease, neither of which are related to our exposure.

3.	Given a choice of how and where to obtain cases, what do you think would be the best source of subjects with bladder cancer?
		a.	Hospital Records
		b.	Tumor Registry
		c.	Death Certificates

4.	What potential problem could have been introduced if you found out that the interviews with cases took 30 minutes longer on average than the interviews with controls?
		a.	Selection bias
		b.	Information bias
		c.	Volunteer bias

5.	In the study of artificial sweetener and bladder cancer, authors collected information on the consumption of artificial sweetener and diet beverages. The duration of artificial sweetener/diet beverage use differed among cases and controls. Some individuals reported having only consumed these products more than 5 or 10 years prior, whereas others reported current exposure to these products. What potential problems could arise when trying to measure exposures that happened over different time periods?
		a.	Misclassification bias
		b.	Volunteer bias
		c.	Surveillance bias

6.	What effect (if any) would you expect if the interviewers were aware of the disease status of the study subjects?
		a.	It would benefit the study since the interviewers would collect better data for the cases and thus understand more precisely how the exposure is related to the disease.
		b.	It would definitely not change the results of the study.
		c.	It could irreparably damage the data by introducing the interviewer bias.