Epid 5310-Lecture 6

Front 1

systematic error

Back 1

Validity

Front 2

Random error

Back 2

Precision

Front 3

Precision reflects...

Back 3

Sampling variability

Front 4

Result of systematic error

Back 4

Bias

Front 5

Bias

Back 5

The result of systematic error in the design or conduct of a study

Front 6

Prevention of bias is accomplished by:

Back 6

1. Ensuring the study design is appropriate
2. Establishing valid and reliable measures for data collection

Front 7

Selection bias

Back 7

-Occurs when the relations for sampling validity of causal contrast study design do not hold
-Results from procedures used to select subjects and from factors that influence study participation such that the relation between exposure and disease is different for those who participate in the study and target population.

Front 8

Selection bias does not result from:

Back 8

Does not result from differential selection by exposure status alone or by outcome status alone
-the impact of the bias on the measure of association can be towards or away from the null
i.e., the exposure effect may be underestimated or overestimated

Front 9

Selection bias in a cohort study

Back 9

Disease status influences selection or participation to a different degree in exposed and unexposed groups (attrition)
-worried about loss to follow up.

Front 10

Selection bias in a case-control study

Back 10

Exposure status influences selection or participation to a different degree among cases and controls.
worried about your controls being different from your cases.

Front 11

Selection bias in a cross-sectional study

Back 11

Either variable (disease or exposure) influences selection or participation to a different degree in comparison groups

Front 12

Common sources of selection bias

Back 12

Bias is introduced it the disease of interest influences the loss of subjects to follow-up differently for exposed and unexposed cohort members, or stated differently, if exposed and unexposed subjects who are lost to follow-up differ in the relative disease frequencies from exposed and unexposed subjects who are followed.

Front 13

Cohort study-Selective loss to follow-up

Back 13

The amount of bias may differ for a given amount of loss to follow up but the greater the absolute loss, the greater the maximum possible bias

Front 14

Case-control studies: Incomplete ascertainment of cases/ diagnostic bias

Back 14

-Bias is introduced if exposure of interest influences the detection of disease