A systematic error is an “error” or bias that is not affected by increasing the size of the studies. Therefore, incorrect estimates are maintained throughout the entire study. Keep in mind that “bias only occurs when sampling fractions for cases and controls [the people not exposed] is related to exposure”.  [Citation 41.3, 41.5]

 

Types of Systematic Error:

Selection bias: bias in the way study subjects are chosen; the main concern is that subjects have a different probability of being chosen due to exposure or interest

 

Solutions for selection bias:

• use case controls that have been detected under simalar procedures

• obtain sujects that have simalar disease severity

• obtain information on medical care and the type the subjects receive

• obtain controls from the same source, such as clinics, hospital, etc.​

​

Information bias: error in data due to validity, sensitivity (the ability to idenify those with disease), specificity (“ability to identify those who have the disease/factor”), reliability (the ability to replicate results)

 

Some examples:

• recall bias

• exposure identification bias (when exposure data is misclassified and what what classifies as “exposed” is incorrect”)

• interview bias (when interviewers “conduct interviews differently for cases and controls)

​

Solutions for information bias:

• get verification of exposure data“use ‘objective’ markers of exposure” (ex: biomarkers, which are “measurable substances in an organism that indicates disease, infection, or environmental exposure”)

• change the study design​

 

[Citation 41.3, 41.5]

​

 

 

 

​