L24-25: Bias

Question 1

This is a form of systematic error that results in an incorrect estimate of association between exposure and risk of disease.

Answer 1

Bias

Question 2

This is a type of systematic difference between the people who are included in a study and those who are not, or when study and comparison groups are selected inappropriately or using different criteria.

Answer 2

Selection bias

Question 3

This is systematic error due to differences in accuracy or completeness of recall to memory of past events or experiences.

Answer 3

Recall bias

Question 4

This is the result of the tendency of authors to submit, organisations to encourage, reviewers to approve and editors to publish articles containing ‘positive findings’; in contrast to findings or reports that do not report statistically significant or positive results.

Answer 4

Publication bias

Question 5

This is systematic difference between how data on exposure/outcome is obtained.

Answer 5

Information bias

Question 6

What are ways to minimise recall bias?

Answer 6

• providing visual prompts/memory aids
• using objective measures
• validating self-reported information with other information (like medical records)

Question 7

What are ways to minimise information bias?

Answer 7

• measurement instruments: standardised and accurate
• blinding and using objective measures (esp during interviews)
• using clearly defined protocol (eg., having a very strict line of questioning)
• validating self-reported info using objective measures (eg., backing up info with other info, or confirming with measurement instruments)

Question 8

What are ways to minimise interviewer bias?

Answer 8

• using clearly defined protocols and measures (eg., using structured questionnaire and standard prompts)
• blinding
• training of interviewees

Question 9

What are ways to minimise selection bias?

Answer 9

• identify potential sources of bias during design and data collection phase
• minimise loss to follow up (so ensuring a high response rate)

Question 10

What 2 study designs are more susceptible to selection bias?

Answer 10

• case-control studies
• cohort studies

> have to make sure that comparison/control group is representative of source population.

Question 11

How is selection bias reduced in case-control studies?

Answer 11

• high participation
• minimise loss to follow up
• having a clearly defined population of interest (picking cases and controls from same population at same time)
• same inclusion/exclusion criteria for cases/controls
• use reliable measurement instruments

Question 12

The selection of cases and controls must be independent of what to minimise presence of selection bias?

Answer 12

Must be independent of exposure

Remember: case-control studies measure outcome, and exposure likelihood. So how likely are people with the outcome to have the exposure?

Question 13

The selection of comparison and exposed groups must be chosen in what manner of way to minimise selection bias?

Answer 13

Comparison and exposed groups must be chosen together.

Remember: cohort studies measure people based on their exposure status, and outcome likelihood. So how likely are people with the exposure to develop the outcome vs those who don’t?

• must be selected together so that the healthy worker effect doesn’t come into play.
• if they’re selected separately, this introduces selection bias

Question 14

What is the main reason for selection bias in study designs such as Case-Control, Cohort and RCTs?

Answer 14

Loss to follow up

Question 15

What is the main reason for information bias in Cohort studies?

Answer 15

Misclassification of exposure/outcomes

Question 16

This is when the measurement error is different across multiple groups.

Answer 16

Differential misclassification

Question 17

Differential classification can occur when:

Answer 17

• classification of exposure is dependent on outcome (common in historical cohort studies)
• classification of outcome is dependent on exposure (usually introduces interviewer bias)