Bias and Cofounding Flashcards

1
Q

What is bias?

A

Bias refers to systematic error, meaning that multiple replications of the same study would reach the wrong answer on average.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What is the effect of bias on validity?

A

The concept of bias is the lack of internal validity or incorrect assessment of the association

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

When may bias be introduced?

A
  • Literature review
  • Sample specification and selection
  • Executing experimental manoeuvre or exposure
  • Measuring exposures and outcomes
  • Data analysis
  • Data interpretation
  • Publication of results
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What is a selection bias?

A

Arises from procedures used to select study subjects and from factors that affect study participation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What is the consequence of selection bias?

A

The relation between exposure and disease is different for those who participate in the study (the study population) and those eligible for inclusion (the target population)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What types of selection bias is there?

A
  • Inappropriate definition of the eligible population
  • Lack of accuracy of the sampling frame
  • Uneven diagnoses in the target population
  • During study implementation
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How to deal with selection bias?

A

Choose a study population with care.

Maximise take up and follow up.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What is information bias?

A

Arises during data collection

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What types of information bias exist?

A
  • Observer (interviewer) bias
  • Recall bias
  • Reporting bias
  • Hawthorne effect
  • Detection bias
  • Protopathic bias
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What is observer bias?

A

An interviewer’s knowledge may influence the structure of questions and the manner of presentation, which may influence responses

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

What is recall bias?

A

Those with a particular outcome or exposure may remember events more clearly or amplify their
recollections

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is reporting bias?

A

Participants might give answers they perceive to be of interest to the investigators

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What is the Hawthorne effect?

A

An effect first documented at a
Hawthorne manufacturing plant; people act differently if they know they are being watched

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is detection bias?

A

If greater effort is given to diagnosing disease in an exposed group than in an unexposed group

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What is detection bias?

A

Exposure is influenced by early (subclinical) stages of disease

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

How to deal with information bias?

A

Blinding

Standardised, definitions, exposure and data collection.

Max response to the questionnaire.

Objective data where possible

17
Q

What are some examples of analysis and interpretation biases?

A
  • Post-hoc significance bias
  • Data dredging bias
  • ‘Mistaken identity’ bias
  • Cognitive dissonance bias
  • Significance bias
18
Q

How to deal with analysis bias?

A

*State hypotheses to be tested in advance
*Follow the analysis plan

19
Q

How to deal with interpretation bias?

A

*Treat all interpretations with caution
*Look for evidence from systematic reviews

20
Q

What is confounding?

A

The effect of an extraneous factor is mistaken for or mixed with the effect of the exposure(s) under investigation

21
Q

What does extraneous mean?

A

Introduced or coming from without; irrelevant

22
Q

What is Simpson’s Paradox?

A

A correlation present within different groups may disappear or even be reversed when the groups are combined

23
Q

How to control confounding in study design?

A

*Restriction
*Matching
*Randomisation

24
Q

How to control confounding in analysing the study?

A

*Stratification
*Standardisation