3.5.1 information bias

40
Information bias Information bias Bias resulting from flawed definition of study variables or measurement of study variables Results in erroneous classification of subjects with regard to exposure and/or outcome – this is called misclassification

Upload: a-m

Post on 15-Apr-2017

785 views

Category:

Education


0 download

TRANSCRIPT

Page 1: 3.5.1 information bias

1

Information bias• Information bias

– Bias resulting from flawed definition of study variables or measurement of study variables

– Results in erroneous classification of subjects with regard to exposure and/or outcome – this is called misclassification

Page 2: 3.5.1 information bias

2

Information bias• There are two types of misclassification:

– Non-differential misclassification– Differential misclassification

• Definitions of these terms depend on the variable being measured (i.e., exposure or outcome)

Page 3: 3.5.1 information bias

Information bias• Types of misclassification of outcome variables

– Non-differential misclassification of outcome• The degree of outcome misclassification is not related to

exposure status– Differential misclassification of outcome

• The degree of outcome misclassification depends on the exposure status – this is a more serious problem

3

Page 4: 3.5.1 information bias

Information bias• Types of misclassification of exposure

variables– Non-differential misclassification of exposure• The degree of exposure misclassification is not related to

outcome status– Differential misclassification of exposure

• The degree of exposure misclassification varies by outcome status – this is a more serious problem

4

Page 5: 3.5.1 information bias

Information bias• Some specific exposure related information

biases– Recall bias: occurs when participants are asked

about past exposure after the outcome in question has occurred (or not), as often happens in case- control studies

5

Page 6: 3.5.1 information bias

6

Information bias– The respondents’ memories vary according to

whether or not they experienced the outcome, especially if the exposure is a commonly known risk factor for the disease they have experienced• Those with disease and the exposure more likely to recall

exposure– Increased sensitivity

• Those with disease and not exposed more likely to report exposure– Reduced specificity

• Will explain use of sensitivity and specificity to quantify information bias shortly

Page 7: 3.5.1 information bias

Information bias• Some specific exposure related

information biases– Recall bias example:

• Case-control study of gestational pesticide exposure and offspring developmental delay

50

Page 8: 3.5.1 information bias

8

Information bias– Recall bias example (cont.):

• Mothers with developmentally delayed children may more comprehensively recall their exposures during pregnancy or may over-report them, having spent time thinking about what might have caused their child’s disability

• Control mothers with typically developing children have not spent time pondering prenatal exposures, and thus may be less likely recall exposure

Page 9: 3.5.1 information bias

Information bias• Some specific exposure related information

biases– Interviewer bias: occurs when interviewers are not

blinded to participant disease status

9

Page 10: 3.5.1 information bias

Information bias– Interviewer bias:– Interviewers may question diseased and non-

diseased differently, for example emphasizing some words or questions, or asking more clarifying questions of those with disease in an attempt to elicit information on the exposure

10

Page 11: 3.5.1 information bias

11

Information bias• Some specific outcome related information

biases– Observer bias: occurs when observers/raters are

not blinded to exposure status (analogous to interviewer bias, except affects disease classification)

– Observers/raters may be more likely to count cases among participants with high risk/exposure profiles

Page 12: 3.5.1 information bias

Information bias• Some specific outcome related information

biases– Observer bias example:

• A sample of nephrologists were sent patient case histories with a simulated race randomly assigned to each case

• When the case history identified the patient as black, the nephrologists were twice as likely to diagnose the patient as hypertensive end-stage renal disease, as compared to patients labeled white

12

Page 13: 3.5.1 information bias

Information bias• Some specific outcome related information

biases– Respondent bias: participants with high

risk/exposure profiles may be more likely to report the outcome of interest

13

Page 14: 3.5.1 information bias

14

Information bias• Effects of non-differential versus differential misclassification

– In practice, it is impossible to correctly measure/collect all variables: some misclassification is inevitable

– Thus, it is important to thoroughly evaluate your exposure and outcome definitions, study protocol, and data collection procedures to evaluate what likely measurement error exists

– Then, think about the extent and direction of bias

Page 15: 3.5.1 information bias

15

Information bias• Non-differential misclassification

– Results in a bias toward the null when the exposure or disease that is misclassified is binary

– For example, when a binary exposure is measured with equal amount of error between case and control groups, it washes out the exposure-outcome association

– This is a conservative bias, and the investigator at least knows that she/he is not presenting an artificially large association

Page 16: 3.5.1 information bias

16

Information bias– Non-differential misclassification when there are

more than two categories of the exposure or disease does not necessarily result in bias towards the null

– Categorization of a variable that has non-differential misclassification can generate differential misclassification

Page 17: 3.5.1 information bias

60

Information bias• Differential misclassification of exposure or disease results in a bias

in an unpredictable direction – it may be toward the null or away from the null

• It is possible to evaluate the bias on a case-by-case basis and speculate the direction of the bias, however the possibility of bias away from the null is problematic

• Generally considered a more serious problem than bias towards the null because

– (a) the investigator does not know the direction of the bias with certainty, and– (b) if the bias is away from the null, the investigator risks presenting an

artificially inflated effect estimate vs. an attenuated one

Page 18: 3.5.1 information bias

18

Information bias• Misclassification of a confounding variable

– Bias in an unpredictable direction

Page 19: 3.5.1 information bias

Information bias• Numerical example of non-differential

misclassification

19

Page 20: 3.5.1 information bias

20

Information bias• Measures useful for quantifying information

bias– Sensitivity

• P(classified positive|true positive)– Specificity

• P(classified negative|true negative)

Page 21: 3.5.1 information bias

Information bias

• Case-control study data – the true distribution of exposure

• OR=?

21

Page 22: 3.5.1 information bias

Information bias22

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 23: 3.5.1 information bias

Information bias23

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 24: 3.5.1 information bias

Information bias• True positive (exposed)

cases = TP• Classified positive

(exposed)– = TPx(class pos|TP)–

= 80 x 0.9• Classified negative

(unexposed)– = TPx(1-(class pos|TP))– = TPx(class neg|TP)– = 80 x 0.1– Or = TP-class. positive

24

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 25: 3.5.1 information bias

Information bias• True negative (unexposed)

cases = TN• Classified positive

(exposed)– = TNx(1-(class neg|TN))– = TNx(class pos|TN)– = 20 x 0.2– Or = TN-class. negative

• Classified negative (unexposed)

– = TNx(class neg|TN)– = 20 x 0.8

25

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 26: 3.5.1 information bias

Information bias26

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 27: 3.5.1 information bias

Information bias27

Page 28: 3.5.1 information bias

Information bias28

Page 29: 3.5.1 information bias

Information bias29

Page 30: 3.5.1 information bias

30

Information bias• Non-differential because sensitivity and

specificity the same for cases and controls• Resulted in bias towards the null

– True OR = 4– Misclassified OR = 2.6

Page 31: 3.5.1 information bias

Information bias• Numerical example of differential

misclassification

31

Page 32: 3.5.1 information bias

Information bias

• Case-control study data – the true distribution of exposure

• OR=4

32

Page 33: 3.5.1 information bias

Information bias33

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 34: 3.5.1 information bias

Information bias

34

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 35: 3.5.1 information bias

Information bias35

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 36: 3.5.1 information bias

Information bias36

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 37: 3.5.1 information bias

Information bias

80

Se:P(class. positive|true positive), Sp: P(class. negative|true negative)

Page 38: 3.5.1 information bias

38

Information bias• Differential because sensitivity and specificity

NOT the same for cases and controls• This example resulted in bias away from the

null– True OR = 4– Misclassified OR = 5.7

• Can result in bias in either direction– Exhibit 4-6 in Szklo – differential misclassification

resulting in bias towards the null

Page 39: 3.5.1 information bias

39

Information bias• Information biases types summary

– Non-differential misclassification of exposure• Sensitivity and specificity of exposure

assessment not both 1.0 but the same for diseased and non-diseased

– Differential misclassification of exposure• Recall bias• Interviewer bias• Sensitivity and/or specificity of exposure

assessment NOT the same for diseased and non-diseased

Page 40: 3.5.1 information bias

40

Information bias• Information biases types summary

– Non-differential misclassification of outcome• Sensitivity and specificity not both 1.0 but the

same for exposed and unexposed– Differential misclassification of outcome

• Observer bias• Respondent bias• Sensitivity and/or specificity NOT the same for

exposed and unexposed