ch 10

Question 1

What should we do before we make a judgment about causation and generalization?

Answer 1

Determine if associations are valid

Question 2

internal valdity v generalizability

Answer 2

• Internal validity - do observed results accurately reflect the true association?
• Generalizability (external validity) - to whom can results be applied?
  Requires internal validity

Question 3

What are two key components of validity that are needed?

Answer 3

1. If a study lacks internal validity, external validity is irrelevant
2. We do not compromise internal validity to achieve external validity (generalizability)

Question 4

2 types of error

Answer 4

• Random error - error is due to chance
• Systematic error - error due to recognizable sources

Question 5

Three types of systematic error?

Answer 5

selection bias
information bias
Confounding

Question 6

What is bias?

Answer 6

• systematic error in the design or conduct of a study that leads to an incorrect estimate of the association

Question 7

How can bias occur? which studies? not inherent of? rare?

Answer 7

• Can be caused by the investigator or study participants during the design or conduct of the study
• experimental, cohort, case-control, cross-sectional, and ecological studies
• study population
• few have no bias or error

Question 8

what are the two effects of bias?

Answer 8

1. Creates the appearance of an association when there is none or masks one that exists
2. Selection bias (SB) and information bias cannot be fixed; confounding can be fixed to a point.

Question 9

What are some solutions to reduce bias? Effects what?

Answer 9

• Limit in study design.
• Limit in the study cohort (during data collection)
• Critically evaluate after the study has been completed and discuss its effects
  = effects on :
  Type and sources
  Direction
  Magnitude
  Impact on study results -> Impact on study interpretation

Question 10

What are the two types of bias?

Answer 10

selection bias (who is in the 2x2 table)
information bias (where in 2x2 table)

Question 11

What is selection bias? why does this happen?

Answer 11

• occurs when the study population is not a random selection of the target population.
• Results from procedures used to select subjects for a study
  At the time of recruitment
  During the process of retaining subjects in a study

Question 12

Which study designs does selection bias occur in? Why does it happen in each?

Answer 12

• case-control: selection/participation of cases and controls is related to exposure status
• cohort/experimental: selection/participation of exposed and unexposed subjects is related to disease status

Question 13

When is SB more likely to occur? why?

Answer 13

Case-control and retrospective because D and E have already occurred by the time the subject is selected for a study

Question 14

What are the multiple ways SB can happen?

Answer 14

• Control selection bias
• differential participation
• Loss to follow-up
• Refusal
• Nonresponse
• self-selection bias
• Healthy worker effect
• Different surveillance, diagnosis, or referral of subjects according to their E and D status

Question 15

What is control selection bias? Happens when? prevented by? Which study uses this?

Answer 15

• a bias that occurs if controls are more or less likely to be selected if they are exposed (or unexposed)
• This happens when controls fail to represent the exposure distribution in the source population from which the cases arose because controls do not accurately represent the same population as the case.
• Prevented by using identical selection criteria for cases and controls
• case-control

Question 16

What is Differential participation? prevented? When does different participation happen?

Answer 16

• ( case-control and cohort)
• Bias that can occur if willingness or ability to participate is related to both exposure and disease status
• Prevented by obtaining high participation rates for all groups

Question 17

what is differntail loss t follwo up? Issue because? Prevented by?

Answer 17

• (cohort and experimental)
• The bias that occurs if study participants exit the study for reasons related to both exposure and disease
• reduces the power of the study and doesn’t know if they develop disease or not
• since the outcome cannot be known without good follow up, must maintain high participation rate (Maximmzin retention and trace study subjects)

Question 18

What is self-selection bias? prevention method

Answer 18

• When participants voluntarily choose t participate in a study or program group rather than being randomly selected into one
• Conducting a community case-control study or limited sever cases

Question 19

How can we reduce selection and information bias?

Answer 19

• Little (or nothing) can be done to fix selection bias once it has occurred
• Must be avoided through careful study design and conduct (see strategies for individual biases)
• can not be “controlled” in the analysis

Question 20

What are the two directions of bias? What do the positive associations do?

Answer 20

• Toward null
  Positive association is biased towards H0, and true association is underestimated.
• Away from null
  Positive association is biased away from null, and true association is overestimated.

Question 21

What is information bias?

Answer 21

as a bias that occurs when the information collected for a study is inconsistent or flawed, potentially leading to inaccuracies in the study’s results
- Occurs after participants have entered your study

Question 22

Which two cases do information bias get into?

Answer 22

1. Case-control study - different techniques are used to collect information from cases and controls
2. Cohort study - different procedures are used to collect information from exposed and unexposed groups

Question 23

Which study does information bias occur more likely in?

Answer 23

case-control and retrospective because exposure and outcomes (retrospective) have already occurred when the subject is selected for the study

Question 24

What are the types of information bias?

Answer 24

1. recall bias
2. interviewer bias
3. measurement error or misclassification

Question 25

what is recall bias?

Answer 25

- Bias occurs if people with the disease remember/report their exposure differently than people without the disease.

Question 26

Which two studies is recall bias seen in?

Answer 26

- Case-control - cases are more or less likely to recall prior exposures than controls - The number of cases and controls do not change. - Retrospective cohort - exposed participants are more or less likely to recall prior diseases than unexposed participants.

Question 27

What is the solution for recall bias?

Answer 27

1. Use sick controls 2. standardized, close-ended questionnaires 3. Use pre-existing data (medical records)

Question 28

which direction is information bias go towards for null hypothesis?

Answer 28

Depends on which subjects have less accurate recall

Question 29

What is interviewer bias?

Answer 29

- Bias occurs if there is a systematic difference in soliciting, recording, or interpreting information. - Toward or away from null

Question 30

Which study designs can interviewer bias be seen in?

Answer 30

1. Case-control - the interviewer is influenced by the participants' case or control status (the interviewer is aware of the disease status of the subject and questions differently about E) 2. Cohort and experimental - participants influence interviewer treatment or exposure status (interviewer knows about E or Treatment status)

Question 31

What are the solutions for interviewer bias?

Answer 31

- masking but is impossible because the information they learn while talking with subjects. - Train interviewers - standardized questionnaires - good data

Question 32

what is a misclassification error?

Answer 32

- Toward or away from null - Bias that can occur if study participants are placed into the wrong exposure or disease category (error in classifying )

Question 33

What are the sources for misclassification errors?

Answer 33

- self-report s - Errors in medical records - Errors in how data are captured - Non-specific disease or exposure definitions

Question 34

which study designs can misclassfication error be seen in?

Answer 34

Case-control and retrospective studies: Relevant exposure may have occurred way before the study started and was hard for subjects to remember - poor recall

Question 35

what are the two types of misclassification errors?

Answer 35

1. non-differntial 2. differential

Question 36

What is non-differential bias?

Answer 36

- bias towards the null - That the extent of the misclassification is the same for both groups - Classification on one axis (E or D) is not related to the other

Question 37

what happens when the expsoure or disease is misclassified?

Answer 37

- E: the exposure misclassification is the same for cases/controls or diseases/not diseased - the extent of the disease misclassification is the same for exposed and unexposed persons

Question 38

what is differential bias?

Answer 38

- Bias towards or away from null - Classification of one axis (E or D) is related to the other axis (E or D)

Question 39

what is the solution for nondifferential and differential bias?

Answer 39

Accurate definition of E and D Accurate data sources Multiple measurements validation