Lecture 7 (Bias + Confounding) Flashcards

1
Q

Bias (def)

A

Systematic errore in design/conduct of study –> mistaken asso
-Major prob in obsv epi studies

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

Selection bias (2 main reasons, major issue in ____)

A
  • Error resulting f/selection procedures or factors that influence participation
  • Major issue in case-control studies
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Info Bias (def)

A

-Error due to collection of incorrect info about subjects (misclassification)

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

Compensating for selection bias (case-control)

A

-Equalizing selection bias by selecting cases/control using same selection process

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

Minimizing selection bias

A
  • Carefully consider selection criteria

- Maximize participation rate

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

Selection bias and cohort studies (when is likely/unlikely)

A
  • Unlikely w/internal comp grps
  • External comp grps more prone (diff pops, might have diff risks aside f/exposure)
  • Healthy workers effect: issue when general pop = external comp grp, healthier
  • Loss to follow is an issue
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Nondiff exposure/disease misclassification (bias towards of away)

A

Biases asso towards Ho (no asso)

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

Diff misclassification (related to both exposure and disease)

A

-Biases towards/away Ho (depends)

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

Nondiff exposure missclass (cohort)

A
  • Biases RR toward Ho if exposure misscl is unrelated to the future development of disease
  • Not likely
  • Won’t bias if happens to same degree among those who develop/don’t
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Disease missclass (cohort)

A
  • Issue when based on self-report, esp for more subjective things
  • Concern if those collecting data are aware of subjects’ exposures
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Exposure missclass in case-controls (what’s the effect)

A
  • Important source of this bias
  • Bias towards null if missclass unrelated to disease stat (non diff)
  • Bias either direction if it missclass depends on disease status (diff)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Types of info bias

A
  • Recall bias
  • Reporting bias
  • Observer bias (sub: interviewer bias, abstractor bias)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Reducing bias

A
  • Carefully define things
  • Choose valid measurement methods
  • Train ppl
  • Quality control
  • Maximize participation
  • Apply methods in same way/detail to everyone
  • Don’t improve quality of data in one but not other
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Detection bias

A

=surveillance bias

-When close med surveillance –> higher prob of detection of disease

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

Most common confounder

A

Age

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

Difference btwn confounding and bias

A

confounding is real, bias is not

17
Q

Checking for confounding

A
  • Compare unadjusted and adjusted ORs (weighted average of stratum-specific ORs)
  • If difference is more than 10%: confounding
18
Q

Confounding (def)

A
  • Distortion of asso due to another extraneous exposure associated w/both disease and exposure
  • Cannot be intermediate variable in causal pathway though
19
Q

Positive confounding

A

-Overestimation of asso

20
Q

Negative confounding

A

-Underestimation of asso

21
Q

Qualitative confounding

A

-Inversion of direction of asso

22
Q

Prevention of confounding (design) (3)

A
  • Randomization
  • Restriction (single category of confounding exposure)
  • Matching
23
Q

Controlling for confounding in analyses

A

Stratification methods:

  • Direct adj (cohort)
  • Indirect (occupational retro cohort)
  • Mantel haenszel method (most common, case control/cohort)

Multivariate models:

  • Logistic reg (case-control)
  • Cox proportional hazards model (cohort)
  • Poisson regression (cohort)
24
Q

Limitations of controlling for confounding using stratification

A
  • Can only adjust for 1-2 w/small # of categories each
  • Or else data would be sparse
  • Only categorical variables
25
Q

Residual confounding (def and sources)

A
  • Confounding that remains after adjustment

- Sources: categories in adj too broad; variable was imperfect surrogate