ch 11 Flashcards

1
Q

confounding?

A
  • systematic difference between the groups that distort the true association between exposure and disease (want to control it)
  • bias toward or away from the null
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

what are the sources for confounding?

A
  • Experimental and cohort studies occur when the exposed and unexposed groups differ by more than just exposure—they differ on some other variable.
  • Case-control study: occurs when cases and controls have different characteristics
  • This can occur in all types of epidemiological studies
  • Unlike bias, it is an inherent characteristic of the population
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

what effect can confounding have ?

A
  • This results in a distortion of the true association between E and D and can bias either towards or away from the null.
  • Can be adjusted or fixed to a point
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

what can confounding be thought of as?

A
  • mixing of effects
  • The estimate of the effect of exposure on disease is distorted because it is mixed with the effect of other factors associated with E and D.
  • Sometimes referred to as the third variable problem
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

counterfactual ideal?

A
  • comparison group would be the exact same people who are exposed in the group had they not been exposed
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

how do epidemiologist select for people that are similar as possible with respect to?

A
  • With respect to other factors that could influence the outcome
  • With respect to the collection of comparable and accurate information
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How are confounder and counterfactual ideals related?

A
  • Confounding can be thought of as a failure to come close to the counterfactual ideal.
  • he difference in disease risk between the exposed and unexposed is because of factors other than just the exposure.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

When is a variable a confounder? Three criteria

A
  1. Independent predictor of the outcome - The confounder is a risk factor for disease among unexposed people
  2. Associated with the exposure
    - The confounder occurs more or less often among the exposed than unexposed
  3. It cannot be an intermediate on the causal pathway between exposure and disease
    - The confounder cannot be caused by exposure
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How do I know what might be a potential confounder for my research question?

A
  • Know your subject area
  • Complete a comprehensive literature review and read - Previous research will help you identify known confounders
  • Historical confounders - Some variables are always considered potential confounders (age, sex, race)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

mediator?

A

a variable that is a step in the causal chain
- is in between exposure and disease

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

what is in the design and analysis phase?

A
  • design: randomization, restriction, matching
  • analysis: standardization, stratified. multivariate
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is randomization?

A

Randomly allocates study subjects to treatment groups so each subject has an equal chance of being assigned to the treatment or comparison group.

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

when does randomization work?

A
  • study is large enough
  • The investigator does not influence the treatment assignment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What kind of characteristics should be similar in randomization?

A
  • baseline numbers should be similar, and the only different numbers should be what is the treatment
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

what are the strengths and limitations of randomization?

A
  • There is no limit on the number of confounders that can be controlled for
  • Do not need information about unknown confounders
    -You do not need to know what they are or measure them

limitation
- Limited to experimental studies
- Less efficient with a smaller sample size

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

what is restriction

A

Limits the study to people who are within one category of the confounder
- sex is a confounder then only men or women in study

17
Q

Strengths and limitations of restriction?

A
  • Simple, convenient, low-expense, and effectiveness
  • conceptually and practical
  • Effective control of characteristics being restricted
    limitation
  • Only possible for known, measured confounders
  • Incomplete control for confounding (residual confounding) if the restriction is not narrow enough
  • Cannot evaluate restricted variables
  • Limits sample size
  • Limits generalizability of results
18
Q

what is matching

A
  • selects study subjects so that confounders are distributed identically among exposure and unexposed (cohort study) or cases and controls (case-control study)
  • Used more often in case-control studies than in cohort studies
19
Q

two types of matching

A
  • Individual matching - identifies individual subjects for comparison that resemble a matched subject (Cases: 65-year-old male with lung cancer/ Control: 65-year-old male withouy lung cancer )
  • Frequency matching - balances the proportion of people with a confounding factor in the compared group (20% were 40-49 (exposed) makes sure that the unexposed who were 40-49 also make up 20% )
20
Q

What are the strengths and limitations of matching?

A
  • Simple and effective control of characteristics being matched
  • Good for when case series are very small
  • Useful for variables that are complex or difficult to capture (ex, neighborhood or occupation )

limitation
- Only possible for known measured confounders
- It can be difficult, expensive, and time-consuming to find appropriate responses.
- Cannot evaluate matched variables
- It can overmatch on a confounder - which is problematic because it makes the matched analysis inefficient.

21
Q

what is stratification?

A
  • stratify (separate) your study population into subgroups where one group has the confounder characteristic and one group does not, then calculate a measure of association for each subgroup
22
Q

strengths and limitations of startifcation?

A
  • Straightforward and easy to perform
  • Effective control of characteristics being separated
    limitation
  • It is difficult to control for many confounders simultaneously because of sparse data problems.
  • Difficult presentations, especially if many confounders
  • Continuous variables not easily stratified
23
Q

what are the five steps startfication?

A
  1. Calculate crude measures of association
  2. Divide subjects into strata of the confounder
  3. Calculate stratum-specific measures of association
  4. Calculate adjusted measures of association
  5. Determine whether the crude measure of association and by how much (magnitude)
24
Q

how can you determine if a confounding has occurred?

A
  • ompare a crude (unadjusted) measure of association to a measure that has been adjusted for confounding
25
Q

if crude adjusted or not RR=? confounding present or not?

A
  • if crude RR = adjusted -> no confounding is present
  • If crude RR does not equal = adjusted RR -> confounding is present
26
Q

How do we determine how much confounding is present?

A
  • use the ten percent rule
  • The variable is considered a confounder if the adjusted association measure changes by 10% or more compared to the unadjusted measure.
27
Q

positive confounding?

A

exaggerates the true association (is a bigger number)

28
Q

Negative confounding?

A

hides it (is a smaller number than the true association)

29
Q

What is a multivariate regression?

A

Involved construction of the statistical model (requires a computer) that describes between exposure, disease, and confounder

30
Q

advantages and disadvantages of multivariate regression?

A

Advantage: simultaneously adjusts for several variables
disadvantage: difficult to conceptualize, data need to fit into an available statistical model (assumptions needed)

31
Q

What is residual confounding?

A
  • confounding persists despite efforts to control or adjust for confounding.
  • Residual confounding should be acknowledged and addressed in the discussion section of a published paper.
32
Q

What are the sources of residual confounding?

A
  • Confounders for which no data were collected
  • Inaccurate data on a confounder
  • Use of broad categories of a confounder in your analysis