Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

EPI PR3144 > Association vs Causation > Flashcards

Association vs Causation Flashcards

1
Q

Exposure -> Outcome considerations

A
  • validity of association observed

- if valid, is the observed association causal?

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

Assessment of validity of association between exposure & outcome (3)

A

Need to rule out 3 alternative explanations

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

Chance definition

A

Observed result may be due to chance

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

Relationship between

  • inadequate sample size
  • variability
  • chance
A

smaller the sample size -> larger the chance of variability -> higher probability of observed association is due to chance

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

p-value definition

A

Probability that the observed result or extreme result is by chance assuming Ho is true

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

95% CI

A
  • suggests statistical significance & precision

- affected by 3 factors (sample size, SD & confidence level)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Bias definition

A
  • any systematic error in the design, conduct or analysis of a study, resulting in a mistaken estimate of an exposure’s effect on the risk of an outcome
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

Can bias be fixed once it has been introduced?

Can confounding be fixed if present?

A

Bias cannot be fixed once it has been introduced.

Confounding can be accounted for by various method (in-study design & in-data analysis)

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

2 major types of bias

A
  1. Selection bias

2. Information bias

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Selection bias definition

A
  • groups in the study are not comparable due to the biased nature of the selection of groups (case & control, exposure & non-exposure)

**
eg cohort study investigating effects of health screening on overall health outcome
((exposure))
case : individuals voluntarily participate in annual health screening
control : individuals who do not participate in annual health screening

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

Examples of selection bias **

A

eg cohort study investigating effects of health screening on overall health outcome
((exposure))
case : individuals voluntarily participate in annual health screening
control : individuals who do not participate in annual health screening

eg case-control study on potential protection from use of NSAIDs against colorectal cancer
((outcome))
case : patients with colorectal cancer
control : patients without colorectal cancer
eg patients with arthritis / peptic ulcer

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

Reason for chance affecting the validity of observed association (4)

A
  • random sampling variability
  • inadequate sample size
  • p-value
  • 95% CI
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Relationship between

  • p-value
  • chance
  • Ho
A

smaller p-value -> lower probability that the observed association is due to chance -> stronger evidence against Ho

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

Why selection bias affect validity of the observed association?

A
  • results in an apparent association between exposure & outcome is observed even if in reality there is no association
  • reduce validity of the observed association
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Information bias definition

A 
Stems from incorrect determination of 
- exposure 
- outcome
- both 
, resulting in misclassification
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Examples of information bias (3)

A
  1. Recall bias
    - quality of recall
  2. Reporting bias
    - reluctant to report information accurately
  3. Poor sensitivity of test kits
17
Q

Types of misclassification (2)

& its effect

A
  1. Differential misclassification
    - misclassification rate differ in different study groups
    - lead to bias in both direction
    eg association observed but in reality there is no association & vice versa
  2. Non-differential misclassification
    - similar misclassification rate in the study groups
    - lead to results supporting null hypothesis & diluting association
18
Q

Confounding definition

A
  • a third variable that is associated with the exposure & can influence the outcome
19
Q

Factors of a confounder (3)

A
  1. Associated with the exposure
  2. Risk factor for the outcome
  3. Not an intermediary step in the causal process of exposure to outcome
20
Q

Methods to control confounders (2)

A
  1. In-study design

2. In-data analysis

21
Q

In-study design (2)

A
  • restriction of confounder
    eg pick participants without confounder
  • matching
    eg match participants from both groups with confounders
22
Q

In-data analysis (2)

A
  • stratified analysis
  • multivariable analysis
    eg multiple regression
23
Q

After controlling for confounders, how to assess for association? (2)

A
  • compare crude RR/OR & adjusted RR/OR

- interpret association based on adjusted RR/OR

24
Q

How to assess if the observed association is causal?

A

Bradford Hill Criteria

- do not need to have all 9 factors

25
Q

Bradford Hill Criteria

A

9 factors
ACCESS PTB

Analogy
Coherence
Consistency
Experiment 
Specificity
Strength 
Plausibility
Temporality 
Biological gradient
26
Q

Analogy

A
  • any similar exposure to support observed association?
27
Q

Coherence

A
  • is the relationship between exposure & outcome coherent with the general knowledge?
    eg natural history & biology of the outcome
28
Q

Consistency

A
  • is it consistent with other studies?
29
Q

Experiment

A
  • presence of experimental evidence to support the causal relationship
30
Q

Specificity of association (2)

A
  • weakest factor

- single exposure lead to a specific outcome

31
Q

Strength

A
  • strength of association
32
Q

Plausibility

A
  • biological system mechanism logical & possible?
33
Q

Temporality

A
  • factor occur first then outcome
34
Q

Biological gradient

A
  • is there a dose response relationship?

EPI PR3144 (24 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions