Lecture 14: Association, Causal Inference, And Causality

Question 1

What is the relation between association and causation?

Answer 1

• Causal relationships are Complex phenomena
• Commonly, multiple-component (factor) issues
• Epidemiological studies may yield (statistical) ASSOCIATIONS (a.k.a., relationships; OR/RR/HR) between Exposure and Disease
• Yet, does the Exposure “CAUSE” the Disease? (Pssst…not often)

Question 2

What is the definition for causality?

For associations?

Answer 2

• An precursor event/condition/characteristic REQUIRED for the occurrence of the Disease
• Recall, Associations are relationships between an Exposure and an Outcome/Disease

Question 3

What are the three types of association?

Answer 3

3 Types of Associations (Relationships)

1. Artifactual (a.k.a.; False) Associations
2. Non-causal Associations
3. Causal Associations

Question 4

What causes artifactual associations?

Answer 4

Artifactual Associations can arise from Bias and/or Confounding

Question 5

How do non-causal associations occur?

Causal associations?

Answer 5

Non-causal Associations can occur in 2 different ways:
A. The Disease may cause the Exposure (rather than the Exposure causing the Disease)
- Example: RA leading to physical inactivity
B.The Disease and the Exposure are both associated with a third factor (Confounding)
- Example: The positive association shown between:
– Coffee drinking & CHD, or
– Down’s syndrome & Birth-order

• Causal associations are the ones they’re usually looking for:
• Exposure -> Outcome

Question 6

Name the 3 types of causal relationships

Answer 6

1. Sufficient Cause
2. Necessary Cause
3. Component Cause (Risk Factor)

Question 7

Describe Sufficient Cause

Answer 7

• A set of minimal conditions/events that inevitably produce disease
• A CAUSE which precedes a disease, and if present, the disease will ALWAYS occur
• • Quite rare; apart from genetic abnormalities
• • Sufficient causes can still have multiple, required ‘components’ (termed COMPONENT CAUSES; a.k.a. RISK FACTORS)) that collectively act to induce disease

Question 8

Describe Necessary Cause

Answer 8

• A Cause which Precedes a disease and has the following Relationship with it:
• Cause must be present for the disease to occur, yet the Cause may Also be present Without the disease occurring
• Example: Mycobacterium tuberculosis; a necessary cause for TB to be diagnosed, yet can be present in individuals without clinical symptoms of the disease

Question 9

Describe Component Cause (Risk Factor)

Answer 9

Component Cause (a.k.a. Risk Factor)

• Something that, if present/active, increases the PROBABILITY (OR LIKELIHOOD) of a particular disease
• Example: High LDL levels are RF for AMI
• Example: Smoking is a RF for lung cancer
• Some patients must be “primed” or “susceptible” to disease before Component Causes induce disease (multi-factorial)

Question 10

What are the two types of interactions in causal research?

Answer 10

• Synergism

- Parallelism

Question 11

Describe Syngergism

Suggested Very NTK

Answer 11

• (FACTORS WORK TOGETHER; BOTH)
• The biological-interaction of 2 or more component causes such that the combined measure of effect is greater than the sum of the individual effects
• Example: If gene- & environmental-factors acted together (in synergy), infants would only get the congenital disorder if exposed to BOTH factors

Question 12

Describe Parallelism

Also quite NTK

Answer 12

• (FACTORS WORK IN PARALLEL; ‘EITHER’)
• The biological-interaction of 2 or more component causes such that the measure of effect is greater if EITHER is present
• Example: Infants would only get the congenital disorder if exposed to either the gene- or environmental-factor but would not get the disorder if exposed to neither
• See Slide 10

Question 13

Describe Multiple Causation

Answer 13

• Multiple component-causes working in concert to collectively become sufficient causes
• Example: CHD

Question 14

What does inductively-oriented criteria

Answer 14

• Inductively-Oriented criteria are used:
• Hill’s Criteria (But really Guidelines)
• Derived following U.S. Surgeon General’s 1964 report on smoking
• “In what circumstances can we pass from this OBSERVED ASSOCIATION to a verdict of CAUSATION?”
• Hill disagreed that “hard-and-fast” rules of evidence could be generated by which to judge likelihood of causation

Question 15

What is the Causal Inference Process

NTK

Answer 15

• AN INTERPRETIVE APPLICATION PROCESS
• HILLS CRITERIA: (guidelines)
  1. Strength
  2. Consistency (Specificity)
  3. Temporality
  4. Biologic Gradient
  5. Plausibility (Coherence, Experiment, Analogy)

The higher the number of criteria met, when evaluating an ‘ASSOCIATION, the more likely it MAY be causal.

Question 16

Describe Strength in relation to association

Answer 16

• Strength refers to the size of the measure of association (RR/OR/HR)
• The greater the association the more convincing it is that the association might actually be causal
  Example: Smokers have up to a 20 times greater; risk of developing lung cancer compared to non-smokers
• A STRONG ASSOCIATION is Neither Necessary Nor Sufficient FOR causality and WEAKNESS OF AN ASSOCIATION is Neither Necessary Nor Sufficient for ABSENCE OF CAUSALITY.”

Question 17

Describe Consistency

Answer 17

• CONSISTENCY (a.k.a., REPRODUCTIBILITY) – the repeated observations of an association in different populations under different circumstances in different studies (not just once!)
• Example: Cause-effect relationship between cigarette smoking & CHD greatly strengthened by the fact that a large number of observational studies have consistently demonstrated an increased risk
• CONSISTENCY may still obscure the truth!

Question 18

What was MHT

Answer 18

Menopausal hormone therapy

• Claimed it did a bunch of stuff and cured a bunch of symptoms
• Through the 1990’s, data collection occurred
• NIH/NHLBI set forth to formally, officially evaluate claims with a national, large, long-term rigorous clinical trial
• Women’s Health Initiative (WHI) study
• Turns out it does only about half of this, and makes other things even worse.

Question 19

What is temporality?

Answer 19

• TEMPORALITY is the necessity that the cause Precede the effect/outcome in time
• Time-Order also describable:
• Proximate cause (short term interval)
• Distant cause (long term interval)
• See Slide 18

Question 20

What is the biological gradient?

Answer 20

• Biologic gradient – presence of a Gradient of Risk (DOSE-RESPONSE) associated with the degree of Exposure
• Example: Light smokers (<10 cigs. per day) are 5 times more likely to develop lung cancer than non-smokers, whereas heavy smokers (>20 cigs. per day) are 15 times more likely than non-smokers
• Caution: Some biologic factors demonstrate a threshold effect (i.e., no effect until a certain level of exposure is reached)
• lead exposure and mental retardation
• alcohol consumption and mortality

Question 21

Describe Plausability

Answer 21

• PLAUSIBILITY – presence of a biological feasibility to the association, which can be understood and explained (biologically/physiologically/medically)
• • Is the event/exposure BIOLOGICALLY PLAUSIBLE, if really true?
• At Issue: – Plausibility decision on criterion-based from existing/known beliefs, which may be flawed or incomplete
• • Example: Infection (H. pylori) causes PUD