Association Causation

1
Association Causation

• Marshall Tulloch-Reid
• Epidemiology Research Unit
• Tropical Medicine Research Institute

2
Objectives

1. Differentiate Association from Causation
2. Understand the causal pie model for disease
   causation
3. List and discuss some of the criteria that have
   been used for deciding between association and
   causation Henle Koch Postulates the
   Bradford Hill Criteria
4. Understand the limitations of these criteria

3
Association Causation

• Variables are associated if a statistical
  relationship exists between them
• Statistical association does not imply a causal
  relationship between the variables

4
Association Causation
Population Hat Wearing Obesity
A 20 20
B 40 40
C 60 60
5
Association Causation
r 0.5
6
Association Causation Hats and Obesity
Percent Hats
Percent Obese
Group A B C
20 40 60
20 40 60
7
Concepts in Causation

• The trigger of an event is often thought of as
  its cause
• Most events are the result of the interaction of
  several causes
• These causes do not need to occur simultaneously

8
Concepts in Causation

• A 88 year old women with uncontrolled type 2
  diabetes and poor vision from diabetic
  retinopathy and glaucoma falls and fractures her
  hip.
• What caused the hip fracture?

9
Concepts in CausationWhat caused the hip
fracture?

• The fall
• Poor vision
• Hypoglycemia
• Neuropathy
• Non-compliance with treatment
• Age

• Osteoporosis
• Genetic factors
• No non-skid rug
• Not wearing a hip protective device
• The dog

10
The Causal Pie
Single Component Cause
Sufficient Cause
11
The Causal Pie
SF II
SF I
SF III
SF sufficient cause
12
The Causal Pie

• Sufficient cause
• Component Cause
• Necessary Cause

13
Causality and the Scientific Method

• Inductive method
• Observations induce in the mind of the observer a
  suggestion for a more general statement
• David Hume inductive method depends on circular
  reasoning

14
Causality and the Scientific Method

• Refutation
• Karl Popper statements about nature can be
  corroborated by evidence but it does not amount
  to logical proof however statements can be
  refuted by deductive logic

15
Criteria for Causation

• A checklist is often employed to resolve issues
  of causation vs. association
• Koch and Henle proposed criteria for infectious
  agents causing disease
• The Bradford-Hill criteria is most often used for
  this purpose and can be more generally applied

16
Criteria for Causation

• First questions to be answered
• Are the findings due to chance?
• Are the findings due to bias?
• Are the findings due to confounding?

17
Criteria for Causation Henle-Koch Postulates

• The organism is always found with the disease
• The organism is not found in any other disease
• The organism, isolated from one who has the
  disease, and cultured through several
  generations, produces the disease (in
  experimental animals)

18
Criteria for CausationBradford Hill Criteria

• Temporal relationship
• Strength of the association
• Dose-response relationship
• Replication of findings
• Biologic plausibility

• Alternative explanations
• Cessation of exposure
• Specificity of the association
• Consistency with other knowledge

19
Criteria for Causation

• Not all the criteria have to be meet for
  causality to be implied
• Some criteria carry more weight than others