SUSSERS CAUSAL CRITERIA

1
SUSSERS CAUSAL CRITERIA

• Nigel Paneth

2
CAUSAL CRITERIA COMPARED
3

• 2. This means we must first make sure the study
  or studies we examine are sound, i.e. unbiased.
  It makes no sense to apply causal criteria to
  studies which are badly flawed. Causal criteria
  try to address the possibility of confounding
  more than they do bias. They assume that the
  data are in fact correct, i.e. unbiased.
•  
• 3. At the level of the individual study,
  consistency does not usually apply.
•  
• 4. Best use of causal criteria is to assess
  groups of studies. Meta-analysis usually
  summarizes odds ratios, thus assessing strength,
  and also formally excludes problematic (i.e.
  potentially biased) studies.

4
CAUSAL CRITERIA AS PER M W SUSSER (Am J Epid
1991133635-648)

• Three absolute requirements
• Association. The exposure and outcome are
  associated more commonly than would be expected
  by chance.
•   Time order. The exposure can be shown to
  precede the outcome
•   Direction. A change in the outcome is a
  consequence of change in exposure. (Not the same
  as directionality in a study).

5

• To best determine whether the above criteria are
  met, we then look at five additional criteria,
  defined slightly differently by Susser
• Strength (same meaning in Susser)
• Specificity
• Consistency
• Predictive performance
• Coherence

6
2. SPECIFICITY

• When we ask if relationships are specific, we
  can mean either
• a. Specificity of effect
• b. Specificity of cause

7
3. CONSISTENCY

• Susser defines consistency as "persistence of
  the association upon repeated testing"
• How is this persistence asssessed?
• a. survivability
•   The association persists even with the
• most rigorous study designs and analysis
•   b. reliability
•   The association persists in many diverse
  study settings

8

• 4. PREDICTIVE PERFORMANCE
• The association observed can predict a
  previously unknown observation

9
5. COHERENCE

• a. theoretical
• Compatible with pre-existing theory
• b. factual
• Compatible with pre-existing knowledge 1.
  biologic
• Compatible with current biological knowledge
  from other species or other levels of
  organization (e.g. cellular in humans)
• 2. statistical
• Compatible with a reasonable statistical model
  of the relationship of cause to effect
  (e.g. dose-response)