Association to Causation Sequence of Studies Clinical observations

Association to Causation

Sequence of Studies Clinical observations Available data Case-control studies Cohort studies Randomized trials

Types of Associations • Real • Spurious

DEFINITIONS • • • Observational study Causation Etiology Your Assignment: Define these terms Association as they apply to epidemiology. Necessary Sufficient

Koch and Causation • • Postulates Why study association? Web of causation Do we need a better way?

Twelve Criteria for Causation • Cause distributed at same level • Incidence much higher in exposed population • Exposure more frequent • Disease should follow exposure • Dose dependent • Expected response • Association should be the same • Other cause-effect ruled out • Control results in decreased disease • Modification of host results in decrease • Human vol. always + • Findings should make sense

Factors in Causation of Disease • • Predisposing Enabling Precipitating Reinforcing Your Assignment: Define these terms as they apply to epidemiology.

Web of Causation

Are Associations Always Connected to the Disease? NO, BUT ……………. . üCigarette smoking and lung cancer üAge and prostate cancer üCar accidents and alcohol üTribal customs and kuru üAgriculture and antibiotic resistance

Association • • Deals with ………………. Is concerned with ………. Is the degree of …………. Has to be scientifically proven …. . Your Assignment: fill in the blanks

Degrees of Association • No association Direct association üNo possibility for association üPossibly associated üAssociated üDirect cause and effect

Guidelines for Judging Whether An Association Is Causal • • Temporal relationship Strength of the association Dose-response relationship Replication of the findings Biologic plausibility Consideration of alternate explanations Specificity of the association Consistency with other knowledge

Deriving Causal Inferences • Arriving at causation from association ü For example, showing that Helicobacter pylori is directly linked to peptic ulcers

Perspectives • • • Causation may be limited Subject to modification Perhaps more complex than realized Sometimes not measureable “Criteria” are really guidelines ümay be subjective