15. Causation in epidemiology: association and causation
- Causation: whether an exposure actually causes a specific disease
- or: whether the association between an exposure and a disease is causative
- Bias, confounding and chance must all be eliminated to establish a causal relationship
- Koch’s postulates for causation (of infectious diseases) states that the causative agents can:
- Be isolated from the affected organism in a culture
- Be cultured
- Produce the same disease in a second animal
- Be isolated from the second animal
- Epidemiological triangle
- Host factors, environmental factors and the causative agent all play a role in disease development
- Causative agents can be nutrients, poisons, microorganisms, radiation, trauma, etc.
- Host factors can be genetic susceptibility, immunological state, age, lifestyle, etc.
- Environmental factors can be crowding, housing, atmosphere, etc.
- Five rules (of causation) in epidemiology (similar to the Bradford-Hill criteria)
- There must be a high relative risk
- The results must be consistent
- There must be a graded response to a graded dose
- I.e., a higher dose increases risk
- There must be a temporal relationship, i.e. the disease must occur after the exposure
- There must be a plausible mechanism by which the exposure causes the disease
- The British Doctors’ study was historical, as it proved the causation between smoking and lung cancer