Problem
Recall the causal criteria presented by Sir Austin Bradford Hill in 1965.
• Strength of association.?When the research problem is established, a research hypothesis is formulated about an expected association between variables. A study design is then selected, and an appropriate statistical test applied to the data. The statistical association is then deemed to be significant or not. In general, a strong statistical association between an exposure and health outcome provides greater evidence of there being a causal association because it is more likely to be real (valid).
• Consistency of association.?This occurs when associations are replicated by different investigators, in different settings, with different methods.
• Specificity.?Specificity of association means an exposure is associated with only one disease or the disease is associated with only one exposure.
• Temporality.?In order for an exposure to cause a disease, the exposure must precede the disease.
• Biologic gradient.?An increasing amount of exposure increases the risk of disease.
• Biological plausibility.?Is the association biologically supported? Biological assessment often involves experiments in controlled laboratory environments.
• Coherence.?Causal inference is consistent with known epidemiologic patterns of disease.
• Analogy.?Analogous situations with previously demonstrated causal associations provide support of there being a causal association.
• Experimental evidence.?The randomized, double-blind experimental study design is the best for establishing cause-effect relationships. This is because randomization is effective in balancing out the effect of known and unknown confounders and blinding is effective at controlling for bias.
Select a disease or condition and identify the potential exposure and the outcome.
1. Discuss the causal criteria that apply to your selected outcome and exposure.
2. Compare a direct causal association with an indirect causal association. Use specific examples.
3. Define and compare the difference between statistical inference and causal inference.