How Science Works (4): Epidemiology and Causality

In the previous post in this series, I discussed the difference between association and causation. Now I want to point out the role of epidemiology in establishing causality.

Epidemiology is the study of how disease is distributed in a population and of the factors that influence or determine that distribution. Epidemiological studies can help identify the causes and risk factors of a disease in a community. Causality is often inferred from epidemiological studies by using the following criteria:

• Strength of association: The greater the difference in rates between the treatment and the control groups, the more likely there could be a causal relationship.
• Consistency of association: The more studies that show similar results using different populations and differing study methods, the more likely there is a causal relationship.
• Dose response: When it can be shown that there is increasing risk for an adverse drug reaction with increasing dose, the more likely that there could be a causal relationship. If the drug—or a drug component—is the cause of an adverse event, removal of the drug or the component should decrease the occurrence of more adverse reactions.
• Cessation effects: Discontinuing a drug and having the adverse response go away suggests a causal effect.
  
• Statistical significance: The likelihood that the results of a study were due to chance is measured by the P value—the smaller the P value (0.05 or lower), the higher the statistical significance of the study, thus, the more likely the results were not found by a chance occurrence. See a more detailed explanation of statistical significance here.
  

Not all types of epidemiological studies carry the same weight in establishing causality. A rule of thumb to determine what type of studies are better for establishing causality would be something like this: Clinical trials, then case-control studies, then cohort studies, then ecological studies, and then case-series reports.