1.15—Health risk data, correlation and causation

Describe what the data show before explaining it: identify the population, units, trend and comparison. A correlation means two variables change together; it does not by itself show that one causes the disease.

Confounders such as age, income, activity or access to healthcare can influence both the risk factor and the outcome. Relative risk, absolute risk and sample size answer different questions.

If disease prevalence rises with smoking exposure, state the size and direction of the association, then ask whether dose, timing, biological mechanism and alternative explanations support a causal interpretation.

Use confidence intervals or statistical tests where supplied, avoid extrapolating beyond the population studied, and distinguish an individual prediction from a population estimate.

A statistically significant association can still be biased; a non-significant result does not prove no effect. “Linked to” and “causes” are not interchangeable.