Every number in the tools on this site rests on a few ideas from epidemiology. None of them needs mathematics. Read them once and a health headline becomes much easier to weigh.
Sunlight, alcohol and asbestos are all hazards: each can cause cancer. How likely that is for any one person depends on how much, for how long and at what age. The IARC Monographs identify hazards and leave the size of the risk to other studies.
A relative risk compares two groups. An absolute risk says how many people are affected. Doubling a risk of 1 in 10,000 adds one case among 10,000 people; a 10% rise in a risk of 1 in 10 adds a hundred among the same 10,000. Ask for both numbers.
People who carry a lighter develop more lung cancer, and the lighter is innocent: smoking explains both. Epidemiologists call this confounding. Before they call a link causal, they look for the same result across many studies, for more disease where there is more exposure, and for the exposure coming first.
A study might report a relative risk of 1.2 with a 95% confidence interval from 0.9 to 1.6. The data fit a small increase, and they fit no effect at all just as well, because the interval includes 1. A narrow interval from a large study deserves more weight than a wide one from a small study.
When IARC places an agent in Group 3, “not classifiable”, the studies were usually too few or too weak to decide either way. That describes the state of the evidence. Showing that something is safe takes large, careful studies that looked for harm and found none.
A risk factor shifts the odds across thousands of people. Most people with a given exposure never develop the disease, and some people without it do. For anything about your own health, the people treating you know what no population study contains.