Risk Factors and Causation

#Risk describes probability

A risk factor is a characteristic or exposure linked to a greater chance of an outcome. It may be changeable, such as a particular exposure, or not readily changeable, such as age or inherited variation. A risk factor does not guarantee that a disease will develop.

Disease often results from several influences acting together over time. Some people with a recognised risk factor never develop the associated condition, while some without it do. Reducing a causal risk factor can lower probability without removing every possible pathway to disease or guaranteeing protection for an individual.

#Association is not proof of cause

An association means that two things occur together in a pattern. One may cause the other, but there are other explanations. A third factor can influence both, creating confounding. Reverse causation is also possible: an early, unrecognised disease may change behaviour or measurements rather than result from them.

Researchers assess causation using multiple lines of evidence, including whether exposure came first, whether findings repeat, and whether proposed mechanisms fit established biology. Well-designed experiments can help, but they are not always practical or ethical. Observational studies remain important, especially when different methods support the same conclusion.

#Reading risk claims in context

Relative risk compares probabilities between groups, while absolute risk describes the probability within a defined group and period. A large relative change can correspond to a small absolute difference when an outcome is uncommon. Both perspectives matter, along with the population studied and the uncertainty around the estimate.

Population findings do not predict one person's future with certainty. Their relevance depends on factors such as age, exposures, background health and the quality of the evidence. A useful risk claim identifies the outcome, comparison and time period, while separating established causal effects from associations that remain uncertain.

#Common misunderstandings

A risk factor is not a prediction of what will happen to one person. People with several risk factors may never develop a condition, while people with no recognised risk factors may still develop it. “No known risk factors” does not mean “no risk”.

A cause also does not have to act alone. Many health outcomes arise through several interacting influences, including biology, environment and circumstances. A factor can contribute to an outcome without being either necessary or sufficient to produce it.

Another misunderstanding is that every risk factor is a useful target for prevention. Some factors, such as age or family history, cannot be changed. Others may signal an underlying process rather than drive it, so changing the measured factor may not change the outcome.

Finally, identifying a contributing behaviour does not establish personal blame. Choices are shaped by resources, living conditions, access to care and other constraints.

#Questions worth asking a clinician

• How does this risk factor change my absolute risk, compared with someone without it, and over what time period?
• Does the reported relative risk represent a large or small absolute change in the chance of this outcome?
• Could other factors explain the association, or could the outcome itself cause the apparent risk factor?
• Which reliable lines of evidence agree that this factor helps cause the outcome rather than merely being associated with it?
• Is there evidence that changing this risk factor lowers the likelihood of the outcome, and how certain is that evidence?