How to read a health headline: trials, observational studies and what each can show

Health headlines often announce that something "causes" or "prevents" something else. The type of study behind the headline determines how far such a claim can be trusted.

A randomised controlled trial assigns participants by chance to receive a treatment or a comparison, so that other differences between the groups tend to balance out. It is the strongest design for showing that an intervention works, though it can be costly and not always possible.

An observational study follows people or looks at existing data without assigning anything. Epidemiology, the study of how disease and health are distributed in populations, relies heavily on this approach. Such studies are valuable, but they can show an association without proving cause, because people who differ in one way often differ in others.

Questions to ask: how many people were studied, for how long, compared with what, who funded it and has it been repeated?

Limits: no single study settles a question. Collections of studies, such as systematic reviews like those from Cochrane, are generally more reliable than any one paper.

This is a general guide, not medical advice.