Large Scale Randomised Evidence
The effects of important causes of disease are sometimes so extreme that the cause-and-effect relationships can be reliably inferred from sufficiently large observational studies. Unfortunately, when assessing the treatment of some disease, there may well be only moderate improvements in outcome. Just a moderate survival improvement in a common disease might, however, save thousands of lives a year (and prevent much disability), so it is important not to get wrong answers. The best way to obtain reliable results about moderate treatment effects is by getting large-scale randomised evidence, as large numbers avoid being misled by the play of chance and proper randomisation avoids being misled by any biases.
|Meta-analyses of trials||Mega-trials|
Large Scale Observational Epidemiology
Previous observational epidemiologic studies have helped to identify a number of causative factors for the main chronic diseases of middle age and there is, perhaps, the perception that little more can be learnt from further such studies, particularly for established risk factors (such as smoking and blood lipids). But, the effects of such factors can vary enormously from one population to another, and there is still substantial uncertainty as to how important these are in different settings, and as to how their importance is changing with time.
|Meta-analyses of observational studies||Large observational studies|