Trials need to be reported correctly: it’s important that researchers report the measurements they said they would.
Before a clinical trial starts, researchers are required to specify exactly what measurements they’re planning to take, in a ‘trial protocol’ or ‘registry entry’.
When the trial has finished, the researchers publish their measurements. Sometimes they don’t publish the measurements they said they would. This can lead to a biased and exaggerated picture of the benefits – and risks – of a treatment. It’s called ‘outcome switching’, and both trialists and journal editors have committed to address it, through guidelines like CONSORT.
With COMPare Trials, we’re holding researchers and journals accountable. Initially we’re focusing on outcome switching: but this is just the beginning.