Positive findings are around twice as likely to be published as negative findings. This is a cancer at the core of evidence-based medicine.