Temporal discontinuity trials and randomization: success rates versus design strength

Brian Knaeble, Erich Kummerfeld

We consider the following comparative effectiveness scenario. There are two treatments for a particular medical condition: a randomized experiment has demonstrated mediocre effecti…