If the therapy had not worked would that be evidence that the therapy is not effective? If so the fact that it did work must be at least some evidence that the

If the therapy had not worked would that be evidence that the therapy is not effective? If so the fact that it did work must be at least some evidence that the therapy is effective. Simple Bayes factor, you can't have one without the other. Additionally, if there is enough similar cancers we can know the rate of spontaneous remission within a particular band of time. This would be much lower than spontaneous remission at any time so that a spontaneous remission at the specific time of treatment could be a rather unlikely event and start to look like decent evidence. Not disagreeing that cancer treatment isn't a shit show but I would suggest that such evidence cannot be easily dismissed due to the lack of the gold standard large scale double blind study.