Let's say, hypothetically, 5000 people have astrocytoma and are being treated. Out of that, 1000 haven't responded to treatment, so they seek alternatives to conventional methods. Now, let's say spontaneous remission (it does happen) is 1%, so out of the original 1000, ten were "treated" successfully. They tell their family members "Hey, XYZ from Dr. dumbshit really worked".
And the other 990? Well, they had grave odds anyway, but they would be like the other ten had they come sooner! This is the basis of crock-as-shit alternative methods with no causal evidence.