I was kinda measuring. My estimate was that given random flips in quantum states (many), it would be easiest (take order of magnitude fewer quantum flips) to change Job's decision, somewhat harder to change the type of his cancer or toggle whether he had cancer (for example, by changing the stochastic cell changes that led to his cancer) and require such completely different universes to change his type of cancer that it isn't even clear whether he could be considered Steve Jobs anymore.
Another way of looking at it is with subsets of universes. Consider the subset of universes where Steve Jobs was the CEO of Apple. In some of those universes, he made a different decision about his cancer. In a smaller subset, he had a less invasive cancer. Finally, the mass of universes where he had a different cancer is well outside him being the Steve Jobs we know of, since that would require a significantly different life history or different genetics profile. On the other hand, Steve Jobs indubitably had some other potential cancers that could have fired up, so the set isn't entirely empty.
So you meant something like "counterfactually, it's more likely that Steve Jobs changed his mind about treatment than it is that he didn't get pancreatic cancer." Because the evidence that could support this claim is mostly macroscopic statistical evidence, I feel like phrasing this in terms of quantum mechanics ("the amplitude of outcomes where steve jobs changed his mind etc. etc.) actually makes it more difficult to support.
From Quora: