>>What sort of 'exciting change' are you referring to ? I'd understand having to prove something was not harmful but having to prove it is beneficial ?

If you want to introduce (say) a new cardiac catheter, you need to show a) it is safe and b) that it carries enough benefit for practitioners to retool and retrain. Blogs and anecdote aren't enough, you need reproducible hard data or the change will be rejected by practitioners and regulatory authorities.

Even if you have come up with a really hot new product, if you've been promoting something else and practitioners have invested significantly in your offering, you need to subsidize their retraining and reinvestment if you want them to stay happy and/or enough backward compatibility needs to be designed in so that existing tools and methodology are still relevant. E.g. there are standards like AO so that ten-year-old tools are still useful on the very latest offerings- you don't have to start over just because the vendor wants to promote a new material that advantages itself.