I would add to that something we're not doing so well. Because of the limitations of the evidence that comes to us when we're trying to decide about approving a drug, probably about 60% or 70% of the time now we're saying that we need to follow this up.
Where we lack evidence is in real practice. Has that benefit that has been suggested by the evidence actually played out in practice? There are worries that in some cases, either because it's utilized beyond what we initially proposed or for other reasons, the benefit may not have played out so well.
I think in any ongoing forward system, it's really important that we have some post-marketing follow-up on what is actually happening in terms of use and cost.