Hey AJ, sure, I can answer that. So on the first related inpatient, you are correct. When we cited some of the inpatient pressure that we were seeing earlier in the year, we attributed that to the fact that we had a lot of new enrollment growth and have limited information from which to predict the level of medical cost utilization. So risk scores is one example. And we were seeing that relative to our expectations, the medical costs were slightly higher than we would have expected based on what the [indiscernible] indicated. So you are correct about that. I would say what we saw in the fourth quarter is completely unrelated and different than that, and very much more widespread, and particularly for the months of November and December. And what we've seen so far is an increase in short stay inpatient authorizations in particular. They are being upheld through our utilization management processes at a higher rate than you would typically expect as well. And given that the positive seasonality you would typically see in the months of November and December, it was that much more surprising. The absolute level of authorizations was up relative to what you would have expected for all those dynamics. Coincidentally, and while the data is still very early, as you know, on non-inpatient we're reliant on the paid claims to get some visibility. We are seeing indications that starting in November, we are also seeing a decline in observation stays. So that's something we will continue to analyze and understand better. But quite frankly, we are going to need more time, more claim development to fully understand the underlying sort of admission, diagnosis codes and other things to fully assess the nature of the uptick in inpatients and corresponding decline in observations and understand if they're in any way related, and then how we think about that again on a go forward basis. In terms of retention mix, I would say a little bit too early for us to fully assess that and look at, we've looked at the AEP enrollment data, I would say from what the team has looked at so far, nothing of concern in terms of the retention mix versus those members who disenrolled. But generally I would say given the benefit reductions we did make in 2024, I would certainly expect individuals who have an intent to highly utilize those benefits were probably ones that were looking at what other options might be available. And to the degree another plan maintained a richer level of benefit, certainly would expect that that was someplace we would see higher disenrollment. To your question about 2024, so I would say it a little bit differently. While certainly this trend was not fully contemplated in our pricing, I think the entire industry would agree with that statement. All of us saw unexpected trend after we filed our bids. We, in spite of that though, did make more benefit adjustments than the industry. And so if you remember, we talked earlier in the year that when you look at the changes made, we did make a higher level of benefit adjustments than others and in some cases we were very surprised to see actual net incremental investment in 2024 despite the stars and v28 and other headwinds that they were dealing with. So we don't feel like this is an issue in terms of underpricing, it's just the occurrence of higher than expected trend late in the year, unfortunately, after benefits were filed, and again, a lot to learn in terms of the persistency of that not only through 2024, but 2025 that we will continue to evaluate as we do all of our 2025 pricing. But based on everything we know now, our intent would be to assume those trends persist and make sure it's covered in our 2025 pricing.