and we -- this goes partly to joe antos' talk, we thought that accounted for 29% to 43% of the change in growth. and we thought that interacted with technology, again, to go to joe antos' point, and that interaction was another 27% to 49%. the insurance changes mattered less. i do think there's some in our data. we're looking now across the oacd and the demographic changes matter just a little bit. so, it -- and now chronic diseases, to go to ken's point, are also increasing, i think, throughout the developed world. certainly diabetes is increasing throughout. so that in, i think, going forward, that's also going to be driving cost, as he said. but again, i think all of these things are different aspects of the same reality. and i do think there is a mechanism that's driving the growth rates. >> i think dallas salisbury had a hand up. >> just a quick follow-on that set of points is, in work that paul friedsten's been doing with the database that we now have had seven years of data pre and post-design change by a major employer. this goes both to joe's point that it makes sense, and jo