The number of people using dialysis - the prevalent dialysis population - had been going up as much as 6 and 7% a year in the late '90s and early '00s. I remember the articles predicting that by 2010 there would be over 700,000 people using dialysis in the US - here for example
http://www.nature.com/ki/journal/v67/n94s/full/4496031a.html In fact in 2010 there were about 420,000 people using dialysis in the US, that's quite a feat - 300,000 people who are not in the chair. It isn't entirely clear what happened, which is too bad. When the news isn't good it gets analyzed from every angle, but when the news is good it goes by almost entirely unremarked.
The prevalent dialysis patient population at the end of the year is the product of everyone who started the year dialyzing, less the number who die or are transplanted, plus the number who begin treatment. If fewer people died or were transplanted the number of people dialyzing would be even greater, just as if many more people transitioned into needing treatment the number in the chair would be greater. What has happened instead, it appears, is that fewer people died, but at the same time many fewer people have needed to start dialysis. Even with more people living with diabetes and hypertension fewer are transitioning into needing renal replacement. I don't think there is any one explanation but I think we should try to figure out what has been working and do more of it.