If long term's just another 20 years, at the outside, that'll be enough. Short of lots more people getting sick and dying or living with chronic illness, untreated (so, not generating medical bills, and see also: civil unrest), I don't see what's going to slow it down in that time frame. I think effective legislation, which could theoretically do it, isn't likely for at least another 10 years or so if we're lucky. If it takes longer than that, well, I just hope the rioters blame the correct things/people for their problems. And hopefully the retirement crisis hasn't reached first-page-story status by that time yet, either, or it'll really be a mess. I suspect that one will hold off a bit longer, though. Pretty sure mid- to late-Gen-X folks will be the ones who make that one pick up the pace, especially if inheritance amounts have been trending significantly down for those same folks due to EOL healthcare bills for their parents.
So yeah, I agree it won't maintain a steady growth rate in the long term. I just don't think it'll stop before it causes some pretty big problems.
[EDIT] incidentally I'm not aware of high-profile projections of this stuff that have HC costs doing anything but continuing up way faster than inflation through at least 2030, provided the regulatory environment remains fairly similar. Even Warren's plan isn't projected to do much about that—need price controls in one form or another (like AFAIK every other OECD state uses) and a fix for our doctor supply problem, at least, probably, to do much about the rising costs.
Look at healthcare as a percentage of gdp over the last 50 years. Costs have been rising at a much slower rate over the last 10 years, and have dropped over the last 3. There's no reason to think the rate will stay constant over 20 years.
If anything naive extrapolating shows falling costs as a percentage of GDP, and sustainable rising real costs.
The only way we get crazy numbers is extrapolating using the average from 2000-2020, but there's no reason to do that.