I did a similar calculation but came back more optimistically (although still on scale of a year to two).
You only need count over 18s as very few cases in this group. It is believed by the UK experts to have a lower fatality rate of 1% and likely lower admission rate due to asymptomatic infections (estimates at 16-75%). Ventilator capacity can increase with stopping surgery and CPAP/other measures can be used in some cases. Treatments developed later should improve efficiency of treatment. Extra ventilators should be manufactured and help ease burden. I think 80% is too high, likely more 60% given an R0 of 2-3. Finally, the chronically unwel elderly are unlikely to benefit from ITU care significantly and would never qualify. With these exceptions it gets down to 1-2 years depending on assumptions.
I'd like to see more discussion around this point. I don't care where you got the numbers from. It's been clear from a glance that 3 weeks of cancelled school isn't going to do anything. So now we're looking at 8 weeks. Umm.... According to what? Which is more likely, 8 weeks or 2 years? I don't know, but I'm guessing 8 weeks isn't nearly long enough.
People are talking about it, that talk is driving policy, and the answer is “with some local adjustments possible based on caseload, until there is an effective vaccine developed and deployed, which will probably be 18+ months”, and the alternative is a truly catastrophic number of fatalities.
I ran the numbers for Canada, and it looks like we'd need isolation measures for 2+ years.
I don't see how that's a feasible solution long-term.