The article in the OP of course refers to a model... likely complex and likely one of many, many models.
Here is an article on the CIDRAP site:
Modeling study suggests 18 months of COVID-19 social distancing, much disruption | CIDRAP
I have not looked at any models, but I am pretty sure there are many. And they likely take into account not just death totals and %rates, but also look at the things like elements of the capacity of healthcare delivery systems and projected costs.
Public health officials that work in governments likely include infectious disease and epidemiology experts on the medical side, and the science side. Healthcare and epidemiology economists, regulatory folks, ethics folks and lawyers.
All very complex.
On both the healthcare side and on the economic side.
And of course the virus itself might be able to change in ways we would all find helpful, or in ways which make things tougher.
Again, there are likely many models-- that article refers to one of them, and seems to report on the results of some input variables, some of which are likely assumptions and some of which are likely informed by learnings achieved and shared thus far.
On the Rx side, when we have studies of, for example medicines that demonstrate high efficacy in pounding down the morbidity and mortality and that also have sufficient safety, we may get some "woo hoo!" moments.
A lower level of illness, a lower duration of illness requiring care in hospital and hospital-like settings (especially ICU settings), and a lower risk of mortality from such RX, from a sufficiently robust study or family of studies... then you plug those into models and voila you may have lower duration requirements for things like social distancing.
That is presuming the medicine(s) are abundantly available and at costs that make the models fly.
Then as testing (collecting and analyzing samples) has higher and higher throughput and availability resources can be more judiciously applied.
Of course all of these things might lay over geographies in ways we may need to deal with, so I kind of think that just like the implementation of measures currently taken has been somewhat regional, relaxation of the measures might be as well.