I think the ultimate result from our current spike will be several mostly rural states with situations similar to Northeast's first spike. States with larger absolute ICU capacity (even if lower per capita) will skate by with using PACU and stepdown units for ICU capacity.
I have been thinking about clinical treatment and triage schemes, since that time is approaching us and I may be put into the COVID call (probably just lines at first, then ICU to relieve the intensivists). I thought it'd be worth sharing with you what my friends in critical care with COVID experience have been talking about:
One triage scheme that we think is the most just is:
Scenario 1: A new unstable (needs intubation) patient arrives in the ER, we will try to rapidly stabilize them but if we can't get to adequate oxygenation in ___ hours, we will have to withdraw care as futile, as that vent will be needed by someone who may have a better outcome.
Scenario 2: A new stable patient arrives in ER, we maintain them on high-flow or non-invasive ventilation. If they deteriorate within __ hours then they are intubated if ventilators and ICU capacity is available.
Scenario 3: A previously stable patient is deteriorating at the same time as a new unstable patient arrives. We probably should intubate the new patient as we don't know their trajectory (and they potentially may recover), but do know the previously stable patient's trajectory is worsening (and high risk of not recovering). The other alternative is using known clinical prognostic factors as a tie-breaker.
Scenario 4: Non-COVID patient (stroke, heart attack, trauma) comes in unstable. They would need to be triaged and probably assessed for survivability. If they can be stabilized and sent to the floor, that's the best even if standard care would be ICU. If they deteriorate rapidly, they'd fall into the above algorithm for recoverability.
This triage scheme doesn't a-priori deny treatment to people based on clinical factors, but does give them a shot to prove their stability. It does assume that failure at early rescue is predictive of death, but I don't know if that's necessarily true with COVID. What we can't afford is to have a bunch of people with anoxic brains taking up ventilators, and the risk of anoxia increases exponentially with time. UCSF has a mean time to stable ventilation of a few hours (compared to >24hrs) with aggressive ventilator adjustments, so this early decision scheme could be done. The problem is that there are a lot of physicians with some critical care experience, but not enough experience to manage ARDS or multi-organ failure. In NYC they had standard protocols that pretty much anyone could follow (until the person fell out of that criteria, then it's just crank everything up and wait until arrest). They did not have standard protocols for triage, though, since those hospitals are huge and had (barely) enough ICU surge capacity. Many rural hospitals face both an over-capacity situation and a high average acuity situation. They will not do well.
I never imagined we would, outside of a mass-casualty incident, ever have to think about these things in the US. But here we are, again, within the space of a year.