I'm pretty skeptical about a lot of the "curves" that people are throwing around because they're typically based on confirmed cases, our testing capacity keeps changing, and we still aren't testing everyone. Even if we were, we don't have an anti-bodies test that I'm aware of to identify people who had the virus but have already recovered. Without controlling for the availability of testing or throwing out data from before tests were widely available in the current format, the curves we are seeing all over the place can be misleading or can even paint a completely inaccurate picture if we mistake "newly confirmed cases" for new infections.

I've been very disappointed in how liberally some pretty alarming projections were circulated without an adequate discussion of the likelihood of the worst case outcomes or any discussion of the most likely outcomes with measures that were being proposed. I think the public health community as a whole felt it was better to raise awareness through worst case scenarios to compell compliance with social distancing than it was to provide information that was not misleading.

The raw numbers for deaths and hospitalizations are probably pretty accurate at this point, but you should be careful trusting most "trends" even with those numbers because there could have been (and almost certainly were) hospitalizations and deaths from this virus before testing was available.

Hospitalizations and deaths are a function of the total infections to some extent and so it seems pretty evident that infections are increasing, but a hospitalization or death today could mean an infection from a week or even a month ago. It's like we're driving a car using only the rearview mirror. We know where we were just on hospitalizations and deaths (we're guessing on infections) a week or two ago, but where we are and where we are going are blind spots in the absence of wider testing. Even testing an entire population in a small area like a neighborhood for present infections and anti-bodies would go a long way toward improving the crazy variance in the projections that keep getting thrown around by the WHO, the CDC, and academics.

The hospitals are really struggling in NY, but we still have basically no idea what the infection rate is and so it's tough to project whether the social distancing and "stay-at-home" measures will have any impact. For example, if the exposure rate is already 90% in NY with an infection rate of 45%, then new cases and hospitalizations are going to tail off in the coming days and then decrease substantially over the next two weeks and any distancing measures will have zero impact on that outcome. In spite of this, those who pushed for these social distancing measures will be quick to take credit for "flattening the curve" just in time to avoid a catastrophy that was never actually coming. That's assuming a very high exposure and infection rate with a hospitalization and mortality rate on the far low end of most projections, but, again, it's anyone's guess what the true infection rate is at this point and the trend is just not reliable.

The studies in epidemiology that will come out in the next 12-36 months will be really interesting. Fortunately, these viruses typically leave a stamp in the form of anti-bodies that build up in recovered individuals, so we can trace it later and find out what really happened. I think that will be an eye opening experience for all of us as history will judge these measures based on the outcomes from those studies.