Over the last couple of weeks, there have been several articles discussing COVID ‘off ramps’, that is, when and how will we decide to (start to) return to ‘normal.’ What I find depressing about most of these articles is that most seem to have not learned a damn thing over the last twenty months. We need an approach grounded, not in some pundit’s instincts–because that worked really well when too many of them were acting as if the pandemic was over in March and April 2021–but in what is an acceptable level of disease. And we need good metrics for that acceptable level so we can have clear guidelines.

For me, there are two metrics we should be using: the percentage of the total population that is vaccinated, and the prevalence of infections (i.e., how many people are infected). Let’s deal with vaccination first.

I’ve seen estimates of R₀ for Delta in the U.S. of around 5*, so to limit infection–not hospitalization or death–I don’t think we’ll even put a dent in transmission until we hit 85% of the entire population vaccinated. And to prevent hospitalization and death, we should have an even higher rate among those aged fifty and older. So I think we shouldn’t really consider lowering our guard until 85% of the entire population is fully vaccinated.

Onto the prevalence of infections. If we had a highly vaccinated population, and the prevalence was one new case per 100,000 people per day, that would be a good place (and some asshole with a blog Twitter feed routinely refers to that as the “good place”). In a city like D.C., population roughly 700,000, that would be roughly 50 new cases per week, or 2,500 new cases per year. In terms of severe effects, including long COVID, that would likely be fewer than 100 hospitalizations per year, a handful of deaths, and, at most, a couple hundred cases of long COVID (long-term or permanent disability). While that’s obviously not ideal, that’s bearable.

So if we were to set a prevalence level of 1-2 new cases per 100,000 people per day, that would be good. It also has the advantage of if (or maybe when) that threshold is exceeded, of things not getting too bad: a doubling would be 2-4 new cases per 100,000 people per day. The reality is, between the lag in observing infections, assessing what is happening, and then altering policy (and then having people actually living those changes), there’s likely a month long lag in reacting. Compared to the upper bound of the CDC ‘moderate’ zone, which is around seven new cases per 100,000 people per day, any spikes with the lower threshold will be smaller and cause far less damage.

While we can argue about the specific thresholds, rather than going with our guts, or having Thinky Thought Leaders declare they are tired of this and It Is Over (how has that worked so far, by the way…), let’s make some good thresholds and then stick to them.

For fucking once during this godawful mess.

*In a talk by Trevor Bedford I saw a few weeks ago, which used some pretty good methods, his group estimated that R₀ in the U.S. was slightly under 5.0. This was using a different approach than the SAGE group in the U.K. uses (as best as I could tell and, honestly, can remember with my not young brain).