19 JULY 2021 — Not everyone who contracts the virus has symptoms; not everyone with symptoms gets formally diagnosed. Consequently, covid test-positive diagnostic counts have always underestimated the number of infections. Death counts are more accurate, and so they’ve been used as lagging indicators of infections. Look up the recorded number of covid-related deaths that occurred during the past week, then divide the number of deaths by the covid fatality rate, calculated from random-sample serology surveys, in order to estimate the number of infections occurring 3 weeks ago that would have led to the number of deaths recorded this week.

For maybe 9 months in the US this calculation yielded a pretty stable ratio. The covid fatality rate for the US, based on a median age of 38 years, was around 0.65%, which meant that there were about 2.6 times as many estimated covid infections as diagnostic test-positive case counts. Then, at the beginning of 2021, people started getting vaccinated, and the ratio began to shift. Older Americans have disproportionately gotten the shots, so they’re now much less likely to get infected. Younger people are less likely to have been vaccinated and so are more likely to be the ones getting infected, but they’re also less likely to experience symptoms severe enough to get formally diagnosed or to die from the disease.

How much have the fatality rate and the infection rate shifted seven months into the vaccine phase of the pandemic? Time to look at the numbers.

On a daily basis the *Washington Post* updates the number of Americans who have been vaccinated, broken out by age. Multiply these age-specific vaccination rates by the US population totals, broken out by age category, in order to calculate the estimated median age of Americans who *have not yet been vaccinated* and who are therefore still vulnerable to infection. That number is 28 years — 10 years younger than the overall median age of the American population.

In prior work I analyzed a number of international serology surveys and the CDC’s mortality data to derive an age-adjustment algorithm for estimating the covid fatality rate; i.e., the average percentage of those infected with covid who die from the infection. For the US, with a median age of 38, the fatality rate stabilized at around 0.65 percent. For populations with a younger median age, the algorithm for computing the fatality rate was 0.65 divided by 1.1 to the Nth power, where N is the difference of the median population age from 38. As of now, the still-vulnerable US population is 10 years younger than 38; plugging that number into the algorithm, the estimated covid fatality rate among the unvaccinated Americans is around 0.26 percent.

Over the past week (11-18 July) 1,800 Americans died of covid. Divide that number by .0026 = 700K: that’s the estimated number of new covid infections three weeks ago (20-27 June). During that one-week interval the dx-positive case count was 91K. So the estimated current ratio of actual new infections to dx-positive case counts is around 700K/91K = 7.7. That’s *three times* what it was at the end of 2020, just before the first vaccinations were administered. Over the past week the new case count was 203K, or more than double the count of 3 weeks ago. 203K x 7.7 = 1.56 million new covid infections from 11 to 18 July in the US.

That sounds like a lot — is it? At the peak of the pandemic, around 4.5 million new infections were hitting America weekly. The current rate is about one-third of the peak. However, if the upward trend continues, we’re liable to reach the prior peak rate by the end of August. But it’s actually worse than that. About half of the American population has been vaccinated, so contagion is circulating among only the unvaccinated. In effect the new infection rate is twice the national number, since it’s concentrated in only the unvaccinated half of the population. At the current rate of increase, infection rates among still-vulnerable Americans will reach peak pandemic levels by the end of July.

How long can this go on? The new variants seem to be quite a bit more contagious than the earlier versions of the virus, so earlier estimates for reaching herd immunity at around 75 percent of the population are probably overly optimistic. Half of the American population has been vaccinated, while around one-third of the unvaccinated have already been covid-infected. So that’s 0.5 + (0.5 x .35) = around two-thirds of the American herd is immune. That leaves around 110 million Americans vulnerable to contagion. Lately the vaccination rate has plateaued at around 3.5 million per week, while an additional 1.5 million are being infected weekly. If both rates stabilize, that’s an additional 5 million acquired immunities per week. 110/5 = 22 weeks before the entire US population achieves immunity through either infection or vaccination. That takes us to the end of 2021. We’ll get there sooner if vaccination rates go up (not very likely) and if infection rates go up (quite likely).