Reopening. Three weeks ago most US states began reopening. Since then newly diagnosed cases have stayed steady while deaths have decreased. But these metrics are lagging indicators. It takes at least a week before a person who’s been infected becomes ill enough to get tested, with test results obtained and reported a few days later. Death would come another week or more after testing, and a few more days would pass before the death is reported. So the impact of reopening should start manifesting itself about now.
There’s some chance that the first wave of reopenings won’t lead to big increases in infection. Recent studies indicate that most new infections can be traced to superspreading events, where one or more infected people join a large indoor gathering, in close quarters, for an extended duration. Nursing homes, meatpacking plants, prisons — these have been the main conduits for superspreading in the US. Most states limit crowd sizes, though last Sunday witnessed the reconvening of indoor church services. Looming on the horizon: schools, spectator sporting events, political conventions…
Mortality and Prevalence in the US. In a recent post I estimated that around 3 percent of Americans have been infected with corona, and that the mortality rate is around 1 percent. Recent high-quality studies suggest a revision: prevalence is likely greater — maybe as much as 6 percent — while the mortality rate may be lower — around 0.6 percent. Running some numbers on covid-related deaths stratified by age, I found that the mortality rate rises at a steady geometric rate for every additional year of age. National antibody studies in France and Spain showed mortality rates of around 1%, but their populations’ median age is about 4 years older than that of the US. The age-based adjustment would drop the US mortality rate to around 0.6%, which is what both the New York and the Indiana statewide surveys came up with.
Nationally there have been 316 corona deaths per million population. If death befalls 0.6% of those who have been infected, then each covid death represents 1/.006 = 167 infected cases. Multiply infections by death rate to estimate a national infection rate of 167 x 316 = 52,800 cases per million, or 5.8 percent. That’s 19 million cases nationwide, or ten times the number of confirmed diagnoses — a multiplier that both the New York and Indiana studies confirmed. Yesterday 25 thousand test-positive cases were reported in the US; the actual number is likely ten times that, or 250 thousand new infections.
Prevalence and Mortality in Underdeveloped Countries. Covid is killing a much higher proportion of younger people in Brazil, Mexico, and India than in the US. Almost certainly more younger people die from disease in places experiencing widespread poverty and overloaded healthcare systems, as this WaPo article argues. But wouldn’t dire societal conditions also increase the death rates of older covid patients? Another factor: these countries have a higher proportion of young deaths because they have younger populations. The median age of Brazil’s population is 32.6; Mexico’s, 28.3; India’s, 28.1 — all much lower than the US’s median age of 38.1. Applying the age-based adjustment, the US would have a covid mortality rate 60 percent higher than Brazil and 250 percent higher than Mexico and India.
Prevalence in Brazil, Mexico, India — none of these three countries has done much testing. Do they lack testing resources, or do relatively fewer of their infected populace experience symptoms severe enough to warrant testing? Mexico reports 657 test-positives per million population — about a tenth the US’s rate. But what if Mexico’s age-adjusted mortality rate is 0.25%*: then each death would represent 1/.025 = 400 infected people. Mexico’s death rate is 73 per million: multiply that by 400 and you get 29,200 infections per million, or 3 percent. That’s half the infection rate of the US, but with only a quarter the death rate.
*Analysis of US data indicates that each additional year of age increases the covid mortality rate by about 10%. E.g., the covid mortality rate for 55-year-olds is about 0.5%, while for 65-year-olds the mortality rate is about 0.5% times 1.1010, or 1.3%. So, if the median age in Mexico is 10 years younger than the US, and if the overall US mortality rate is 0.6%, then the mortality rate for Mexico would be 0.6% divided by 1.1010, or 0.23%.