Governor Cuomo today discussed the release of the first results from a widespread, randomized antibody study of the state’s population. (CNBC Story) 3000 residents were tested at grocery stores and other locations so not exactly well-randomized, but we need to see a paper to understand the methods used. Across the state the rate was 14% and in New York City, 21%. Again, some people expressed alarm at how many asymptomatic carriers there must be, but to me, this is great news. Everyone of those positive tests were presumably people with asymptomatic or mild cases. So the rate of serious illness and deaths is far lower than we are told from reported infection tests. And the state is a good way toward natural slowing of transmission.
About 264,000 positive tests have been reported in New York. The population is around 19.4 million. The antibody testing results suggest that 2.7 million have antibodies. So the reported rates are missing 90% of the cases. Without knowing if they exclude people with positive infection test results, it appears that as many as 3 million New Yorkers have been or are infected. There have been around 19,450 deaths (New York is one of the places where too many deaths have likely been attributed to coronavirus.) That is a case fatality rate of .65. Not great but far lower than what the models projected. Because of the sampling methodology, the researchers may be under-sampling infected peoples.
And I think that the infections are front-loaded with the sickest, most susceptible people. One very rough indicator of this is that on the Worldometers coronavirus site, on the front page, it gives you the percent of active cases that are “critical or serious”. That percent has steadily dropped, from 5% to 4% to 3%. This indicates that in fact, the most susceptible of getting very sick were disproportionately infected early on and that the remaining population to be infected is not going to have the same rate of serious or critical illness and deaths.
And here is some more evidence on the true infection rate, straight from the CDC. (CDC Report) The CDC infection tested residents and staff at homeless shelters in Boston, San Francisco, Atlanta and Seattle. They found rates of 17% for staff and 17% for residents in Seattle; 36% of residents and 30% of staff in Boston; 66% of residents and 16% of staff in San Francisco and 4% of residents and 2% of staff in Atlanta. Most of these residents must have been asymptomatic. These are not the healthiest people. They may be more susceptible to infection, but they are asymptomatic or mild, it appears.
So what does this mean for Minnesota. If the rate of missed infections is 90%, the detection rate used in the model should be 10%, not 75% and the deaths in the model would be around 3000, not 22,000. Pretty pathetic modeling assumptions.