This is an update to my post from March 21st,
http://zogprime.blogspot.com/2020/03/coronavirus-math-and-our-future.html
In the interest of not burying the lead, I’ll start by saying that we are still in a very bad place.
We are not doing enough to slow the rate of spread, and we are headed for a nightmare scenario.
If the United States can’t get the rate of new infections below 1% daily we will exceed a million deaths nationwide by January or February 2021.
The US now has over 1% of the total population with confirmed cases. The growth rate for new cases, or Daily Rate of Spread (DRS), is hovering around 2%. At 2% the time to double the number of cases is a fraction over 35 days. [1]
At that rate, we will hit roughly 5% of the population with confirmed cases by the end of September, 10% by early November, 20% by mid-December, and 40% by mid-late January. At that density of infection, the entire country will hit or exceed 100% of the available hospital capacity, with many areas over 200%.
https://projects.propublica.org/graphics/covid-hospitals
Phrased in the terms used in that article we will hit 20% within the first 9 months of the outbreak and 40% within the first 12 months.
When I published the original post the US was still under 25,000 confirmed cases, but the Daily Rate of Spread (DRS) in the US was over 40% and we were just days away from breaking 100,000 cases. I ran the numbers are various DRS to try to illustrate the severity of the situation. Any DRS in double digits would be disastrous. By early April the national DRS was below 10% and falling thanks largely to social distancing, quarantine, and mask orders. In spite of that, the US broke 1,000,000 confirmed cases in late April. By early May the DRS was below 2%, and in early June briefly dropped below 1%.
During the first week of June, many states (Particularly those with Trump loyalist governors) decided to declare victory and start re-opening. June 15th was the last day we saw a sub 1% DRS. Since then the DRS has slowly been climbing and broke through 2% on July 1st.
In March, starting under 25,000 cases had the US managed a DRS of 2% and a doubling rate of 35 days we would have had a realistic shot at containment. The US would still be under 250,000 cases with the 1,000,000 case mark over 4 months away and hospitals able to keep up.
From where we are now a DRS of 2% is a disaster. Right now the mortality rate when you compare deaths to new cases for the day is at an all-time low. July 10th saw 72,311 new confirmed cases and only 826 deaths or 1.1%. That is an artificially low number that reflects the case rate 2-3 weeks earlier, but as such, it can be considered a best-case scenario. If you combine a 40% infection rate with a 1.1% mortality rate and plug in a population of 330,000,000 you have 132 million people infected and 1.452 million deaths by mid to late January.
Let me restate that. At the current growth rate, and with the absolute, most optimistic, artificially deflated mortality rate I can come up with we lose almost a million and a half people by early next year. If you use the global mortality rate of 4% we are looking at over 5 million dead by January.
Most of the remaining news is also bad. The early spike in COVID-19 cases was centered around a few large outbreaks. Mostly in the New York/New Jersey area. When the northeastern US started to get its outbreak under control the numbers dropped rapidly. The current surge of cases is spread across the country with dozens of hotspots or potential hotspots. A concerted effort in Houston, Los Angeles, Miami, or Phoenix will not have the impact that the “all hands on deck” in New York had.
http://www.gregorybufithis.com/2020/07/03/the-trillion-dollar-question-why-are-covid-cases-increasing-in-the-us-while-deaths-are-decreasing-the-answer-is-simple-simpsons-paradox/
My conclusion remains largely unchanged from what I said in March…
Are we overreacting? No. If anything we are still underreacting. At the current DRS, even the current mortality rate will be devastating. As hospitals are stressed beyond capacity the mortality rate will only increase. If we don’t get out ahead of this and contain the spread millions will die. That’s not hype, that’s not alarmism, that’s math.
[1]
Doubling rate calculated using
log (2) / log (1+x) where x is the rate of increase.
E.g. for x = 2%:
log (2) / log (1+.02) = 35.003
Numbers on CDC website 7/12/20 ca 9:30 pm:
ReplyDeletehttps://www.cdc.gov/coronavirus/2019-ncov/cases-updates/cases-in-us.html
cases (confirmed and probable) 3,236,130
of these, new since yesterday 62,918
deaths (confirmed and probable) 134,572
of these, new since yesterday 906
But the actual number of cases is probably much higher:
https://www.cdc.gov/media/releases/2020/t0625-COVID-19-update.html
"But all in all, I think you’re in the right range, that somewhere between 5%, 6%, 7%, 8% of the American public has experienced infection, whether they recognized it or not. The estimate that we have given you at this point is it appears that the rate is — and this is CDC’s serology data — that the rate is approximately ten seropositive antibody individuals for every one case."
https://www.msn.com/en-us/health/medical/2-charts-estimate-the-true-scope-of-the-uss-coronavirus-infections-and-deaths-the-portrait-is-grim/ar-BB16xP2M
"MIT researchers, similarly, calculated that by June 18, around 17.5 million people in the US had gotten COVID-19 — around eight times the official number on that date."
If they're right, then around 30 million people in the U.S have been infected, of whom around 150,000 have died or will die, a death rate of around one half a percent.
Projecting out the 2% DRS and a 0.5% death rate leads to everyone in the U.S. having been infected by about November 10, and about 1.6 million deaths.
Obviously the question is whether those rates will continue. What percent of us will manage to avoid infection? Will overwhelmed hospitals mean the death rate will rise? After enough especially susceptible people have died, will the death rate decline?
All valid follow up questions and all make it harder to project.
DeleteBack in May someone I know posted a link to a Slashdot article that referenced a CDC study. The Slashdot takeaway (or at least their key point) was the the mortality rate was more like 0.4%.
I read the referenced CDC report
https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html
and my takeaway was essentially what you said, that the CDC is saying that the morbidity rate is much higher than the confirmed infection rate (possibly as much as 10x) and that the mortality rate is lower.
About a month after I had that conversation the Washington Post came out with an article
https://www.washingtonpost.com/health/2020/06/25/coronavirus-cases-10-times-larger/
reaching the same conclusion.
At this point I don't think anyone (or at least anyone with a shred of credibility) is disputing the number of deaths reported from COVID-19 (almost 143,000 as of today). If anything that number is on the low side too.
Given that, from a math point of view your numbers are right on target. I think though (and I need to go back and find this) that in epidemiology there is an expectation that infection rates won't go to 100%. I'm forgetting where I saw this, and it may be tied to herd immunity which is in question as instances of re-infection rise, but I think that historically pandemics start to plateau in the 60-80% range.
That is scant comfort because at 80% we would still be over a million deaths at a 0.4% mortality rate and it is highly likely that rate would increase as medical infrastructure gets overwhelmed and collapses.
As bad as the numbers I posted are I think that if we do not change course drastically they represent the best case, not anywhere near the worst. I intentionally tried to stick with working off of confirmed case numbers in the hopes that people might understand how grim the situation is even without the unconfirmed but highly probably cases.
My concern with the CDC studies is not the validity, but that people will cherry pick out of context. I think that the people with their heads in the sand want to believe that the current confirmed case rate is the extent of the infection, but that the mortality rate is misreported.
The fact that the only way that works is to say that there have really only been 14,300 deaths and not 143,000 seems to be one of those little details that gets lost.
So take your pick. Either we are seeing a mortality rate of 3.7%, an infection pool of 3.8 million and a rate of spread around 2% or we are seeing a mortality rate of 0.4%, an infection pool of 38 million and a rate of spread that is harder to estimate but likely at least 2%.
Either way it's bad.
Thanks for doing the math - I really find the information helpful.
ReplyDelete