Categories
COVID Archives

Daily COVID model update for Friday, April 10

North Carolina and Florida Peak

  • Current Prediction of Reported and Still Active Case Peak: April 10 or 11
  • Confidence: High
  • Model projection for tonight: 504,000

Key Stats:

  • Total Test Results reported yesterday: 180,746
  • National reported case Growth Rate yesterday: 7.9%
  • NY reported case Growth Rate yesterday: 7.1%
  • WA reported case Growth Rate yesterday: 4.8%
  • FL reported case Growth Rate yesterday: 5.9%
  • NC reported case Growth Rate yesterday: 6.6%

The data from the various sources were more congruous yesterday, so I’m feeling better about that. The national peak cases point is very close, and looks like it will happen tonight or tomorrow. The IHME model predicts national peak deaths on Sunday, so I’m surprised I’m calculating peak cases so close to peak deaths. I would have thought it would be much earlier, but I do not have good data on the average length of time from case reporting to death reporting. Just to reiterate, my assumptions are:

  • Mean time from symptoms onset to sample collection: 3 days
  • Mean time from sample collection to test reporting: 3 days
  • Mean time from symptoms onset to resolution (recovery or death): 13 days

If any of you can improve on this, let me know.

The big news (for some of us, anyway) is that it looks like Wednesday was the day NC experienced its peak active cases at 1,842, and Florida experienced its peak active cases at 8,500. NC has seen little COVID, so active cases may bumble around this number for a few days but this is good news. We’ll know how solid this for both states as we watch over the next couple of days. The IHME model projects peak deaths in NC next Wednesday, and FL in 13 days.

We hit a new low on national growth rate for cumulative reported cases yesterday at 7.9%, and the growth rate in active cases is close to zero (1.0%). It looks like my model is ahead of the actual data by about 3 days. I’ll continue to report model/actual results, and it’s premature to refit the model. However, I’ve changed the short-term case prediction above to reflect reality, as new cases have gone flat, and quite predictable for the moment.

Categories
COVID Archives

Daily COVID model update for Thursday, April 9

  • Current Prediction of Reported and Still Active Case Peak: April 8
  • Confidence: Medium, High Confidence between April 8 and April 11
  • Model projection for tonight: 456,000

Key Stats:

  • Total Test Results reported yesterday: 125,304
  • National reported case Growth Rate yesterday: 8.3%
  • NY reported case Growth Rate yesterday: 7.5%
  • WA reported case Growth Rate yesterday: 3.6%
  • FL reported case Growth Rate yesterday: 4.8%
  • NC reported case Growth Rate yesterday: 6.4%

Well, we may or may not have seen peak active cases in the U.S. yesterday. For the first time, the various sources of data are not correlating well. www.infection2020.com, my usual data source, reports 435,609 as of end of day yesterday. Wall St. Journal this afternoon reports 432,000 cases, citing Johns Hopkins. The CDC reports 427,460 as of 4pm yesterday, and the Covid Tracking Project reports 423,637 as of 4pm yesterday. I’m not happy with this, but will continue to use my primary source and hope this resolves itself. As a result, a little less discussion today until I see more underlying data credibility.

The IHME model is still projecting peak deaths on April 12, just 3 days from now, and total deaths at about 60,000 by August. The national growth rate (using www.infection2020.com data) is 8.3% yesterday, well into single digits. I’m keeping my eye on California to when active cases peak, and will introduce a California graph below.

Categories
COVID Archives

Daily COVID model update for Wednesday, April 8

NY peaked Monday

  • Current Prediction of Reported and Still Active Case Peak: April 8
  • Confidence: Medium, High Confidence between April 8 and April 11
  • Short Term Projection for tonight: 431,000

Key Stats:

  • Total Test Results reported yesterday: 137,367
  • National reported case Growth Rate yesterday: 9.2%
  • NY reported case Growth Rate yesterday: 6.3%
  • WA reported case Growth Rate yesterday: 5.0%
  • FL reported case Growth Rate yesterday: 10.7%
  • NC reported case Growth Rate yesterday: 12.2%

The good news today is that the state of New York Active cases seemed to peak on Monday (more below). It looks like California will peak very soon, but Florida still lags by a couple of weeks. I also noticed that the IHME model (from University of Washington) today revised total U.S. deaths from over 80,000 to 60,000. I’m still thinking the 60,000 number is too high based on what I’m seeing. Let’s hope so. That puts the total deaths in the range of a bad flu season (not good, but some perspective helps).

We’ve been down into single digit growth rates for several days now, so the disease is now spreading at a snail’s pace compared to 2 weeks ago. I’m looking for Peak Active Cases in the U.S. any day now. My model predicts anytime between yesterday and tomorrow (more below).

Check out the growth rates in NY and WA. Washington is over the active case peak, and it happened at something above a 5% growth rate. In both NY and WA we are on the downslope, and probably a few other states where I haven’t run the data.

Categories
COVID Archives

Daily COVID model update for Tuesday, April 7

Peak active cases today?

  • Current Prediction of Reported and Still Active Case Peak: April 7 or 8
  • Confidence: Medium
  • Short Term Projection for tonight: 403,000

Key Stats:

  • Total Test Results reported yesterday: 155,045
  • National reported case Growth Rate yesterday: 9.0%
  • NY reported case Growth Rate yesterday: 7.1%
  • WA reported case Growth Rate yesterday: 5.2%
  • FL reported case Growth Rate yesterday: 9.7%
  • NC reported case Growth Rate yesterday: 11.0%
  • Total currently hospitalized for COVID in U.S. N/A – data unreliable

I started doing my own modeling of COVID growth when I noticed that the news coverage didn’t bear much relationship with the reported cases. The numbers I was seeing pointed to a much smaller footprint than a lot of the sensationalized news. Now that we’re nearly over the peak (and yes we are – we are NOT in the infancy as some news reports still claim) I expect the news to gradually approach reality. In any event, I’m happy for the IHME model (funded by Bill Gates, btw). It’s in the public sphere, and anyone can look at it: https://covid19.healthdata.org/united-states-of-america You can easily look up your state to see when the model predicts peak deaths (some states have ALREADY peaked). In any event, I expect to send a daily report so long as I have something interesting to add, maybe until we’re clearly on the downslope.

As promised, I refit my Logistic model to historical cumulative reported case data today. I’ll continue to track against the updated model. The model currently projects that the cumulative total of all reported cases will steadily, and then slowly approach something less than 600,000 total in the April 23 timeframe. As I mentioned yesterday, there is no reasonable data I’ve been able to find on recoveries, but I’ve made the assumption that the mean time between symptom onset and resolution is 13 days, the mean time between symptom onset and sample collection 3 days, and the mean time from sample collection to test reporting 3 days. I’ll change any of these assumptions with more information, but using these, I believe we will reach peak active cases today or tomorrow.

Site powered by   

Foundation for Sane Public Policy

Visit John F. Groom on Amazon and Groom Media Membership

Publicty