Today, the city of New York released data for April 15, and the cumulative totals for the confirmed and probable Covid-19 fatalities = 11,477. Modeling indicates, that those fatalities represent that 23 days before, March 23, there were an estimated 1,712,317 folks infected. 1/

That total of infected people on March 23rd, represents 20.29% of New York City. It is a crossroads, as Gov. Cuomo said, because we have to stabilize the curve so as not to challenge the health care system. He proposed another month, to confirm that future compounded daily...2/

growth rates (CDGR) stays close to 1-2%. On, this thread I will keep track of CDGR of COVID-19’s terminal indicator. As of March 23 (again based on April 15 cumulative fatalities, and modeling) we see too, that the levels of fevers returned to levels before 3/

Clearly, today is as good a day to begin keeping track of the CDGR in NYC, the biggest driver of deaths in the state. Among those aged, 0-17, in NYC, 0.68% were infected as of March 23rd; among 18-44, 31.630%; among 45-64, 23.388%, among 65-74, 10.871%; ...4/

and, among those 75 and older, 13.752%. Fortunately, the city of New York has done reasonably well at limiting exposure among the oldest. Still, this coming month we can do better to protect elders, and vulnerable people older than 44 years old. I will track those CDGR’s. 5/

Clearly, the group of people who are least at risk for terminal outcomes, has had the fewest exposure so far. That is the group between 0-17, who represent 1.744 million people and only 11940 are estimated to have been infected - have had 3 fatalities.
6/

Today, NYC released the totals for confirmed and probable cumulative fatalities as of April 16, 2020 = 12,199. As noted yesterday, this thread will serve to track the compounded daily growth rate (CDGR) since April 15, 2020 going forward. That is 4.72% overall from yesterday.

The key variable of interest on this thread, will be the lagging *estimated* indicator of CDGR - for infections: two days ago that was estimated for March 23rd, which was the presumed baseline = 1,712,317 (20.29 of NYC and end of peak) for infections. 3/24/20, CDGR = 4.72% 8/

That is, the way to estimate growth in infections in NYC and figure out if the CDGR is too high; we want to be able to estimate the compounded growth rate from some peak. The arbitrary peak, is assumed for purposes of this thread to have ended on March 23rd, when the KINSA..9/

..fever data began dropping 23 days after it began. That period, from the last day of February till March 23rd produced an estimated growth of to 1,712,317. We certainly don’t want to see that kind of compounded growth again, as it caused chaos, in NYC’s HCS. 10/

The key point as Gov Cuomo has alluded to in his statewide pressers, we don’t want to see that in the days following March 23rd, estimated value 23 days of growth coming too close to an estimated 1,712,317 value again. Yesterday’s value of CDCG of 4.72% is a bit too high. 11/

Certainly a doubling from 1,712,317 is close to catastrophic in NYC. Because in this phase, a doubling over the next 23 days, starting with the estimated value of the 1.7 million requires a CDGR of only 3% and we the chaos caused by that much growth. We’d like to add way less.12/

The value for March 24 (day 1 of the 23 day window) the calculation done for April 16 - 4.72% CDGR, was just over 80k new infections. That’s scary. It will require vigilance, and continued social distancing to see what this brings. We’d like to see this sequence be below 2.5%13/

Today NYC released cumulative fatalities of confirmed and probables totaling 12,712. This is used to calculate the infected numbers for March 2017 (previous posts explained the heuristics and rationale), two days after the peak is suppose to have crested. 14/

The calculation for the day growth percentage of infected individuals for the group is 3.86% (March 24-25) and the two day CDGR, March 23-25), = 4.29%. 15/

For NYC the single day (since March 24) and cumulated daily growth rate infections (by age groups since March 23. Current total number estimated infected in NYC -
0-17 0 (0) 11,940
18-44 3.78 (4.14) 1,159,722
45-64 4.37 (4.45) 515,634
65-74 3.60 (5.24)
Over 74 3.47 (5.13)

16/

Completing the table from the previous tweet

0-17 0 (0) 11,940
18-44 3.78 (4.14) 1,159,722
45-64 4.37 (4.45) 515,634
65-74 3.60 (5.24) 85,334
Over 74 3.47 (5.13) 89,785

Total infections in NYC 1,862,415 and 22.07% as of March 25th.

17/

Infections by age group as a percentage of all in that group (New York City) as of March 25, 2020.
Age Total % in group in NYC
0-17 0.648
18-44 34.300
45-64 25.515
65-74 12.039
Over 74 15.201
18/

Percentage of COVID-19 deaths in NYC by age group -
Age group %
0-17 0.0236
18-44 3.9411
45-64 22.3647
65-74 24.2448
74 < 49.4257
19/

Number of COVID-19 deaths by age group per 1 million infections (NYC)

Age group #
0-17 27
18-44 432
45-64 5,514
65-74 36,117
74 < 70,000

20/

If the population were evenly distributed between age categories and the attack pattern of the virus were identical regardless of the age group, the infection fatality rate, would be 2.2418 % based on this particular model. Just sayin’.

21/

Age related odds of dying from COVID-19 after being infected
Age Odds ratio
0-17 1 : 37036
18-44 1 : 2314
45-64 1 : 180
65-74 1 : 27
75 < 1 : 13

22/

Estimated fatalities if NYC were to be completely saturated with COVID-19 infection = 79,595, fatalities. IFR = 0.9423
Age # Fatalities
0-17 47
18-44 1461
45-64 11,142
65-74 25,599
74 < 41,346
23/

NYC, with population distribution held constant as of (April 17, 2020); reaching the approx. 80% population herd immunity est. for an R0=5.7
Age Fatalities
0-17 11
18-44 1,461
45-64 10,193
65-74 11,050
74< 22,527
Sum Fatal 45,242
24/

The preceding tweet (24) has some inconsistencies - 100% infection rates for the age group 18-44; 23.52% for those 0-17; and a calculated 75%, presumed total immunity. Assuming folks have protecting antibodies, the % among the youngest has significant protective value for them25/

And, it turns out that most of the youngest of those still that could be exposed are very unlikely to have bad outcomes. Still, even though there would essentially, presumably be herd immunity in NYC, particular care, must be taken around those around seniors. 26/

4/18/20, cum. confirmed and probable NYC fatalities (CF), and, est. infections by age group for 3/26/20
Age CF Infected
0-17 3 111,940
18-44 508 1,175,926
45-64 2,957 536,310
65-74 3,201 88,629
74< 6,571 93,871
27/

As of March 26, for the primary model, total infected in NYC was 2,006,676 = 23.757623 % of the total estimated population of 8,446,451.The CDGR over the last day for infections was 2.253%; and the CDGR since the presumed last day of the peak, 3/23/20 = 2.58% 28/

During today’s presser Gov. Cuomo announced that the state will be conducting surveys, of thousands of NY State residents to detect antibodies for COVID-19. Consequently, I will provide model predictions each day, for NYC based on model projections with intervals based on..29/

the estimated CDGR projection going forward from the day of analysis out to the next calendar day. Therefore in the following threads I will make five predictions of the percentage of those in NYC infected by the day of the survey. This, when the data collection is....30/

complete, will, as Gov. Cuomo noted today, will be the first hard proof of the prevalence of COVID-19. Since NYC’s data is significantly different then the rest of the state, I’m assuming the data collectors will differentiate between the City and the rest of the state, ...31/

allowing for a direct comparison between my model projections, and, the actual NYC antibody data as collected on that day or days. So soon I will be estimating data for tomorrow first, using five narrow increments. 32/

Monday 4/20/20 based on range of projections to have occurred between 3/23/20 out 28 days later at given growth.
CDGR Antibodies % NYC
1.9 3,068,358 36.33
1.6 2,825223 33.45
1.3 2,600,719 30.79
1.0 2,393,467 28.34
33/

4/19/20, cumulative confirmed and probable NYC fatalities, and infections by age group for 3/27/20
Age Fatalities Infected
0-17 3 111,940
18 -44 526 1,217,593
45-64 3,030 549,550
65-74 3,308 91,591
74< 6,816 97,371
34/

One day growth in estimated infections from 3/26/20-3/27/20 was 5.38%; the four day CDGF since 3/23/20 to 3/27/20 was 3.37%. The average daily increase in infected people was 64,146. The daily average of 64,146 new infections was 81.45% of the daily pace between 2/29 to 3/23. 35/

The total estimated % of people infected by 3/27/20, in NYC was 24.484%. Next we will estimate the total number of people using CDGR’s of 2.2%, 1.8%, 1.4%, and 1.2% since March 23 going out to tomorrow estimated the % of each. 36/

Date CDGR% Infected 4/21 % of NYC DailyAvgInf
3/23 2.2 3,404,905 40.31 54,946
“ 1.8 3,038,887 35.98 42,325
37/

Date CDGR% Infected 4/21 % of NYC DailyAvgInf
3/23 1.4 2,711,001 32.10 31,019
“ 1.2 2,560,141 30.31 25,817
38/

The reason I’m tracking estimated infections by day is to alert those folks that 23 days hence, approximate 1 in 151 New York City residents will die, should the distribution of those infected remain the same. At the peak load of around 800 per day, it is presumed that 120k..39/

got infected 23 days earlier according to the model I’m using. 120,000 infections per day almost, in fact did, break NYC’s HCS. 40/

I’ve yet to break out the estimated daily rates of infection in NYC, yet, I wil provide that information if I see the lagging indicator, fatalities, start to broach those numbers again. This is why, I with trepidation notice when fatalities tick up at all. 41/

4/20/20, cum. confirmed and probable NYC fatalities (CF), and, est. infections by age group for 3/28/20
Age CF Infected
0-17 3 111,940
18-44 560 1,296,296
45-64 3,200 580,383
65-74 3,504 97,018
74< 7,160 102,286
42/

There was an uptick on the day ending April 20 of deaths. The number of fatalities on that day was 744. I will make reevaluations about the possibility of an increase in ICU patients remaining high based on the estimated increase in infections among the most vulnerable, soon. 43/

In any event, yesterday’s report (4/20) showed a cumulative total of confirmed and probable fatalities in New York City of 14,427, with a dramatic uptick in single day fatalities of 744. And the expected % of those producing SARS-CoV-2 antibodies as of 3/28/20 was 25.90 %. 44/

My next tweet will estimate the % producing antibodies in NYC as of tomorrow. I’ll do the calculations, still I need a rest and some Hello Fresh before all of that. I promise to tweet a picture of today’s meal - Pork Bibimbap. 45/

Estimated totals and percent infected with antibodies for April 22. High and Low bounds using CDGR’s from March 23 to April 22.
CDR Infections NYC % Antibody
2.9 4,270,613 50.56
2.5 3,799,662 44.99
2.1 3,379,102 40.01
1.7 3,003,690 35.56
46/

4/21/20 - cum. confirmed and prob. NYC fatalities; est. antibodies % by age for 3/29/20
Age CF Antibodies %
0-17 5 10.69
18-44 575 39.37
45-64 3,314 29.74
65-74 3,629 14.17
74< 7,473 18.07
Total 14,996 27.54
47/

High and Low bounds using CDGR’s from March 23 to April 23. Percentage of NYC with antibodies on 4/23/20.
CDR NYC % Antibody
2.956 52.90
2.700 48.97
2.440 45.30
2.160 41.60
1.473 33.74
48/

4/22/20 - cum. confirmed and prob. NYC fatalities; est. total numbers. With. antibodies % as of 4/22/20
Age CF Antibodies %
0-17 9 335821
18-44 588 1272727
45-64 3379 612848
65-74 3736 103442
74< 7699 109986
49/

Total number of cum. fatalities in NYC as of April 22, 2020, and total number dead as of that day, and daily compounded daily growth from the day ending 4/21/20
Cum. Fatalities Tot. Day dead CDG 4/21-22
15,411 451. 2.69%
50/

After analyzing today’s reported 21.2% estimated antibody rate among NYC residents aged 18 and over - fits my modeling of one week before the data collection - 4/20-21/20 - which puts us to April 13 data, where the group over 18, in fact estimated 22.89% of the population. 51/

Data from 4/13/20 fits 21.2% NY State estimated infection rate in NYC, based on data collected 4/20-21/20. Data below is that fit.
AG CD TI
18-44 413 956018
45-64 2417 438371
65-74 2503 69302
74< 5031 71871
52/

It seems theoretically reasonable that most folks who present antibodies detectable by the measuring method NY State used in their serological SARS-CoV-2, takes approximately one week. Given that, a calibration to my modeling is now much more secure. In other words future..53/

serological surveys will represent the state of my model estimates, 7 days earlier. In other words if NY City or state were to do a retest next Monday/Tuesday (equal data collection) the estimated infection rate in NYC, ought to be based on data collected and analyzed 4/20/20.54/

Ag = Age group; CD = Cumulative deaths on the day data was collected; TI = total estimated infections by age group, represented from the total deaths by age group, 0.0432% in group 18-44; 0.55136 in group 45-64; 3.6117 in group 65-74; 7.0% over 74< https://twitter.com/dmmf7/status/1253510720026796032?s=21 55/ https://twitter.com/dmmf7/status/1253510720026796032

Next, now our revised model (estimated IFR in lieu of surveys) needs to make adjustments about when we need to estimate the deaths divided by the total infected. Given today’s estimates based on data collected on 4/20/20, Cum deaths from that date were 14,427 and total..56/

infected above 17 was 21.2% (survey serological SARS-CoV-2 antibodies) of approximately 6701349 = 1420686. So the estimated IFR as of Monday using the survey estimates was 1.015% using the denominator using NY State data about NYC. 57/

Using the estimate from 4/13/20 https://twitter.com/dmmf7/status/1253510720026796032?s=21 the total from column TI is 1535562. In the numerator, 14427/(denominator 1535562 gives a IFR = O.932. Meaning, the IFR is around 1% in NYC for folks over 17 years old. With no vaccine in sight. Ten times more deadly/flu. 58 https://twitter.com/dmmf7/status/1253510720026796032

Now that the parameters of the model are more established I will track the data, prospectively without having to resort to Cumulative daily growth rates to estimate infection rates. I will start a new thread which incorporates these new model parameters. 59/

The true IFR will be lower, given the distribution as observed within the groups over 17 - because those numbers would skew proportionately by including those younger than 18 in the population estimates - given the very low IFR present in the NYC data. 60/

I believe that Gov. Cuomo misspoke when he mentioned yesterday that the infection fatality rate was about 0.5%, based on his serological survey of antibodies to SARS-CoV-2 take time to develop- my data analysis finds that seven days before his survey, were all those infected,..61

so that, if we assume no new infections were accruing, we’d note that number would be lower, than it is if new infections were still occurring, and the cumulated fatalities would represent a larger proportionate number consequently. Therein lies the difference between,..62/

his IFR of approximately 0.5 and my models number of approximately 0.9. In fact depending on the age distribution of those succumbing to COVID-19, the IFR will vary, more older folks, proporltionately, model estimates of IFR go up; fewer proportionate older folks, the IFR down63/

Another factor, that my model accounts for, which works in the opposite direction, he collected data only on folks over 17. That would work in favor of inflating the IFR, because of those factors, I believe NY state will eventually adjust the IFR upwards when the they balance.64/