Worldometer Analysis of Covid-19-SARS(2) infection rate and inflection points: A comparison of Six (6) Countries

Vanity | March 17, 2020 | Gas_dr

[gas_dr]

I have spent the last several days like most Americans somewhat baffled, perplexed and even a little nervous at the rapidly unfolding events as to CoVID-19. The community seems to have been divided into two groups, those who are advocating that there is nothing more than the common flu here and that there is vast overreaction, and those that advocate that this is a serious illness that for the safety of all requires drastic measures in order to prevent mass death and tragedy. If one looks closely, it is possible to find everything from death rates calculated at 0.1% - 14% and beyond.

Fair warning, I tend to fall in the previous camp and think that this is likely an overblown situation that appears to be gathering inertia, and I have gotten into several heated and admittedly emotional arguments about this current situation. For full disclosure, I am a critical care physician practicing in the United States, and the fear of the unknown probably drives to a small degree my emotion. So this evening, I have decided to put emotion aside, and truly conduct a retrospective analysis of the data that is present in a multivariant system and see where the data lead us.

I have discovered something interesting, and will simply present the data without commentary. I have analyzed from the worldometer website six countries: China, Italy, United States, South Korea, Spain, and France. I have been curious as to the timelines and increase in daily cases in countries that appear to have rampant disease, and countries that have flattened the curve.

Methodology: Utilizing a base case rate of 100 infections, I standardized the time to 6340 infections, which happens to be the current report of United States infections updated at 1800 EDT. Datasouce: Worldometers.info

CHINA: Time to equivalent cases of United States infections: +7 days. (571 cases -> 7,711 cases). Time to Peak of newly reported cases (2nd inflection point) +15 days from current US case load. After inflection point, total new cases fell dramatically

South Korea: Time to equivalent cases of United States Infections: +15 days (100 cases - 6593 cases). Time to peak of newly reported cases (2nd inflection point) +4 days. After inflection point, total new cases fell dramatically.

Italy: Time to equivalent cases United States Infections: +15 days. (75 - 6,387). Time to peak of newly reported cases +7 days. This is where the data may become predictive. Between March 15, 16, and 17 there was a relatively flattening of new cases and decline in the last 24 hours reported period. What will be a test of this model is what happens to the cases in this 24 hour period. According to the other models, this should be the peak.

United States: Time from 100 cases - 6340 cases +15 days. If this model is predictive, we can ascertain in the next 4 - 7 days based on number of cases and shape of curve if we appear to follow South Korea numbers, or Italian and China Numbers. The current acting head of HHS repeatedly states out numbers track South Korea.

Spain: Time from 75 cases - 6391 cases +13 days. Today is day +2 from the relatively time constant 100 -> 6500 cases period.

France: Time from 100 cases - 6633 cases +16 days. Today is day +1 from relative time constant too >6000 cases.

Analysis: At this time, in the six major countries actively reporting all cases and under greatest scrutiny, there appears to be a constant of +14 days to grow from 100 - 6500 cases. The percentage variation in this is extremely small. From the time this 15 day time constant occurs, there is divergent data to the second inflection point in countries who are reporting a decrease in overall numbers of new cases (curve flattening). That time constant is from 4 - 7 days (South Korea (4) - China (7) - Italy (7)) If this is predictive model, what we should see it a flattening of the Italian curve starting today -- that would track with China. In the United States we are between (4) and (7) days from flattening if the data hold. The divergence of cases added between day 4 and day 7 are as few as 1500 new cases as in South Korea, and as many as 20,300 new cases in Italy, which fits an exponential growth from day 4 - 7.

If the Italian numbers are flat tomorrow and the next day, the model appears will become increasingly predictive. Questions that remain to be answered: Are aggressive measures able to reduce the time to second inflection as in the case of South Korea? If so, then we can predict in America that maximal saturation will be March 21. If it trends toward China and Italy, then it will be March 24.

I will update this as data become available. I look forward to reasoned discussion regardless of personal opinion as to the scope of this pandemic.

[griswold3]

Hmmm March 21-24 for maximum saturation mark

[Sacajaweau]

A graph would be a super help but the consistency is easily followed.

[Buckeye McFrog]

Fascinating that a group of countries with such varying populations all ended up at around 6500 cases in the timeframe. Thanks for posting.

[TexasGator]

“Methodology: “

hmmm .... BuzzFeed posted these graphs FOUR days ago!

[fireman15]

I assume that you may have seen this article that explored the situation on the Diamond Queen that was quarantined off the coast of Japan for weeks. The situation was not as dire as it might have been assumed from the news reports.

https://wattsupwiththat.com/2020/03/16/diamond-princess-mysteries/

[SaxxonWoods]

I’m no doctor but I’m pretty good with stats, have education in it. I got a similar answer doing similar though more shallow research.

That’s why I’m sitting here hopeful we will know a lot more in 5-7 days.

[gas_dr]

I can do pretty careful analysis, I suck at importing graphs, I included the reference site so all could peruse graphs at their leisure.

[dangerdoc]

When antibody tests are available, it will be interesting to see how many people were infected.

[gas_dr]

Thank you for your comments. I thought about including Diamond Princess and agree with that analysis, however, I thought the sample size was too small to include and may have skewed the data in a more optimistic way.

[gas_dr]

I would love to see a serology conversion rate. My thought is that the (+)serologies and much more than we know and that the rates are far less. I attempted to provide an analysis based on data available.

[null and void]

As much grief as I've given you (and you should see the ones I deleted instead of mashing Post!) over the recent past, this is good work!

I fervently hope your read on the peak is accurate and that we cleave closely to the South Korea numbers.

Thank you!

Bring Out Your Dead

Post to me or FReep mail to be on/off the Bring Out Your Dead ping list.

The purpose of the “Bring Out Your Dead” ping list (formerly the “Ebola” ping list) is very early warning of emerging pandemics, as such it has a high false positive rate.

The false positive rate was 100%.

At some point we may well have a high mortality pandemic, and likely as not the “Bring Out Your Dead” threads will miss the beginning entirely.

*sigh* Such is life, and death...

If a quarantine saves just one child's or one old fart’s life, it's worth it.

[TexasGator]

Graphs are on Buzzfeed

https://www.buzzfeednews.com/article/peteraldhous/coronavirus-updating-charts-us-world-compare

[gas_dr]

Thank you for your kind words. As I reflected on some of my more emotional posts, I came to the conclusion that perhaps my fear of the unknown drove a bit of my irrational behavior. So I have attempted to simply scale it back to the analysis of the data we know and produce a work product that regardless of opinion, we could discuss constructively.

[1FreeAmerican]

Can I get all that in a graph please...

[fireman15]

My own guess is that the tests will eventually reveal that far more people are infected that is currently assumed especially in areas like where we live in KIng County, Washington. MOre than 10,000 Chinese people returned to their homes in East King County AFTER the TRump Travel Ban. At least as many returned before the travel ban was initiated but this is not as well documented. Many of these people were likely infected, so this thing had been circulating in this area for quite awhile before it hit a vulnerable population at the Life Care Center of Kirkland and finally got picked up on the radar. With the length of time that has passed and the small number of deaths associated we may not even hit a thousand in this country. The government over-reaction appears to be a huge waste of time and resources. But it will be very interesting to see what happens over the next couple of weeks.

[Sacajaweau]

You did great.

Someone posted a graph and what that showed is that the USA pulled the reins earlier and we are keeping below everyone else in terms of progression.

Many thinks for your info.

[JayGalt]

Thank you interesting analysis.

[fireman15]

excess caps were typos.

[gas_dr]

The source data is at the web site above in the post that are the raw data. They are also updated through today and have a more extensive set of charts. One of the interesting statistical corralories appear to be that there is a parallel convergence for about 48 hours of death rate and resolved cases as a percentage (about 50%) which is immediately followed by a dramatic rise in resolved cases and decline in deaths. This is constant across all data sets. In America, we are on day 2 of that parallel convergence. We will see cases rise the next four days in a dramatic way — the questions is what happens then. The caveat may be if we actually test a ton of patients it may demonstrate a relative decline in true positives or amount tested. South Korea tested a total of 160,000 through the period I specified so I will look at those numbers.