todnewman.com

San Marino is a small city-state of around 33K people completely surrounded by Italy. It is located at 43 degrees N. Latitude. Its climate is described as humid subtropical. Because of its small size, San Marino flew under the COVID-19 radar for a long time, largely due to the disaster that was making the news in Italy. Once San Marino’s deaths and cases were normalized by their small population, however, it became obvious that this country had been the hardest hit per capita in the world. See the values on the Y axis for San Marino and compare to Washington, which I have shown below. San Marino has had to date over 10x the number of cases per 1000 people and over 20x the deaths that Washington (one of the hardest hit in the USA has experienced!

The Good News for San Marino: The charts above show that both the case curve and the death curve have entered a region of deceleration. There are still cases being reported, but they are less each day. Per the San Marino Coordination Group for Health Emergencies, they had a total of 193 cases confirmed as of 4/3 and only 47 of these were hospitalized. Of the 47, only 13 had severe symptoms. That breaks down to 25% hospitalization and 7% in ICU (not sure how many needing ventilators). Is this a picture of what a slowing outbreak looks like? San Marino only had 2 deaths reported on 4/3. These numbers are all down from 3/30 when they had just 192 total cases and 59 in hospital with 16 in ICU. I wish I could find hospitalization numbers like the above for each day of the outbreak, but so far I can’t locate them. I am finding scattered reports from countries and states on this, but in general there’s not much. Iceland has a good portal where they’ve advertised 45 hospitalizations (12 ICU) on 1417 confirmed cases. The denominator is very high for Iceland compared to any other country largely because they’re doing randomized testing. They have indicated that 50% of their confirmed cases have no symptoms. So, of the symptomatic cases, Iceland is seeing 6.3% hospitalization and 1.7% ICU. Much lower than San Marino still, but there’s a good chance that San Marino has a large untested population that isn’t being counted.

Update for Data Nerds. Here’s a table that you might be able to stare at for quite a while. This has all the key parameters that you might care about and give you the data to mentally refute the News Anchor throwing out bad or misconstrued analysis. I’ve heard quite a bit of this.

My Analysis:

Here’s what I see…

1. New York has crazy high numbers across the board. It’s not simply a matter that they’re earlier into their outbreak. At least as far as we know, they’re in the same boat with Washington and California time-wise. Their situation just got far worse for some reason. Their deaths per 1000 people is about twice as high as the next highest (Louisiana). I’m guessing Louisiana is in their situation due to the extensive hospitality industry they host in New Orleans. Lots of travelers from all over the world, and, oh, Mardi Gras was right at the start of the US outbreak. But New York is quite an outlier. Still not as bad as Italy or Spain, but compared to the rest of the US States, it is first in Cases per 1000, Deaths per 1000, and the Confirmed Case Rate of Change (IROC) feature. This means, they’re worse and getting even worse faster than any other state!
2. Louisiana’s Death Rate is getting less steep. This could be a one-day anomaly, but the delta IROC also shows that the slope has stayed consistent for at least 2 days. I’m hoping this means that the acceleration of their death rate has ended.
3. Vermont’s Death Rate, which was high for a while (small states can see faster impact from seemingly-small numbers of deaths), has at least temporarily slowed. You can see this by the dIROC-deaths going to zero and the IROC-deaths being nearly zero. We’ll keep watching this, but I have noticed a trend of smaller states and countries having really bad-looking statistics for a few days and then the outbreak peters out. San Marino may be another example of this. Washington also seems to be in the same category. They’ve been hit hard early in this outbreak. Hopefully this is a good sign and not a short-term data anomaly.
4. New Jersey has the fastest accelerating curve of normalized deaths over time. I’ve noticed this a couple of times from New Jersey… they accelerate then slow down. Could be bad data capture/release practice or could be something else.
5. The number of cases are accelerating for all states. What this means is that instead of something like 20 new cases each day (which would be a constant slope of 20 cases per day), all states appear to be seeing acceleration like 20 new cases, then 30 new cases, then 50 new cases, then 100 new cases. This is probably completely normal due to the way viruses spread. Each viral infection results in 1.2 to 2.5 new infections. Viruses always top out somewhere, though (none infects 100% or even 50% — Spanish Flu came close — of the population). 15% is a number that is common for Flu epidemics. So when we see COVID-19 top out, we should see the accelerations of the Confirmed Cases go to close to zero and the IROC for confirmed cases go close to zero. Keep watching for that.
6. Deaths seem to be staying low or even slowing across the states. This is also reasonable. Lets keep watching to see if this changes for the good or for the bad.

These analytics are all based off of 4/1 data… I’m noticing the groups disseminating the data release it a few hours later every day.

US Update

Above is the top states sorted by the Growth Rate of deaths in their state (inst_rate_of_change). I tend to refer to this as IROC as shorthand. Right now, this is about the most interesting metric to me because it uses deaths normalized by population (thus making it possible to compare apples to apples) and calculates the rate of change of growth. This would be the “steepness” of the curve. When this curve flattens off, it is a sign that the early phase of the outbreak might be nearly over. See image of the Louisiana outbreak below. Louisiana’s death rate of ~0.06 deaths per thousand population is very, very high right now. Not as high as New York’s though (contrary to what the lady on CNN said last night… I hear so many incorrect assertions by the big media talking heads). So this can be considered a very steep curve. The blue curve that I fit to the datapoints is what I use to calculate IROC. Louisiana’s IROC of about .009 is calculated for the 4/1 datapoint. Two days ago, their IROC was about 0.07. So their curve is still accelerating. Orleans Parish alone announced almost 900 new cases yesterday. That’s on a population of around 400,000.

World Update

As you can see above, the largest growth in cases across the world is still Spain, Switzerland, and Italy. Switzerland, however, still has about 1/4 of the deaths (when normalized by the population) of those two. I have discussed this in a previous blog. You can see the numbers for this chart below sorted by Deaths per 1000 people. Austria is also interesting along with Germany and Norway for the extremely low number of deaths in those countries.

Progression of COVID-19

Below is a chart of the top 12 countries (minus the US) by Confirmed Cases. This chart shows the duration of the outbreak from 1/28 to today. Interesting things to note… there is evidence that some of the European countries are decelerating a bit on confirmed cases. This could be for any number of reasons, not just that the outbreak is slowing. But over the last few days, we see the first time that countries like Italy have shown any slowing at all. Germany, Austria, Switzerland, Belgium all seem to be flattening off a bit. Spain, the UK, and Turkey show no signs at all of flattening. Keep watching here and I’ll let you know when the curves really indicate slowed growth in cases.

I’ve seen various references recently to COVID-19 as a disease of the rich — a reference to the fact that it is most active in very developed countries. Typically, the foil compared against COVID-19 is Ebola, which has appeared primarily in the most undeveloped countries. But is this true? Does the wealth of these countries have anything to do with the outbreaks? Lets look at some data. In the image above, one could imagine how a media outlet, hungry for a story, might look at the data and surmise that only rich countries are being affected. This diagram is trying to show the state of the world at this moment and as you can tell by a relatively blank Asia, history is not being captured. So, the recency bias is in favor of the rich country narrative.

The Latitude Effect

I have put great energy into measuring what is going on with COVID-19 by latitude, because I’m aware that in general, latitudes have similar temperatures. You can see the state of the latitude effect in the table below. In general, it’s unclear why 40 to 60 degrees latitude has been so heavily impacted by cases and deaths (particularly 40 to 50 north latitude), but one can tell that many of the countries in these latitudes are wealthier countries. This might be one of the drivers of the over-distribution of COVID-19 to rich countries, but can’t explain it all. If it did, we’d see the effect across all longitudes. So there must be more.

Other Factors

Another subject of a handful of my analyses is the correlation of other factors with outbreak severity. The World Bank and the World Health Organization have made available data about numerous factors measured across countries. This includes medical factors (Body Mass Index, Diabetes prevalence, etc.) as well as environmental factors (average temperature, pollution, etc.). I have blended these data with my Instantaneous Rate of Change (IROC) numbers for each country to see if I can correlate these many factors with the severity of the outbreak as measured by the IROC.

Correlated Factors with the Severity of Case Growth

In the table above, you can see on the factor on the left and the correlation with the IROC of confirmed case growth on the right. Let me discuss a few of these.

1. Female Smoking Rate: This factor has been the most highly correlated factor with IROC of Case Growth ever since day one. Why? All I can figure is that from looking at the world bank data, Male Smoking Rates have little variance across countries whereas Female Smoking Rates do. For instance, Iran has nearly zero rate of Female Smoking whereas in Italy, the numbers of Female Smokers equals the Male numbers. So therefore, if Smoking is correlated with case growth (as this would indicate), high female smoking rates would naturally drive even more cases than would low female smoking rates. Also, in many countries, there may be a non-linear effect on case growth if females who might have been in caregiver roles became sick.
2. Body Mass Index Standard Deviations (STD): We know from reports that doctors are seeing extreme COVID-19 cases in folks with high BMI’s. The data above confirms this and indicates that high male mean BMI results in higher numbers of COVID-19 cases. I have no idea why the male BMI is more highly correlated with case growth than female mean BMI. However, the more interesting thing is that the Standard Deviation (in general English, the STD is a number that explains how data is spread out from the mean. For BMI, a large STD means that BMI numbers are all over the board, not just clustered around the country’s average) of Female BMI is the second-highest factor correlated with COVID-19 case growth. Perhaps this is a similar discussion to female smoking above. It may point to a likelihood for extremely high or low BMI’s, which might be a health factor that could lead one to contract the COVID-19 virus.
3. Nurses per 1000 people: The only reason I could figure that this might be correlated with COVID-19 case growth is that first, circular reasoning would dictate that countries that we know are currently affected are wealthier and have more nurses… the correlations are not causations, remember. Second, however, and more palatable, having more nurses could equal more confirmations because nurses can give tests.
4. Population Density: It seems obvious that population density would be positively correlated with case growth, and it is. But not so much so as the above factors.
5. Income Group: The first negative correlation is Income Group, and this is artificial by the way my data is represented. In my data, the World Bank’s high income group is represented with a 1 and the Lowest Income group is represented with a five. Therefore, when the income group numbers go down (to 1), the case growth increases. This confirms what we already see, that the wealthier countries are where the case growth is higher.
6. Temperature: This factor is the average March temperature in each country. This factor is also negatively correlated. As Temperature goes up, the case growth goes down. This is also pretty obvious, but it’s good to see there are numbers to support the obvious.
7. Other Diseases: Tuberculosis Rates, High Blood Pressure, Diabetes. All of these show negative correlations with case growth. Why? Aren’t these co-Morbidity factors? Well, all I can figure, is that first, things like TB are higher in poorer countries and maybe this is just biased by the wealthier countries that have high case growth. The second thing I can guess is that possibly, these factors lead to deaths for other reasons where they are high and possibly there are fewer people at high risk to contract COVID-19? More to think about.

Correlation with the Severity of Death Growth

Since I have made it clear that I don’t value the quality of Confirmed Case Data, this correlation is more interesting to me. See below.

You’ll notice that some of the factors correlated with Case Growth are also correlated with the Death Rate. Female Smoking and BMI STD are two. This is unsurprising, as there is clearly a relationship between Case Growth and Death Rates. Interesting takeaways are as follows:

1. While Percent of Population over Age 65 is highly correlated with Case Growth, it is a few points more correlated with deaths. This may indicate that people in this age group are more likely to get the disease but even more likely to die from it. That’s probably no surprise, but the numbers align with the media narratives at least.
2. Nurses per 1000 people: This factor was highly correlated with case growth but is much less correlated with Death Rate. The fact that it is positively correlated with Death Rate at all might be confirmation that this is just an effect of the fact that nations with more nurses have already had lots of Cases and Deaths. BUT the fact that this factor is a full 14 points lower on the Death Rate correlation strongly indicates that a large population of nurses is shrinking death rates.
3. Density: Population density is around 11 points lower on the Death Rate correlation table. One reasonable explanation is that high density tends to cause the disease to spread, but high density is also a proxy for the better health services available in dense cities, which tend to keep the death rate down.
4. Temperature: Temperature is much less negatively correlated with Death Rate than it was Case Rate. Apparently, temperature is a contributor to the spread of the disease, but has less effect on the Death rate.

There are more things that could be discussed (feel free to weigh in on the comments) about these correlations, but this is probably a good start.

Conclusion

Is COVID-19 a disease of the rich? I’m pretty sure I haven’t answered this completely, but I’ll reiterate a couple of points that indicate to me that it is not. First, the latitudes the rich countries are in seem to have had a March Temperature in a range that led to quicker spread of the virus. Second, as is probably obvious, richer countries have more resources to test their population. As I have shared in other articles, countries like Iceland have taken a very scientific approach to testing so they can understand the outbreak better. Third, richer/more open countries are being more honest and consistent with data collection. I have zero confidence in the numbers coming out of China and Russia, because they have something to gain from artificially low number. Fourth, there are factors associated with richer countries that might be increasing case and death rate growth: things like numbers of nurses (only correlated with case growth), higher BMI numbers, aging populations, and maybe overall smoking rates.

And of course, the data is still young. Give this another year and we’ll understand these correlations much better.

Currently we’re seeing new states entering the exponential phase of confirmed cases and deaths. As I’ve mentioned recently, I don’t trust the confirmed case statistics much, but they’re interesting as a curiosity. There’s a very wide range of approaches to testing and classification of an individual as a “confirmed case.” For instance, we have Iceland who has a randomized testing strategy that is resulting in large numbers of asymptomatic confirmed cases. On the other extreme we have Iran and China, where reports indicate all confirmed cases were symptomatic and largely hospitalized. So in China’s Hubei province, the 4.7% death rate was calculated largely on a base of cases made of of largely very sick people. If one assumes that Hubei’s 68K confirmed cases was just 40% of the total infected population, the death rate drops to 1.9%. (See HERE for report on how Iceland is finding out that 50% of total cases are asymptomatic). There’s a very big difference depending on how one measures. In the chart above, we’re measuring the number of confirmed cases per 1000 residents (the color of the bubble) along with the instantaneous slope of the curve of deaths per 1000 residents over time. You can see on the map that if the rate of deaths isn’t severe, you can barely see the dot on the map. This way, if you can see it, it’s a region that is really being impacted by COVID-19. Recently there has been a lot of activity in Chicago, Detroit, and New Orleans along with the large numbers of cases and deaths in the NY/NJ region.

The Latitude Effect Still Holds

I have noted a few times that it is strange how the severity of the outbreaks is focused more on certain latitudes. I suspect that much of this has to do with the temperatures at these latitudes during February and March, but it’s possible that there are other reasons too. You can see in the chart above, that when normalized for population, latitudes 40-50 North and 50-60 North have most of the world’s cases. The strange thing to me is that Latitudes 40-50 North has consistently had double or more the number of deaths per 1000 people. This trend continues to hold, as does the trend of few cases between 0 and 20 degrees North (too hot there, likely) and 60 to 70 degrees North (too cold?). See the table below for the actual numbers.

Will the latitude effect hold up? As temperatures in the northern latitudes increase during April, will their rates of new cases and deaths slow? No one knows right now, but I think I have built analytics tools that will detect these trends quickly.

I introduced Instantaneous Rates of Change a few days ago thanks to a good suggestion by Brad Morantz, a fellow AI researcher. I had been trying to represent the severity of the curve of deaths over time through a linear fit, something I knew was sub-optimal, but figured was good enough for the purpose. Here’s a good link describing this concept. Now, the IROC gives me the slope of the curve at any given day. This is very interesting now that we are hoping to see these accelerating death curves turn over and decelerate. In the table above, you can see the IROC for 3/30 as well as the Delta IROC which I’m defining (right now) as today’s IROC minus the IROC from 2 days ago. This metric gives us an indicator when the slope of the curve is decreasing.

By the way, looking at the table above, it’s giving us the indication for the first time that deaths from COVID-19 are starting to hit the heartland of America, a region that had been barely affected up until yesterday. The number of deaths is still small in places like North Dakota, Montana, Alaska, etc., but the fact that these states have moved to the top of the IROC chart indicate that they are now experiencing deaths. Previously states like Massachusetts, NY, NJ, WA, and Louisiana dominated the top of this table.

Example:

I found in the data today the example of the state of Delaware. See table below where I’m sorting from lowest Delta IROC to highest. I have been waiting for a suitable example of a negative Delta IROC.

Now see the time series chart below for Delaware and you can see what a Negative slope looks like. Even though this is just an artifact of the fact that Delaware has few deaths right now, this is the signal that we’re looking for in New York or Italy, or some other hard-hit region to give us the notion that the acceleration of deaths in the region has slowed and should continue to do so.

Conclusion: IROC is a good measure for us to see the slope of the death time series curve numerically. I found the example of Delaware, not by searching through time series graphs, but by looking at the sorted IROC numbers. We’ll continue to evaluate this as an important metric in trying to understand this outbreak.

Update: 3/29 saw large bursts of new cases in Cook County, IL (Chicago), Miami/Dade County, FL, Harris County, TX (Houston), and numerous NY, NJ, CT, and RI counties. Below we see the top states by delta instantaneous rate of change of deaths per 1000 persons. This is essentially the tangent of the curve. As we all suspect, these curves are still in their accelerating phases, so the IROC is a much better way to measure the severity of the situation than a linear fit (my old method). The delta IROC that the table is sorted by shows the change in the IROC over the last two days. As the situation gets further along, I will likely expand that to 4 days. Then we should have a number that will tell us when we leave the accelerating phase and move into the decelerating phase.

At this point, most of the data aggregators (including Johns Hopkins) have failed to keep track of the myriad of reporting organizations and have settled for less comprehensive data collections. So I have had to go off and decide which metrics are the most meaningful and aggregate them myself.

Because confirmed cases are so dependent on whether a country/state is testing and also because cases range from no symptoms all the way to stays in the ICU and even death, I have decided that confirmed cases are just a curiosity. In the chart above, you can see that NY has just over 2 confirmed cases per 1000 people. However, the number of deaths per 1000 people in NY is 0.02, 100x lower. Since hospitals must be required to report deaths (and since death is a fairly straightforward metric, whereas a confirmed case requires a positive test), I think this is a much better way to evaluate the severity of the outbreaks around the world.

The above chart makes the case (that I have been trying to document) that the ratio of deaths to cases is quite different in different parts of the world. However, there’s one more important measure that can also reflect on the severity, and that is the velocity and acceleration of this normalized death rate. The velocity over a one day period is easy to calculate. It’s just today’s normalized deaths minus yesterdays. The acceleration is the more useful number to understand, however, as it relates to how rapidly the problem is going to get out of control (or under control). This is because all of these outbreaks across the world appear to be increasing non-linearly. To understand what a non-linear process might look like, think about the old saying, “The Rich get Richer.” Because they’re Rich, they get Richer, and because now they’re Richer, they get even more Richer!

So here is a table showing the slope of the non-linear curve for today and the change in this slope over the last two days. Any of these countries or states with a positively changing slope is accelerating. Any with a negatively changing slope is decelerating (what we want!)

The countries that are accelerating the fastest (inst_roc_delta) are poor little Andorra, San Marino, Iceland, Spain, and then the trio of states NY, Vermont, and Louisiana. As you can see, even a small number of deaths in a country or state with a low population, can contribute to a big change in the instantaneous slope.

So… takeaway. Deaths in New York, Vermont, and Louisiana are increasing quickly, but are still well below the rates per 1000 people that we’re seeing in Spain, Italy, France, Iran, and even Switzerland (who is also well below the rates of those other hard-hit countries). The point may well be that things will eventually equalize across the world and these numbers will align. Or it may well be that some regions will remain lucky and will be up to 10x less hit with deaths than neighboring countries. It’s hard to tell right now, so we’ll keep watching.

Still wondering if the Latitude effect on COVID-19 is real or just coincidental. I have shown the below chart a few times. In it, you can see that over 90% of the world’s cases have occurred in the band between 30 and 60 N. Latitude.

I have noticed, however, another effect that is interesting. The cases in the more Northern region of this band (between 50 and 60 N. Latitude) seem to have similar numbers of cases, but with far fewer deaths. I looked through the data and built tables and charts to see if this observation was true. Below, you can see a simple bar chart with these results. All the numbers are normalized by 1000 population, which takes into account the fact that the population in the range from 30 to 40 degrees N. Latitude is just over 3x larger than the population in the range from 40 to 50 degrees. Therefore, this chart shows us that per 1000 people, the number of cases in both of these ranges is nearly equal (about .23 cases per 1000 people). HOWEVER, the normalized death rate in 30 to 40 is 3x larger than it is in 40 to 50 (or any other 10 degree range of latitudes for that matter). See the bar chart below and the table in the next block down if you prefer looking at numbers. I think these numbers tell us that — for now at least — this trend is real.

Why might this effect exist? It could be just coincidence… large numbers of countries with aged populations between 30 and 40 N. Latitude? Other factors that amplify the outbreak? Societal factors in the Northern latitudes that enable better social distancing? I have no idea, but am very interested to see if this trend holds.

Today’s numbers are out and we see a new spiking (small) country, San Marino. San Marino (SMR) is a small micro-country fully surrounded by Italy. It shares much of it’s culture with Italy, and is also near the Northern part of Italy where the COVID-19 outbreak started.

It is a very small country, with a population just north of 33,000. To date, they have 187 cases and 21 deaths. This results in nearly 5 active cases per 1000 people and 0.6 deaths per thousand. Their death to confirmed case ratio is 12%, which is extremely high.

The above chart also brings to mind the case of Iceland (ISL), who has one of the larger numbers for COVID-19 cases per 1000. Part of this is because Iceland has created a strategic sampling approach for COVID-19 so they can find out where it is hitting hardest and triage quickly. This article gives us more of the story too…

Fifty-six individuals have recovered from COVID-19, 9,013 individuals are currently in quarantine (681 are in isolation), 2,096 individuals have completed quarantine, and 15 individuals are currently in the hospital diagnosed with COVID-19, two in intensive care, and one in a ventilator. A total of 11,727 tests have been administered. Of those diagnosed with COVID-19 in Iceland, two have died.

This is interesting in light of the chart below, which shows that Iceland’s Confirmed Cases are equal to 70% of the number of hospital beds in the country. But from the data above, only 15 of the ~700 diagnosed people are in the hospital. Because Iceland was randomly testing, they seem to have been able to keep people in quarantine and out of the hospital. This is basically the opposite of the Italy and San Marino experiences. According to the San Marino ministry of health (big kudos to Googles Italian to English translator!), 41% of their confirmed cases are hospitalized (0.2% of their entire population). For comparison, Iceland’s rate of hospitalization is just 2.3% of their confirmed cases and 0.004% of their total population.

Conclusion: This might confirm the strategy of America’s CDC (at least their recent strategy) of “flattening the curve” to spread the infection out and keep the numbers low in the hospitals. But it also makes the STRONG CASE for randomized testing to identify and mitigate more quickly.