Daily Summary & Notes

Today’s report uses the data from the 2:50PM Report from the GA Department of Public Health.

Today we saw 3273 new cases (our record for new cases is 4813), which brings us to 21741 in the past 7 days (9.2% of total cases so far). We also had 96 new deaths (our record for new deaths 122), which brings us to 483 in the past 7 days (10.3% of total deaths so far). We saw 210 new hospitalizations (our record is 447), bringing our 7-day count to 1472 (6.7% of total hospitalizations so far). Lastly, we had 43 new ICU admissions (97 is the record), bringing our 7-day count to 303 (7.5% of total ICU admissions cases so far).

I realize these numbers can be hard to put into context, so here’s an alternative metric. In the past 30 days we’ve had equivalent of 34.63 Cruise Ships full of infections, 4.28 747 Crashes in deaths, 30.73 movie theaters worth of hospitalizations, and 66.8 hotel fulls of ICU patients.

For testing, we saw 18013 new COVID19 tests, bringing us to 190100 in the past 7 days (9.6% of total COVID19 tests so far). We also saw 2129 new antibody tests, bringing us to 12436 in the past 7 days (4.9% of total antibody tests so far).

Today I just want to take a moment and point out that Kemp’s latest executive order on masks is exactly the sort of cosmetic measure that has led GA to having such poorly controlled community spread of COVID19. Rather than allowing municipalities the ability to enact meaningful ordinances, he has chosen to limit them to only controlling city property (something they could already do), and essentially eliminating enforcement abilities by mandating warnings before citations. This is cosmetic at best, and cowardice at worst.

Data

Data Notes

Prior to 5/11, all data is taken from the noonish update from the GA Department of Public Health to present even time intervals between data points which is important for graph interpretation. On 5/11, reporting schedule shifts to being at 9AM, 1PM, and 7PM, so this report will capture to the 1PM reporting time. On June 2nd, reporting was reduced to once a day at 3PM. Data does reflect multiple inefficiencies and inaccuracies in the current reporting system, including showing tests before their results are returned, delays in reporting on weekends that create artificial spikes and valleys in change data. In general, interpretation should examine the general trends, and not focus exclusively on endpoint trajectories, which are highly influenceable by these data variations.

To help visualize the effects of State actions on the outbreak, I’ve added a few sets of lines to several of the graphs. The first - the vertical blue lines - show when the state of emergency went into effect (3/15; solid line) and when we might expect to see first effects from it (dotted line). The second - vertical red lines - is the Friday Shelter in Place was instituted (4/3; solid line) and the date we might expect to see first effects (dotted line). The third - vertical pink lines - show when the shelter in place was lifted (4/30; solid line) and the date we might expect to see first effects (dotted line).

In addition, to help visualize change in graphs using cumulative data that spans large counts, both linear and algorithmic scales are offered. You can read more on interpreting graphs using log scales here.

Where point data is presented, a LOESS regression with 95% confidence intervals is shown to help the viewer interpret overall trends in the data. This is preferred over a line graph connecting all points, which tends to over-emphasize outliers in report.

Cumulative Confirmed Cases

Cumulative Hospitalizations

Cumulative Deaths

Cumulative ICU Use

Change Patterns

Count Level Tracking

Growth Rate Percentage Changes

These charts show the daily growth rate of each counted statistic. The growth rate is calculated using the general formula below by taking the number of new reports today as a percentage of the cases from yesterday.

A note regarding interpretation - because this compares incremental change to the total count, the daily growth rate can be deceiving. To illustrate, suppose we started with just 100 cases and tracked growth rate while adding 10 cases each day. We would offhandedly say the growth rate has remained consistent - it’s a consistent rate of 10 per day, which is 10%, right? On day 1, that’s accurate. But on day 10, that 10 new cases (a steady increase) is only a 5% increase; on day 20 it’s 3%; and on day 30 it’s a measly 2.5%, even though the total number of cases each day hasn’t changed. To maintain the same growth rate, we need to significantly increase the number of new cases per day. By day 10 we need 24 cases to keep the growth rate; by day 20 we need 61 cases; and by day 30 we need 159 cases! Now imagine that our testing capacity was a mere 50 cases per day, and at least some proportion are being given to people are not suspected of infection as a precautionary measure (let’s say 1 in 5, or 20%). In that scenario (which is similar to our real lives), we’d be unable to accurately track growth after just day 12! Unless we dramatically increased testing capacity in this case, we’d be showing just 3% growth on day 30, despite a real growth rate of 10%! Additionally, we would have failed to identify 52% of cases due to lack of testing. Now remember that we’re on day 164.

Z Score Fluctuations

Because percentage growth becomes misleading over time, I’ve added a floating 4-week Z-score visualization for each measure to help put into perspective the magnitude of daily variation in numbers.

For those who don’t spend a lot of time in the world of statistics, a Z score is a measure that describes the relationship of an observation (in this case, a particular day’s number) to the average across the entire group. It is calculated by taking the difference between the observation and the mean, and dividing by standard deviation.

\[ Z = (O - Mean) / SD \]

For example, if the mean score for a group is 50, and the standard deviation is 10, then a score of 60 would have a Z score of (60-50) / 10 = 1, and a score of 20 would have a Z score of (20/50) / 10 = -3.

This can be useful in identifying patterns in data reporting, and help put daily fluctuations in perspective. Because the data is more localized, it doesn’t fall victim to the diminishing returns effect. These visualizations are limited to the data from the last 30 days, which further helps illustrate trends and fluctuations.

New Cases

For today’s cases, the 30-day mean is 3463.1 and the standard deviation is 684.0391373.

Hospitalizations

For today’s hospitalizations, the 30-day mean is 256.0666667 and the standard deviation is 139.0807092.

Deaths

For today’s deaths, the 30-day mean is 52.1666667 and the standard deviation is 34.5823189.

ICU Admissions

For today’s ICU Admissions, the 30-day mean is 43.5333333 and the standard deviation is 24.8592359.

Testing

These graphs contain several markers that reflect the changing nature of the testing data that has been provided over time.

As of 4/28 specific counts of the number of tests administered by the government and commerical providers stopped being reported. Additionally, on this date we began to track data on the number of positive tests conducted by the CDC.

On 5/27, specific counts of serology tests (antibody tests) became available, which had previously been aggregated into the total test count. This date has been marked with a vertical gold line on the graphs. This distinction is important, as positive antibody tests do not result in new cases in the overall count, and thus both suppress the positive test rate and artificially inflate estimates of test prevalence. The daily data for daily COVID19 tests and serology tests is tracked starting on this date.

For today’s new tests, the 30-day mean is 29006.2 and the standard deviation is 7615.6166341.

For today’s new COVID-19 tests, the mean is 26994.8 and the standard deviation is 7287.4326561.

For today’s new antibody tests, the mean is 2011.4 and the standard deviation is 976.3435517.

Is Increased Testing Causing Increased Cases?

A popular talking point recently is that the increase in cases that are being detected is not reflective of increased spread, but rather a result of increased testing. There is a certain logic to this - the more tests that are run the more potential cases we can identify. However, this can lead us to significant logical errors, and these in turn can lead to dangerous behaviors. While our data does not allow a perfect causal analysis, we can examine what associations between testing and cases exist in our data.

Correlations Between Testing and Cases

If we run a simple correlation between total number of tests and total number of cases, we get an initially persuasive graph. Note that this graph includes both antibody and molecular tests.

This gives a correlation of 0.9883368! This is inviting, but it mostly just shows that both of these numbers are increasing. This is potentially misleading because it looks at cumulative data. In fact, if we run a correlation between the total number of tests administered and a simple series of ascending numbers (1, 2, 3, etc.) we get a correlation of 0.9623664. Because our hypothesis (increased testing causes increases in reported cases) is more about fluctuations in these two variables than cumulative growth, we need a different analysis.

If we look at the daily increase in cases against the daily increase in tests, we get a different picture:

This gives us a correlation of 0.8102776. But this number is also misleading, because there are significant time lags in reporting of tests and new cases within the data.

To better assess the relationship, let’s look at 10-day moving averages for both new tests and new cases, and see what correlation exists between them. This will help balance out the issues of delayed results.

This gives us a correlation of 0.89. By the observational nature of our data, we can’t infer causation.

As I’ve watched this plot evolve over the past few weeks, I think it’s starting to become clear that we have two different distributions happening. The first is the relatively flat group of cases you see across the bottom, which has characterized most of our COVID19 response. During this time, we had relatively stable case numbers regardless of whether we saw heavy testing days or light testing days. In this group, it’s pretty clear that testing and number of cases detected aren’t strongly related. However, we also see a second group of cases, which seem to veer upwards rather abruptly at around 12500 cases per day. These data points are more recent. If our data was limited to these cases, you could make a case that there’s a strong association between increased testing and increased case identification. As it is, our correlation estimate ends up sitting somewhere between the two lines.

Having watched the data evolve, my hypothesis is that the increase in testing is a response to the increase in cases. of particular relevance to this hypothesis is the gap between our two “groups” above the 15000 test per day level. I think if testing were the driver of new case identification, then we’d see greater variation at the higher testing levels, rather than two distinct low variation areas forks. When we consider the hypothesis that increased case reports spurred an aggressive increase in testing, the pattern here makes more sense. When the number of cases detected at ~12500 tests per day began increasing, testing itself was escalated to try to keep up. This hypothesis would also account for the differing curves between new tests per day and new cases per day.

To further explore the relationship between testing and outcomes, let’s consider the relationships between testing and both deaths and hospitalizations.

Correlations Between Testing and Deaths

This gives a correlation of 0.9635046! This again shows the issue with simply comparing linear additive functions.

If we look at the daily increase in deaths against the daily increase in tests, we get a different picture:

This gives us a correlation of 0.2258687. Of course, this number may also be significantly misleading - after all people probably aren’t dying the same day their test results come back! Since we don’t have an average time from infection to death, and because that’s probably a wide range, let’s explore a 14-day moving average.

This gives us a correlation of 0.38.

Correlations Between Testing and Hospitalizations

Considering hospitalizations, we get this:

This gives a correlation of 0.9895002! This again shows the issue with simply comparing linear additive functions.

If we look at the daily increase in hospitalizations against the daily increase in tests, we get a different picture:

This gives us a correlation of 0.3541314. Again, there’s probably some lag between hospitalization and when people get test results back - some are likely hospitalized before results are returned, and vice versa - so let’s look at a 10 day moving average.

This gives us a correlation of 0.7.

Correlations Between Cases and Hospitalization

Lastly, let’s look at the correlations between these indicators and cases themselves.

Considering hospitalizations, we get this:

This gives a correlation of 0.9804732! This again shows the issue with simply comparing linear additive functions.

If we look at the daily increase in hospitalizations against the daily increase in tests, we get a different picture:

This gives us a correlation of 0.6516253.

This gives us a correlation of 0.9.

Correlations Between Cases and Deaths

Considering deaths, we get this:

This gives a correlation of 0.9331433! This again shows the issue with simply comparing linear additive functions.

If we look at the daily increase in hospitalizations against the daily increase in tests, we get a different picture:

This gives us a correlation of 0.425242.

This gives us a correlation of 0.44.

Final Thoughts

What do we make of the information from these new graphs? I think there are a few takeaways. First, it’s safe to say that while the increase in testing does create an increased ability to detect cases, it is not the reason that cases are increasing; after all we’re seeing similar escalations in hospitalizations and deaths which couldn’t be caused by increased testing. Second, like with the correlations between new tests and new cases, we can see that there seem to be multiple groupings within this data, which likely reflect periods of escalated testing in response to increased cases and changes in how we treat patients diagnosed with COVID19. Ultimately the story we see here is much richer and more complex than those who want to blame pandemic numbers on testing are willing to acknowledge.

Commentaries

Is Herd Immunity A Viable Solution to COVID-19?

Is Using Common Sense A Viable Solution to COVID-19?

Should People be Protesting During COVID19?

Making Sense of that Mask Study

Comorbidity (Written 7/15/2020)

I think today is a good time to remind people about comorbidity risks. I often see people insist that they have no risk because “only people with pre-existing conditions get COVID”. While pre-existing conditions are associated with increased risk, this misses both that healthy people with no prior conditions get COVID, and that what’s counted as pre-existing conditions is pretty broad. The GA DPH website indicates that the following are considered comorbid conditions in COVID19 data reporting: Chronic Lung Disease, Diabetes Mellitus, Cardiovascular Disease, Chronic Renal Disease, Chronic Liver Disease, Immunocompromised Condition, Neurologic/Neurodevelopmental Condition, and Pregnancy. These are very prevalent conditions here in Georgia - Over 6.9% of adults have COPD or other lung disease, more than 1 in 10 Georgians have diabetes, and more than 1 in 3 Georgians have some sort of cardiovascular disease. I could pull stats fo r the other conditions listed, but the implication is clear - a large proportion of our citizens are at elevated risk. Most people likely either have one of these comorbidities, or are close to someone who does, and don’t recognize the risk.

Physical Distancing (Written 7/31/2020)

Today I want to talk briefly about social distancing. The guideline that’s been shared is to maintain 6 feet distance between people. Unfortunately, many people struggle with this. The struggles tend to fall into two areas.

First, some people are not good at judging what 6 feet away is; most people I see are treating 3-4 feet as 6 feet. Often people also may start at 6 feet away, and slowly close that distance (sometimes unconsciously). With all this in mind, let me give you a few ways of thinking about what 6+ feet looks like:

  • If you could shake hands without moving your feet from where they are, you’re too close.
  • If you could fall face forward - just straight face planting into the ground - and the other person could catch you, you’re too close.
  • If the person could hit you with a baseball bat without leaving where they’re standing, you’re too close.

The second issue is that many people interpret “stay at least 6 feet apart” badly. Much like how people interpret a speed limit of “55MPH” as “drive 55MPH, if not more”, people interpret social distancing stay “6 feet between persons, if not slightly less”. Aside from making the absolute minimum safe distance the norm, this also tends to ignore the reality of human beings as 3D creatures.

Consider the following situation: You have a line 30 feet long in front of a service counter. How many people can stand in that line and maintain social distancing? In the abstract, we might quickly calculate that 30/6 is 5, or even conceptualize 6 if we assume we can put a person at spot “zero”. Now let’s think about actual human beings. Let’s assume the first person in line stands 1 foot from the service counter. If we assume a personal space bubble of about 2 feet, then a minimum of 6 feet away for the next person is 9 feet from the window; subsequent spots are at 17 and 25 feet. Suddenly we’re down to 3 people in our socially distanced line. If some of our “spots” are filled by groups of people, like families or couples, then we have to build even more space.

With those two issues in mind, I encourage you to think about social distancing as 10 feet away rather than 6 feet. This accounts for our poor spatial judgment, tendency to drift closer, and the issues of humans being 3-dimensional. This also creates space to move in and around people if you’re working in a classroom or retail environment.

Final Thoughts

As always, I am not trained in epidemiology, and defer to recognized experts in the field on all issues. These analyses and commentary are soley designed to help lay persons approach the publicly available data and larger public health conversations.

Stay Home.

Wash Your Hands.

Wear a Mask.

Assume You’re Infected.

Documentation

Code and data available here.