The COVID-19 situation is getting better, by the numbers, so why is the state getting even redder when it comes to weekly color-coded ratings?
The answer to that question can be found by breaking down how the state scores counties and what's happening with testing numbers.
If you picked up KPC newspapers on Wednesday, you may have seen the COVID-19 daily update story headlined "State showing continuing improvement in COVID-19 numbers."
But today, you'll find that not only has Noble County turned from orange to red in weekly county metric ratings, but that the state as a whole has gotten worse.
You may be left wondering how could things be getting better and worse at the same time? One of those must be wrong, right?
Well, no, not actually. Here's why:
First, to start, the statement that COVID-19 numbers are improving is backed up by the data. If you take a look at the graphs at coronavirus.in.gov, it's pretty easy to see for yourself.
Since late-November and early-December, Indiana is seeing fewer cases per day, fewer people hospitalized and fewer deaths per day.
The numbers are still high, higher than they were really any time before November, but they're over the peak and trending in the right direction.
That trend tracks both for Indiana as a whole but also for the local northeast Indiana area.
But what you'll also see looking at those graphs is that, during and since Christmas and New Year's, testing has dropped.
And although cases have dropped, they haven't dropped by the same-sized or larger proportions, meaning that positivity is ticking up.
And that's primarily where the disconnect is happening when it comes to county metrics numbers.
Understanding the seemingly contradictory statuses of the state requires understanding how the county metrics ratings are set up, what they're measuring and how.
The state's color-coded ratings depend on two data points in every county — new cases per 100,000 residents over the last week and average positivity rate over the last week.
Counties are assigned either zero, one, two or three points depending on where those measures are and then those two scores are averaged to get an aggregate, which determines the color. A score of 0 or 0.5 is blue, 1 and 1.5 are yellow, 2 and 2.5 are orange and 3 is red.
The first metric, per-capita case rate, has been almost irrelevant in the calculation lately.
Since case counts have been so high over the last several weeks during the late-fall surge of COVID-19, almost all counties went over and have stayed over the highest threshold of 200 new cases per 100,000 residents.
This week only one county — highly rural Newton County — had a rate lower than 200 per 100,000, but there have been numerous weeks recently that all 92 counties have been over that bar.
That means every county, or just about every county, receives a score of 3 on the per-capita side toward their overall rating.
And it's not that counties are over that 200 mark. Many counties are way over it.
There have been some weeks where some counties have topped 1,000 cases per 100,000. Locally, Noble and Steuben counties are over 600 per 100,000 and DeKalb County is over 400 cases per 100,000 this week.
A important reminder: The state uses per-capita rates, which adjusts for population size, so that a populous county like Allen can be compared fairly to a small rural county like LaGrange.
So, even if counties are seeing fewer cases and showing improvement, what's happening is they're not seeing enough improvement to get below that 200 per 100,000 bar yet, which would help them get into an orange rating or lower.
Since none of our local counties have more than 100,000 residents, that per-capita rate of 200 per 100,000 translates to 70 actual cases per week for Steuben County, 80 cases for LaGrange, 87 for DeKalb and 96 for Noble, weekly totals that all counties have been well above on a weekly basis for months now.
When looking at the positivity side, any county over 15% positivity gets 3 points, while between 10-15% is 2, 5-10% is 1 and lower than 5% gets 0.
With case counts so high and almost every county receiving a 3 on that end week-to-week, red vs. orange determination largely falls to where positivity is. Anyone over 15% gets a red rating, while anyone under has been falling into orange.
Positivity is calculated by how many people test positive divided by the total number of all tests. So if you run 100 tests and 10 come back positive, the rate is 10%.
But determining why positivity is changing up or down can be tricky, as you have to look at the underlying data.
For example, 10 cases on 100 tests is 10%, but so is 20 cases on 200 tests or 1,000 cases on 10,000 tests.
Positivity may be going up because more people are getting infected. If suddenly 15 out of 100 people tested are positive, you're now at 15%. But positivity can also increase if you test fewer people. If you get 10 cases but on 90 tests instead of 100, you're now at 11.1%.
Counties don't test the same number of people every day, so those figures fluctuate up and down. The state takes its seven-day average by taking all the positive tests over the past seven days divided by all tests over the past seven days.
But lower testing can also inadvertently increase positivity via observation bias.
People who are symptomatic are more likely to seek testing for COVID-19 than those who aren't. Therefore, you're likely to find more positive people through testing 100 people who are seeking it voluntarily compared to if you randomly tested 100 people from the community.
If people in a community don't widely test, they may find their positivity rates are coming out artificially high.
This effect is part of the reason why Noble County has generally been better than its neighbors in weekly ratings. Noble County tests more and encourages people who are only exposed to people with COVID-19 to get tested. Following that procedures helps catch some people who may be early on in an infection and helps prevent them from spreading, but it also often finds many people who simply aren't infected, which helps gives Noble County a lower and more finely tuned positivity rate.
In summary, a bigger sample size gives a more accurate picture of actual transmission rates.
So, how can numbers be improving but colored ratings aren't?
Simply put, it breaks down to two main reasons:
1) Cases are decreasing but not decreasing enough to drop to a lower level on the state ratings chart yet.
2) Testing numbers are dropping by a larger proportion than case numbers, which inflates the positivity rate, thereby driving more counties toward a red rating.
With the same score for cases per capita but counties possible seeing higher scores for positivity via the state metrics system, it's pushing more areas into the red and making the state picture look worse, even thoughoverall raw numbers on COVID-19 are getting better lately.