INDIANAPOLIS — State officials have recently identified two issues with how the state is calculating COVID-19 positivity rates and will be fixing them before the end of the month, which will impact rates.
If anything, the impact is likely to make the statewide rate look worse — they’re likely to go up potentially by a few percentage points — although rates for individual counties is more likely to be mixed, with some small counties possibly seeing improved rates when the errors are fixed.
Indiana State Health Commissioner Dr. Kris Box detailed the issues that have been found and how they will be addressed as of Dec. 30.
But first, Box stressed that the underlying data itself — how many tests have been done, how many cases have been identified, the number of cases per capita, how many and deaths have been verified — those raw statistics are accurate and unaffected by the errors.
The issues only affect display of the state’s all-time positivity rate and calculations for seven-day positivity rates that the state looks at as a rolling average.
Box clarified that there are two issues, one in coding and on in arithmetic.
On the coding side, Box stated that an error was discovered in the “software logic coding” used to calculate how the positivity rate and how testing numbers were being pulled into calculations.
On the arithmetic side, the state will be changing from taking a straight average of seven days of single-day percentages to instead running a calculation based on the actual raw numbers.
“This will help to minimize the effect that a high variability in the number of tests done each day can have on the week’s positivity, especially in our smaller counties,” Box said.
To visualize how this would work, currently the state is taking a seven-day average of one-day positivity rates. For example, if the positivity rates for the last seven days were 10%, 12%, 12%, 14%, 9%, 10% and 9%, adding those together and dividing by seven would give an average of 9.57%.
But based on the actual numbers, as testing numbers fluctuate up and down and can change, if those were based off of results of positives per total tests of 10/100, 12/100, 24/200, 35/250, 9/100, 20/200 and 18/200, when you add the total positives over those seven days, 129, and divide by the total number of tests, 1,150, you get a different percentage — 11.2%.
To summarize, the first method doesn’t take into account changes in the sizes of the raw numbers going into the calculation.
Box reported that on the statewide basis, the two fixes are likely to result in an all-time positivity rate that’s approximately 2-3 percentage points higher than currently reported.
The impacts on individual counties may be mixed. Smaller counties that test less — a county like LaGrange County, for example — could actually see its seven-day positivity significantly decrease under the new calculation.
Counties that test more frequently and have had more stable day-to-day positivity returns are likely to have less change in their numbers compared to low-testing counties that have higher change in their day-to-day positivity.
Lastly, despite the change in the averages that Hoosiers will see, Box said the end result will be that the up-and-down trends people can see on the dashboards will remain the same, just the lines will be shifted upward.
“If you look at the trend of this percent positivities, even though they’re higher, the trends are exactly the same,” Box said.
Gov. Eric Holcomb stressed that although the state has now identified these errors, finding it earlier wouldn’t have changed the state’s response to the pandemic.
Box also noted that since the issues don’t affect any of the raw data itself, other measures such as the per-capita case rate that is used in the color-coded ratings wouldn’t have been affected.
County color levels may be impacted by the correction in the positivity averages and, especially lately since every county in the state has been over 200 cases per 100,000 residents on the case count side, the main determining factor in whether counties are orange or red depends on whether their seven-day positivity is over 15%.
The update, set to take effect when county metrics scores update next week on Dec. 30, could cause a shift in how many counties and which ones are in orange or red rating.
© 2024 KPC Media Group, Inc.