Indiana COVID-19 positivity rate impacted by software error, fix will mean higher rates

Positivity rates will be changing on Indiana’s COVID-19 dashboard starting Wednesday next week (Dec. 30), but not due to differences in the number of actual positive tests.

At Governor Holcomb’s Wednesday press conference, Indiana’s health commissioner Kristina Box said that a software error had caused an error in past calculations. Statewide the numbers will increase by 2 or 3 points after the fix is made, she said. Counties will also see an impact.

The rolling average positivity rate in Monroe County is now about 5 percent, or about half of the statewide rate. The upward and downward trends for Monroe County and the state are roughly parallel, except for period when Indiana University students were returning to campus.

The rates that are now to be re-calculated show the same trends as those that were previously reported, according to Box, even if the absolute value of the numbers is different.

The new calculations would not have affected decisions that were made, Box said. Data on hospital resources and number of cases or deaths was not affected by the error, Box said.

In simple terms, the positivity rate is the number of confirmed positive tests divided by all the tests done. That’s what the state calls the “all test” positivity rate. It’s the one that the Monroe County Community School Corporation uses to decide whether to hold classes in person or remotely.

A different kind of positivity rate can be computed by taking the percentage of unique individuals who test positive and dividing by the number of people who were tested. Someone who tests negative 9 times, but positive once, would count as just one person who was tested. The other 9 tests would not get included in the denominator. So, the unique-individual positivity rate is defined in a way that makes it bigger than the all-test positivity rate.

Responding to a question from The Square Beacon, state health department spokesperson Megan Wade-Taxter said the “coding error” did not become apparent until a significant number of labs started reporting information. At the press conference, Box put the number of new labs at around 600.

Wade-Taxter added, “Many of the new labs report their data differently. The software was written to pull test results from a specific line, but we discovered that some labs were reporting it on different lines.”

The impact of pulling the wrong data into the calculations will be seen in higher positivity numbers. “As a result, we have undercounted the positive tests in the calculation of our positivity rate,” Wade-Taxter added.

Box also reported that the methodology for calculating rolling 7-day average positivity rates will change from an average-of-daily-rates approach to one that takes the the number of positive tests done for the whole 7-day period and divides that by the total number of tests done for the 7-day period.

The method the state has been using up to now has been susceptible to skewing, based on a very high or very low rate for a day when few tests were done.

The new positivity rate calculations will be included for the first time on Wednesday, Dec. 30, according to Box.

Tuesday’s announcement about the positivity rate glitch came on a day when the state of Indiana’s COVID-19 dashboard reported 139 new COVID-19 deaths, not all for the same date. The statewide daily high of 85 deaths came on Dec. 9.

In Monroe County the total count of deaths now stands at 85, after three more were added in Tuesday’s dashboard report.

The rolling average case count in Monroe County has been trending downward, from around 100 two weeks ago to around 75 on Tuesday.

The rolling average positivity rate in Monroe County is about 5 percent, or about half of the statewide rate. The upward and downward trends for Monroe County and the state are roughly parallel.

Leave a Reply