IOWA CITY, Iowa (AP) — Iowa tried to correct a major flaw in its coronavirus data Wednesday that backdated thousands of positive and negative test results and distorted local infection rates across the state, a spokesman for Gov. Kim Reynolds said.
The announcement comes as school districts and parents have been relying on the data to make agonizing decisions over whether children should start the year with in-person or online education. Now the state is confirming it has been inaccurate, but in a way that made local conditions seem worse than they really were.
Reynolds’ spokesman, Pat Garrett, said that the state’s data system had been erroneously recording the results of people who were tested on multiple occasions.
He said that an individual’s most recent result, whether positive or negative, was “unintentionally attributed to the date of their first test.” That means thousands of recent results were recorded in the system as having occurred in March, April, May and June.
The state updated the coronavirus dashboard Wednesday to fix the errors. Garrett said revisions reduced the current 14-day positivity rate in 80% of Iowa’s 99 counties, while others saw an average increase of less than 1%.
Garrett did not immediately say how many thousands of test results were impacted, or release a county-by-county breakdown.
But some changes were dramatic. Humboldt County had the state’s highest positivity rate Tuesday of more than 22%, but it plummeted Wednesday to 8.9%. Webster County, which recently was among the state’s highest, dropped to near the bottom with 2.6%.
School districts have been using those rates to determine whether they must reopen for in-person instruction in the coming days or whether they can seek waivers to begin the year with online instruction only. Parents have been using the data to help inform their decisions about whether to send their kids back to schools.
The governor has ordered schools to return for at least 50% in-person instruction as long as the local positivity rate is below 15%, a threshold that is three times higher than many experts suggest is safe. But she has also said parents must have the choice to have their kids learn from home.
The state’s practice of erroneously backdating positive tests for someone who had previously tested negative became public in media reports this week. But it wasn’t clear until Wednesday that new negative tests were also being backdated.
Dana Jones, an Iowa City nurse practitioner whose research uncovered the problem, said Wednesday’s changes to the data were stunning and seemed to make no sense. She called on the state to make its database available to the public.
“I don’t understand what they have done at all,” she said. “We need data that we can analyze ourselves.”
Jones had been baffled when she kept seeing new infections added to earlier months’ totals and sent an email to the Iowa Department of Public Health asking for an explanation. To her astonishment, a top official described the system error in an email last Friday and said it would be fixed.
Jones shared the correspondence with the Bleeding Heartland blog, The Associated Press and other news outlets, calling it a “horrifying” mistake that needed immediate attention. Democratic U.S. Rep. Cindy Axne and others demanded that the governor’s office acknowledge the mistake publicly and correct the data.
But the governor’s office and the public health department did not answer questions about the problem or what it would mean for school districts on Monday and Tuesday. Other local journalists and data trackers had raised questions about discrepancies in the numbers for weeks but had not gotten clear explanations for them.
Another critical error emerged Tuesday when Webster County announced that a local clinic had failed to report up to 3,000 negative test results. County public health officials said the omission had artificially inflated the local positivity rate and needlessly prompted Fort Dodge to delay the school year, a decision it reversed on Monday.