Iowa Health Official Knew Coronavirus Data Was Inaccurate Before State Published Positivity Rates
Iowa’s top health official said she knew about the inaccuracies in the state’s public coronavirus data weeks before it was reported.
State Medical Director Dr. Caitlin Pedati knew “toward the end of July” that some COVID-19 test results were attributed to the wrong dates, before Gov. Kim Reynolds announced July 30 that the state would use county-level test positivity rates to make decisions about schools reopening. Then the state used flawed data to put the wrong positivity rates on its coronavirus website for school leaders and community members to see.
In an interview Thursday with Pedati, IPR asked why she didn’t say the data was wrong at the time.
“Our epidemiologists were working along with our informatics folks at that time to help understand what we were seeing and to help clarify and scope what the impact would be,” Pedati said. “Which, of course, it’s hard to know until we’ve completely walked through and done a thorough assessment and validated all of that data, which we were able to do this week.”
IPR also asked Pedati when she informed the governor’s office about the problem, but Pedati said she’s not sure of the exact date.
Reynolds said she found out about the problem last week, and said it happened because of Iowa’s “antiquated” system for tracking diseases. Her office confirmed the data flaws Wednesday and claims to have corrected them.
Dana Jones, a nurse practitioner in Iowa City who has been tracking the state’s coronavirus data since May, uncovered the backdating issue last week. She reported it to health officials who confirmed it.
“I think the fact that the data is wrong, and that’s what we’ve been driving our decisions on, is concerning,” Jones told IPR earlier this week.
Reynolds' office says thousands of cases were wrongly backdated.
Dan Diekema, the director of the division of infectious diseases at the University of Iowa, said the state needs to report and fix errors quickly because many people and organizations, including local and federal health officials and academic institutions like Johns Hopkins rely on state data.
“We shouldn't be relying on the public to detect errors in the data,” he said, referring to the nurse practitioner who reported the glitch. “But the point is that we're all very engaged right now in this pandemic in how to bring it under control. And so having, you know, those statewide data at our fingertips is really important.”
On Wednesday, after state officials said the error had been corrected, some counties saw their 14 day average positivity rates plummet.
According to the AP, Humboldt County, which had the highest 14 day positivity rate in the state on Tuesday, fell from 22 percent to 8.9 percent.
At a press conference on Thursday, Gov. Reynolds said the corrected data resulted in a drop in the 14 day average positivity rates for 79 of the state’s counties. She said the rest of the state’s counties saw an increase with an average of less than 1 percent.
Schools at the center of data problem
Last month, Reynolds announced under the state’s “Return To Learn” guidelines, schools will be required to provide at least 50 percent of instruction in person unless the county has a test positivity rate of 15 percent or more over a 14 day period and more than 10 percent of its students are absent.
On Wednesday, the Iowa City school district and Iowa State Education Association filed a lawsuit attempting to block enforcement of the state’s guidelines.
Iowa City’s request to start the year virtually was denied by the state.
At a press conference on Wednesday, ISEA President Mike Beranek said IDPH has “misled us with bad data.”
“Because of the missteps by the Iowa Department of Public Health and the reporting of the rates within counties, people are even unsure of what those numbers mean,” Beranek said. “So relying heavily on what the county department of public health and the other scientists and medical professionals in that community are saying is what our districts need to be utilizing.”
The state’s 15 percent threshold goes against the recommendations of many public health experts. The U.S. Surgeon General’s Office has recommended only communities with a positivity rate of less than 10 percent should return to in-person classes.
The World Health Organization considers any community that has a 14-day average positivity rate of more than 5 percent to have substantial community spread.
“Most public health experts would say that you need to be below 5 percent to really think about safely reopening schools,” Diekema said.
As of Thursday, according to the state’s website, Plymouth County, which has a positivity rate of 17.8 percent, is now the only county that meets the state’s 15 percent threshold to request moving classes online.
Sixteen counties have a positivity rate above 10 percent while 71 counties are above 5 percent.
Jennifer Lane, a spokesperson for the Fort Dodge Community School District, said the district was preparing to delay the start of its school year two weeks because the state’s flawed data showed a very high positivity rate for the county.
In an email to IPR, Lane said the district sent an email to families on Aug. 14 that they had made the decision to push the school’s start date from Aug. 25 to Sept. 8 following “conversations with local public health officials.”
At the time, the state’s coronavirus website was reporting Webster County’s 14-day positivity rate to be more than 20 percent, one of the highest in the state.
But Lane said the district scrapped the plan, which had not yet been approved by the school board, over the weekend when they learned there was “an error in the reporting of negative tests."
Webster County health officials announced this week that a local clinic failed to report up to 3,000 negative coronavirus tests to the state last week.
The updated data drastically lowered the county’s positivity rate to just 2 percent.
Local health departments, elected officials have concerns
This week, the Siouxland District Health Department announced it will no longer share daily COVID case data with the public due to concerns about the accuracy of the state’s data.
Siouxland District Health Department Deputy Director Tyler Brock said he felt that the state’s data was skewing Woodbury and other counties’ positivity rates higher because the state is under-counting negative cases.
Brock said while positive tests are “eventually” getting reported, he said his concern is that the glitch also means that the state is only reporting one negative test per person, even when that person has had multiple negative tests.
“We’re dealing with a system that was intended to handle a few hundred salmonella cases a year, along with some of the other infectious diseases that we deal with all the time,” Brock said. “It was not necessarily to deal with something of the magnitude of COVID.”
Pedati said for Iowans who get multiple negative tests, their most recent negative test is reported on the state’s website.
Polk County Emergency Management Agency Director AJ Mumm said during a meeting of emergency personnel on Tuesday that the county’s data did not match the state’s numbers.
“We are trying to resolve that with them,” Mumm said. “But right now we are having a difference in the data. So we'll just continue to monitor that and update our numbers as they become available to us.”
Last month, the AP reported that a COVID-19 outbreak at a Tyson meatpacking plant in Columbus Junction was much larger than the state reported. The department announced in early May that 221 employees had been infected when documents showed that days earlier Tyson officials told the state 522 workers tested positive.
Data requested from the state health department by IPR on July 17 still listed the number of confirmed positive cases at the Columbus Junction plant on May 5 as 221.
In a statement released Wednesday, Democratic Congresswoman Cindy Axne called the software glitch a “gut punch.”
“This data is the foundation on which all of Iowa’s critical safety and reopening decisions are being made and justified. Iowans have been told to trust the data – and if that data is inaccurate, it has calamitous consequences for Iowa,” Axne said.
Reynolds was asked Thursday why Iowans should trust the state’s coronavirus data.
“I think one of the reasons they can feel confident in the data is we have worked very hard to provide transparency, to provide real-time data, and we’ve updated it as we’ve heard concerns or issues from Iowans on things they wanted to see reflected on the website,” Reynolds said.
Her administration has neglected to answer several questions from reporters about data discrepancies over the past few months.