Washington state has been reporting incorrect COVID-19 testing numbers for eight weeks by overcounting the number of people who have tested negative.
The state dashboard has shown 13% more people testing negative since April 21 than actually have, the state Department of Health (DOH) revealed Wednesday.
The inflated numbers were a result of a workaround that was used to handle the high volume of negative tests, according to a DOH news release. The state dashboard was supposed to display negative molecular tests (which detect an active coronavirus infection), but accidentally also included negative antibody tests (which show whether a person has been exposed at some point in the past).
DOH spokesperson Lisa Stromme Warren said the error was the result of a missing line of code. She was not immediately able to explain how the department discovered the problem or whether it means the trend line of the virus’s spread in Washington may be worse than previously thought.
“We understand that we made a mistake and we own it,” she said.
This isn’t the first time DOH has had problems reporting its data. DOH’s disease reporting system was flooded with case data at the end of March, which stopped the state from publicly reporting the number of new COVID-19 cases, preventing state and health officials — and the public — from seeing the full picture of the virus’s spread.
“These negative test numbers, while inflated, have not impacted decision-making as it pertains to counties advancing through phases,” the news release said, referring to Inslee’s four-phase plan for reopening county by county.
The problem didn’t affect the state’s reporting of positive testing numbers, according to DOH, but it did lead the state to underestimate the percentage of all test results that were positive.
On Tuesday, DOH had reported that 5.5% of all test results in the state were positive; after Wednesday’s correction, that number was adjusted to 6.2%.
The state also adjusted its official COVID-19 death count on Wednesday to remove seven deaths recorded in King and Yakima counties where COVID-19 was not the primary cause, Stromme Warren said.
With that update, as of 11:59 p.m. Tuesday, 1,226 people in Washington are known to have died of COVID-19, meaning the state’s death rate from the disease is 4.6%.
Over the next few weeks, DOH plans to expand how it reports deaths to include an indication of “whether we can confirm or rule out COVID-19 as a contributing cause of death, including identifying probable and suspected deaths,” according to a news release from the department.
Last month, state health officer Dr. Kathy Lofy told reporters that along with official death rulings from medical examiners and coroners, the state relied partly on hospital reports to local health departments to count COVID-19 deaths.
“If somebody came in [to a hospital] with COVID-19, and then ended up dying of something very, very unrelated, it would not be counted from that perspective,” Lofy said then. “But I do have to say that sometimes it’s pretty hard to tell if the COVID-19 infection could have contributed to the death in any way … So, in determining the cause of death, it can sometimes be a little bit challenging. We do leave it up to the health care provider to tell us whether they think COVID-19 contributed to the deaths.”
If someone dies at home, Lofy said at the time, a medical examiner or coroner can request a coronavirus test from the state if the person had symptoms consistent with COVID-19. At that point, she said, “it’s a little bit of a judgment call from the medical examiners’ perspective as to whether they think COVID-19 caused the death or contributed to the death.”
It’s impossible to know Washington state’s true COVID-19 death toll because medical examiners have had limited ability to look back at early deaths that might have been caused by the disease.
The state has also identified about 100 death certificates in which physicians listed COVID-19 as a cause, but there was no record of the deceased having a positive test in the state’s system. Those deaths do not appear in state numbers.
State health officials believe more deaths are being incorrectly left out of COVID-19 counts than are being incorrectly included.
Washington is not alone in its struggles to accurately report COVID-19 data.
Florida’s Department of Health fired its geographic information systems manager in mid-May after she objected to — and sometimes refused to comply with — what she saw as unethical requests to make changes to the state’s public portal. Rebekah Jones says the state site undercounts Florida’s infection total and overcounts the number of people tested — facts that bolstered officials’ decision to start loosening restrictions on the economy in early May, when the state had not met federal guidelines for reopening. Jones launched her own coronavirus dashboard last week.
Even the U.S. Centers for Disease Control and Prevention (CDC) has been lumping together tests that measure different things, the Associated Press reported last month. Officials at the CDC and in multiple states have acknowledged that they combined the results of viral tests and antibody tests. Doing so makes it difficult to understand how the virus is spreading.
Such errors render the CDC numbers about how many Americans are infected “uninterpretable,” creating a misleading picture for people trying to make decisions based on the data, Ashish Jha, director of Harvard’s Global Health Institute, told the AP at the time.
In Arizona, Gov. Doug Ducey used a graph showing a declining rate of positive tests in May when he announced that barbers, salons and restaurants could reopen. Ducey did not disclose during the televised news conference that the figures combined diagnostic and antibody tests. Positive results from diagnostic tests were declining, according to published state data, but adding the antibody tests made the decline look steeper.
“It is incumbent on health departments and the CDC to make sure they’re presenting information that’s accurate. And if they can’t get it, then don’t show the data at all,” Jha said in May. “Faulty data is much, much worse than no data.”
Seattle Times staff reporters Lewis Kamb, Evan Bush and Asia Fields contributed to this report, which also includes information from The Washington Post and the Associated Press.