CDC coding error led to overcount of 72,000 Covid deaths | Coronavirus
A quiet change in how the US Centers for Disease Control and Prevention (CDC) publicly reports details of Covid deaths underscores the need for the agency to communicate clearly and transparently about the rapidly changing science, according to experts.
The past two years have created many communications challenges for the agency, which works with massive amounts of data from dozens of different sources, including states and territories.
“Mistakes are inevitable because humans are fallible, but always strive to quickly explain what happened and what is being done to prevent it from happening again,” said Tom Frieden, former director of the CDC and president and CEO of Results Save Lives. .
“You have to communicate too much, basically,” he said. “Whenever there’s something that needs fixing, be upfront about it: here’s what happened, here’s what we know, here’s what we don’t know.”
Last week, after reporting to the Guardian on child death rates, the CDC corrected a “coding logic error” that had inadvertently added more than 72,000 Covid deaths of all ages to the data trackerone of the most publicly available sources for Covid data.
The agency briefly noted the change in a footnote, though the note doesn’t explain how the error happened or how long it lasted.
A total of 72,277 deaths across all age groups reported in 26 states were removed from the tracker “because the CDC’s algorithm accidentally counted deaths that were unrelated to Covid-19,” Jasmine Reed said. , spokesperson for the agency, to the Guardian.
The problem stems from two questions the CDC asks states and jurisdictions when they report deaths, according to a source familiar with the matter.
A data field asks if a person died “of illness/complications of illness”, and the field next to it asks for the date of death. When the answer is yes, the date of death must be indicated.
But a problem apparently arose if a respondent included the date of death in this field even when the answer was “no” or “unknown”. The CDC system assumed that if a date was provided, then the answer “no” or “unknown” was an error, and the system changed the answer to “yes”.
This resulted in an overcount of Covid deaths across the demographic breakdown, and the error, when discovered, was corrected last week. The CDC did not respond to a question about how long the coding error lasted.
“Working with near real-time data in an emergency is essential to guide decision-making, but it can also mean that we often have incomplete information when data is first reported,” Reed said.
Death counts in the data tracker are “in real time and subject to change,” Reed noted, while numbers from the National Center for Health Statisticsa center within the CDC, are “the most comprehensive source of death data,” despite delays in reporting, as the process includes a review of death certificates.
Reporting cause of death is difficult even in non-pandemic times, experts said.
“It’s really hard to get hard numbers,” said Glen Nowak, former director of media relations at the CDC and co-director of the Center for Health & Risk Communication at Grady College at the University of Georgia. “It’s not just with Covid – it happens with just about every infectious disease or even foodborne disease and waterborne disease, where there’s a big outbreak.”
There are several reasons for this. Healthcare providers typically report initial data, but treating patients is a more pressing priority. Death certificates take time to complete and the cause of death may be subject to political pressure or difficulty in determining the underlying reason(s). States and other jurisdictions may have delays in sending data to the CDC; currently, a third deaths are reported after 10 days. Then the CDC processes the data, which may also be an involved process.
Most of the CDC’s data comes from state and local governments, and the quality can vary widely “because there has been decades of underinvestment in public health at the national, state, city, and local levels,” said Frieden.
“The public health and health care system that we have in this country makes it extremely difficult to collect data.”
The CDC estimates that more than 968,000 Americans have died from Covid, and this change does not appear to have affected that estimate. The tracker shows demographic data out of about 785,000 deaths, which means there could be more than 180,000 deaths not yet counted in these breakdowns.
The recent shift in demographic data shows the difficulty of offering up-to-date assessments as data reporting and analysis has lagged in the face of a massive outbreak. Data on the same topic across the CDC may also vary depending on the source and how the numbers were calculated.
“The level of precision you see in these numbers makes you think they must be really super precise,” Nowak said. Instead, these are educated guesses that help contextualize the scope of Covid relative to other diseases.
“I don’t think public health and others are doing a sufficient job of reminding people that these numbers have significant margins of error,” Nowak said. “The caveats should be clear that these are our best estimates based on data that has been reported to the CDC.”
As scientific evidence accumulates, adjustments and changes are inevitable and frequent. But significant changes in calculations and records need to be explained clearly, especially in an emergency situation where the public is frequently alert to data – and unexplained changes – like this.
“The best practice, really, is to have almost daily briefings, so you’re updated daily on what you’re seeing and answer questions daily,” Frieden said. This is how the agency has dealt with past outbreaks of Ebola, Zika and H1N1, also known as swine flu.
These briefings should be led by scientists with expertise in many fields, not just the CDC director, he said. The current CDC administration has “gradually returned to the habit of doing this, and I hope that trend will continue.”
The CDC is a government agency that provides data to inform national policy, and public health policies often have a political component, Frieden said. It can never be apolitical. “But you should never worry about the accuracy of the data.”