In a 'sea change' for the agency, CDC is sharing data earlier


She knew his journal was reviewing a CDC study on the Covid-19 vaccine for pregnant women. Traditionally, prestigious publications like his won't publish data if it has already been discussed in the public domain, for fear it will look like old news.
Walensky explained to Rubin that the data was of great public health importance, as it could inform pregnant women's decisions about getting vaccinated, Walensky told CNN. CDC wanted to publicly discuss the data before it was published, as publication can take weeks or even longer.
The two worked it out, and on August 11, the CDC announced the new data on its website, strengthening its guidance that pregnant women should get vaccinated against Covid-19. Walensky further discussed the data and guidance at White House briefing the next day.
The data still has not been published in the New England Journal of Medicine.
The release of the pregnancy data is just one recent example of the CDC bucking a deep cultural norm within the agency by releasing data after it has been analyzed and reviewed, but before it's been published.
"This is a sea change for the CDC," said Rubin, who has worked with the agency's scientists for many years.
The hope is that early release allows for quicker guidance, as happened with the pregnancy data, rather than waiting weeks or even months for publication.
"We have to break norms," Walensky told CNN. "We need to inform people in real time [and] we can do more complex analyses later."
She added that the agency still does a rigorous vetting of data before releasing it to the public, even if it might not be as full an analysis as what would be expected for publication.
"The perfect can't be the enemy of the good here," she said. "Given the moment of the pandemic that we are in, [we] need to have a deep change."
For years, the CDC has been criticized for being slow to release data, waiting for publication or internal agency review.
The agency is large and sprawling -- it has 13,000 employees and more than 10 centers - and guidelines and studies are typically passed through several layers for approval, a process a senior agency official described as often "overly burdensome."
"Even when you're just changing something minor, like a word or phrasing, it has to go back through that monster before it can be finished," the official said. "It's like even if you want to change the word 'happy' to 'glad,' you have to reprocess it -- you have to run it through that big machine again."
"The amount of approval they needed for everything is just wacky," Rubin said.
And that comes from the editor of the New England Journal of Medicine. "We're about the most careful people you can imagine," he said.
Criticism has continued since Walensky became director in January.
A Washington Post article last month about the CDC's handling of Covid-19 vaccine booster shots described the agency's "slow and siloed approach to sharing data."
In July, the Post published an internal CDC document that said the Delta variant appears to spread as easily as chickenpox. The document contained both published and unpublished data, according to a CDC spokeswoman.
It's "problematic" that the agency didn't release all of the data earlier, especially since many Americans have grown to distrust science and the government, said Dr. Eric Topol, executive vice president for research at Scripps Research in San Diego.
"When the CDC has an important data set, they should be out front with it. That's how you build trust and confidence," Topol said. "If they don't share it in a timely way, it feeds conspiracy theories."
But others have a different perspective.
On the one hand, it's helpful that the CDC is moving fast on important data, but on the other hand, presenting unpublished data gives ammunition to the many Americans who don't want to follow the CDC's advice, said Lori Freeman, CEO of the National Association of County and City Health Officials.
People who already don't want to wear masks or get vaccinated have seized on the fact that the CDC is presenting vaccine or mask data before it has been peer-reviewed by journal editors, she said.
"It might seem like a minor crack, but people will dig at that crack and pick at it to indicate those changes weren't based in real science or data because it was not published," Freeman said. "I do watch Fox News, just because it's instructional to see what they're thinking, and I see it all the time -- 'The CDC changed this and they haven't published the data yet.' "
She added that the agency needs to explain to the public that "there is value in provisional data" in some cases so that important decisions can be made in a timely way.
Walensky said she knows however it's handled, people will find fault in her agency's decisions.
"We will forever be criticized for either being too slow or being too fast," she said.
In the past two months, the agency announced several bits of Covid-19 data, including the pregnancy data, before waiting for them to be published.
At a July 1 White House briefing, Walensky said "preliminary data from a collection of states" suggested that 99.5% of deaths from Covid-19 occurred in unvaccinated people. Data from 13 states will soon be published in the CDC's Morbidity and Mortality Weekly Report, according to an agency spokesman.
On July 27, the CDC updated its mask guidance, recommending masks for vaccinated people in certain indoor situations. Some of the data behind that recommendation -- a study of a Covid-19 outbreak in Barnstable County, Massachusetts -- was published in the CDC's Morbidity and Mortality Weekly Report three days later.
At an August 18 White House briefing, Walensky mentioned the results of a study of 4,000 health care and other essential and frontline workers that suggested vaccine efficacy waned with time. That data was published in the agency's Morbidity and Mortality Weekly Report on August 24.
Getting data out earlier has involved a cultural shift within the agency, CDC sources said.
The agency's epidemiologists, physicians and other scientists are conservative and precise by nature, they said -- their professional training has taught them to dot every "i" and cross every "t" before going public with data.
"We're a bunch of scientists and statisticians who will argue about a colon versus a semi-colon or this table versus that table," said a longtime CDC physician who is involved in the agency's Covid-19 response. "We need to be more forward thinking, and sometimes it's hard to get them out of the weeds and see the big picture."
Changing this culture in the middle of a pandemic has taken some work, the CDC sources said.
"There's a lot of caution at the CDC and concern that we might be changing the way we're doing things and might end up releasing something that wasn't properly vetted," the physician said. "But there's an equal or greater number of people who recognize we have a responsibility to publish or share what the public knows as soon as we can."
It's also been a culture change with external academic researchers whom the CDC often relies on to help gather and analyze data, the CDC sources said.
Those researchers strive to publish in prestigious medical journals -- their professional success often depends on it -- and the researchers know that pre-pandemic, those journals might refuse to publish data that was already discussed elsewhere.
Walensky described her talks with academic researchers about releasing their data early as "hard conversations" because discussing data pre-publication "breaks norms" that "had been inculcated for decades."
But she said there has been "little pushback" as researchers feel "properly credited" for their work.
Another physician working on the Covid-19 response said even though academic researchers might be concerned about early release of data, they've taken pride as they've watched their work mentioned at White House briefings and put into practice.
"There's a lot to be said for the immediate feedback of 'I made a difference,' " said another CDC physician who is working on the Covid-19 response. " 'This work I was doing - the tedious work me and my colleagues were doing -- that really made a difference.'"

CNN's Lauren Mascarenhas and Virginia Langmaid contributed to this story.