Dr. Rochelle Walensky Knows the CDC Made ‘Dramatic Mistakes.’ Now She’s Trying to Fix Them

The U.S. Centers for Disease Control and Prevention (CDC) was not prepared for COVID-19. After more than two years, it still isn’t. The CDC’s response to COVID-19 has been widely criticized as slow, confusing, and mostly ineffective.

Now, the agency is taking a long, hard look at itself. On Aug. 17, CDC director Dr. Rochelle Walensky proposed sweeping changes in how the agency communicates with Americans and publishes data—two of its most critical roles as the nation’s leading public-health agency.

“I don’t think moving boxes around on an organization chart will fix the problem,” she tells TIME of the changes, which she has already begun to implement. “What we’re talking about is a culture change. We’re talking about timeliness of data, communication of data, and policies guidance. Reorganization is hard, but I think this is even harder than that.”

The revamp has been months in the making. In April, just over a year after taking the reins, Walensky called for an agency-wide review of the CDC. While previous directors have ordered such reviews to assess the CDC’s operations, this particular analysis was especially urgent because of the pandemic and low trust in the CDC, after the Trump Administration sidelined the agency, ignored its advice, and at times contradicted its guidance. Walensky asked for honest feedback from nearly 200 employees, academics, and other outside experts.

Walensky says the review, which has not yet been made public, was sobering but unsurprising. “To be frank, we are responsible for some pretty dramatic and pretty public mistakes, from testing to data to communications,” she said in a video message to CDC employees, which TIME viewed.

Here’s what Walensky says went wrong—and how she plans to improve the CDC.

A need for nimbler data

The CDC “has been developed on an infrastructure of academia,” Walensky says. Until COVID-19 forced the agency into the spotlight, the CDC’s target audience was mostly other public-health experts and academics, and its main mode of communication was through periodically publishing scientific papers. “In these pandemic moments, we found ourselves having to talk to a broader audience,” Walensky says. “We didn’t have to convince the scientific audience—we had to convince the American people.”

Americans wanted timely, accurate information about how to deal with the new virus. But since the very start of the pandemic, the CDC’s advice has seemed confusing and often contradictory—especially around how the virus spreads, who should wear masks, and what types of face coverings are most effective. The agency was also slow in producing critical information about how contagious SARS-CoV-2 was. “We all didn’t like the headlines, especially when we knew all of the good work that was going on,” says Walensky about media coverage of the CDC’s missteps. “So how do we address the challenge of what people are saying about us?”

Walensky says she is now pushing for the CDC to collect and analyze data in a more streamlined way, in order to more quickly turn that information into practical advice. During COVID-19, researchers began relying more on pre-print servers, which published scientific studies on COVID-19 before the results were reviewed and vetted by experts (the gold standard for validating results). “The peer-review process generally makes papers better,” she says, “but it is also the case that if you’re trying to take public-health action with actionable data, then you don’t need the fine-tuning of peer review before you make [the results] public.”

She and her team are discussing ways to post data that would be relevant to the public earlier—not to replace the peer-review process, but to supplement it, so that both the public and health experts can see the evidence on which the agency is basing its recommendations. They are considering, for example, uploading the data onto a preprint server or publishing separate technical reports to distinguish early data from the final peer-reviewed product.

Currently, the agency’s advice is only official once it is published in the CDC’s publication, MMWR, which requires a relatively lengthy and involved peer-review process. During a public-health emergency, such data need to be made available more quickly, Walensky says. “I have called journal editors and said, ‘I know we have a paper under review, but the public needs to know, and I am going to break this embargo,'” she says.

That happened last July, when data from an indoor gathering in Barnstable, Mass. showed that vaccinated people were getting infected after mask policies were loosened; as a result of the findings, the CDC reinstated a recommendation to wear masks in large public environments before the study was published in MMWR. In another instance, CDC scientists had data on the effectiveness of vaccines under review for MMWR, but revealed the information before publication in a public meeting of vaccine experts convened by the U.S. Food and Drug Administration.

“We can’t be loose with the data,” she says. “But there needs to be something between dotting every I and crossing every T.”

Better, clearer messaging

Key to making such data more accessible is using clear, jargon-free language to convey it. In her video message to employees, she stressed that producing “plain language, easy-to-understand materials for the American people” would become a priority, along with making sure scientists develop talking points and FAQs.

They’ve already started putting this into practice, she says, pointing to the CDC’s revised Aug. 11 isolation recommendations. Compared to past guidance, the new version is written more for the public and addresses people’s practical concerns, such as when to start counting isolation days and which precautions to take in the home, she says.

From her perspective, the culture change Walenksy is hoping to implement boils down to one question that she is urging all CDC staff to consider: will the data they are analyzing, or the study they are conducting, or the advice they are generating, address a public-health need? “We really need to talk about public-health action, and not just public-health publications,” she says.

That won’t happen overnight, she acknowledges. But now that other viral diseases—including monkeypox and even polio—have joined COVID-19, the stakes are high for CDC to catch up fast. The agency continues to receive criticism from public-health experts, doctors, and the general public for repeating some of the same mistakes from COVID-19 in handling the monkeypox outbreak. Data on monkeypox cases are still too slow. “To this day, we have race and ethnicity data on less than 50% of monkeypox cases,” she says. “We’re still working on getting complete case report forms and still working on getting immunization data.” Testing for monkeypox was also not widely accessible for months—delays reminiscent of the early days of COVID-19—because the agency’s testing protocols were too long and inefficient to combat a rapidly spreading virus. But, Walensky says, “within a week of the first case, we were reaching out to commercial labs to expand testing capacity quickly.”

The changes she’s implementing won’t be immediately apparent to the public, but she’s confident they will eventually lead to clearer communication and faster data on emerging outbreaks. “People won’t wake up after Labor Day and think, everything is different,” she says. “We have a lot of work to do to get there.”