You’re reading The Checkup With Dr. Wen, a newsletter on how to navigate medical and public health challenges. Click here to get the full newsletter in your inbox, including answers to reader questions and a summary of new scientific research.
About a year ago, Boston University law professor Janet Freilich and her colleagues began noticing that public health data resources maintained by the Centers for Disease Control and Prevention were disappearing or being significantly altered. So they decided to start tracking them.
I cited some of their research in my column this week, which explored the impact that upheaval at the CDC is having on local health departments. But that doesn’t capture just how extensive these interruptions to federal data have been, as Freilich explained in an interview. So I wanted to return to the topic here to give readers a better sense of the real-life consequences of the administration’s decisions.
By the end of March, Freilich and her co-authors pulled together a paper analyzing a three-week period in January and February 2025, during which the CDC removed more than 200 datasets, roughly 13 percent of its collections available online. Some were later restored, but only after specific terminology was changed, such as replacing “gender” with “sex.” Others halted data collection altogether, including a program tracking the health of mothers and their babies, which researchers had widely used to study maternal mortality.
In another research article, published last July in the Lancet, she and another researcher examined databases from the CDC and the Departments of Health and Human Services and Veterans Affairs that were modified between January and March 2025. They found that 49 percent of the datasets they reviewed had been substantially changed. Yet only 13 percent of those altered datasets included documentation noting that a modification had occurred.
Those revisions matter. Researchers across the country rely on this information to evaluate health policies and outcomes, and unacknowledged changes can alter how findings are interpreted. Moreover, when datasets are secretly revised without a record of what was changed and why, it undermines confidence in the integrity of the information. Transparency is foundational to trust.
Freilich told me that this line of research led her to examine CDC databases that were supposed to be updated weekly or monthly but had not been refreshed for months. What stood out was a clear pattern: Eighty-sevenpercent of the databases that were paused without explanation from May to October of last year were related to vaccination.
“It was really shocking,” she said. Although some of the databases have since resumed updates, others remain dormant, complicating efforts to monitor infectious diseases and assess levels of protection among vulnerable populations.
This poses real problems for local health departments, which must decide where to focus vaccination campaigns based on accurate, up-to-date information about who’s getting the shots (or not). Clinicians also need to understand the prevalence of vaccination in their area to assess risk. “If you don’t have information, you don’t know where the gaps are,” Freilich said.
Researchers, too, rely heavily on federal data, especially when tracking fast-moving infectious diseases. State and local agencies are usually the entities collecting information, but only the federal government can coordinate efforts to aggregate it nationally, standardize data and make analyses broadly accessible.
Freilich commended efforts by nongovernmental groups that are trying to safeguard federal data before it disappears. She pointed to efforts by institutions such as the Harvard Law Library and other archival initiatives that have downloaded and preserved public datasets.
“The challenge is that it’s one thing to preserve, but it’s another to deal with ongoing data collection,” she said. “I don’t think that anyone, other than the federal government, can realistically recreate many of these databases.”
Of course, the erosion is not confined to health data. Other targets include environmental and climate information, with many implications for communities trying to prepare for extreme weather, monitor pollution and assess risks to human and animal health.
Climate.gov, a widely used site maintained by the National Oceanic and Atmospheric Administration, stopped publishing new content last summer. It had provided updates on changing weather patterns, drought conditions and greenhouse gas emissions. The nonprofit Environmental Data and Governance Initiative reports that the administration made more than 630 significant changes to federal environmental websites in its first 100 days, a pace that far exceeds that of the first Trump administration. While groups such as the National Security Archive and other watchdog organizations are racing to preserve removed material, static snapshots cannot replace continuously updated, functioning data systems.
“I think that we underestimate the extent to which federal government data run our lives,” Freilich said. “If the data environment changes, which it has been, it has the potential to cause a lot of chaos and a lot of disruption.”
In response to multiple queries about the missing and altered datasets, an HHS spokesperson said the changes reflected “routine data quality and system management decisions.” The individual added, “Under this administration, public health data reporting is driven by scientific integrity, transparency, and accuracy.” The irony here is hard to miss.
The post Why it matters that the government is messing with health databases appeared first on Washington Post.
