Health datasets represent 40% of affected federal collections
A federal database of law enforcement officer misconduct records, a CDC survey measuring violence against children and youth, and real-time pollution readings collected by air quality monitors from US embassies around the world are among hundreds of federal datasets the Trump administration has taken down or altered as part of what The Guardian characterized as an aggressive campaign to wipe data findings that do not align with the administration’s priorities.
The dataindex.us tracker, which was released last week, includes detailed information about deleted or altered datasets from more than 60 federal agencies spanning the environment, labor, law enforcement, and other topics. The Guardian reported the findings on Tuesday.
Christopher Dick, a cofounder of dataindex.us and former chief of the population evaluation, analysis and projections branch at the Census Bureau, said the group launched the online tracker because it is difficult to know exactly how many federal datasets have actually been deleted. “The question we’re trying to answer is: are certain datasets being taken away that have an impact on the American people?” Dick said.
Denice Ross, the former US chief data scientist under the Biden administration and a dataindex.us cofounder, said federal data infrastructure is at a critical juncture. “We’re at a very vulnerable point as a nation,” Ross said. “We’re facing a future where we are not going to have the information that we need to make decisions that will save lives and benefit American communities.”
Health datasets have absorbed the largest share of the deletions and alterations, comprising about 40% of the total number of terminated or altered data collections tracked by dataindex.us. One of the main documented changes is that these datasets no longer collect information on queer and trans people — a shift the tracker attributes to President Donald Trump’s executive order “Defending Women from Gender Ideology Extremism and Restoring Biological Truth to the Federal Government.”
Multiple CDC surveys have removed questions about respondents’ gender identity. The affected systems include the National HIV Surveillance System, the primary source of case data for HIV infections; the National Violent Death Reporting System; the CDC Diabetes Prevention Recognition Program; the National Intimate Partner and Sexual Violence Survey; and specific surveys evaluating PrEP uptake among gay men and in communities of color.
At least four health datasets are no longer being updated at all. One example is the Substance Abuse and Mental Health Services Administration’s Drug Abuse Warning Network, a surveillance system that doctors and public health researchers had used to track drug-related emergency room visits in near real time to spot emerging overdose trends.
Thomas Babor, a professor emeritus at the University of Connecticut who specializes in public health and addiction, said the Drug Abuse Warning Network and similar tools are essential for identifying health threats. “These data are critical for identifying emerging epidemics and what parts of the population are most vulnerable to different types of health conditions,” Babor said. “We need to know what direction the trend is going, whether it’s the beginning of an end; we need to learn from prior episodes of epidemics and how they were responded to and what impact the responses had.”
Babor described the takedowns as a loss of institutional memory. “Taking down, even temporarily, some of the most important surveys that we have, especially for comparative purposes, is like we’re erasing our institutional memory for public health,” he said.
Other deleted datasets include the US Department of Agriculture’s twice-annual survey of farm employers, which had counted the number of farmworkers employed and their wages and informed the minimum pay rates for H-2A visa workers, and a National Center for Education Statistics survey of the pay and working conditions of teachers and principals at public and private schools across the country.
Dataindex.us identified the deleted and altered datasets through a combination of research on dataset usage, crowdsourcing from media reports, government employees, and dataset users, monitoring social media, and tracking notices posted about proposed regulations from the Office of Information and Regulatory Affairs.
Government agencies are not required to notify the public when data collections are terminated, according to Ross and Dick. Some datasets have stopped updating because of a lack of staffing, while others have been quietly pulled from federal websites with little notice. Some federal datasets also change or update on a regular basis, making it difficult to determine what is permanently deleted versus a routine update.
Ross said the dataindex.us tracker is likely an undercount. “We want people who depend on different datasets, some of which might be sort of obscure, to let us know what corners to look in because we don’t have full visibility,” she said.