Published on
The Trump administration quietly altered massive US health datasets to remove any mention of gender, according to a new analysis.
Scientists around the world rely on health data from US government agencies like the Centers for Disease Control and Prevention (CDC). But since President Donald Trump took office in January, these agencies have been altering data labels, mostly without disclosure, according to the investigation published in The Lancet, a leading medical journal.
The authors argue that the “hidden” changes could cast doubt on the integrity of American science.
The database edits “might be politically motivated and [are] not transparent,” Dr Aaron Kesselheim, a professor at Harvard Medical School, and Janet Freilich, a law professor at Boston University, said in the paper.
Their analysis included 232 US health datasets that were edited from late January to late March, covering veterans’ health care use, global tobacco consumption, stroke deaths, nutrition, exercise, and obesity. It did not include datasets that are updated on a regular basis.
About half of these datasets were “substantially altered” during the two-month period, but only 15 acknowledged that changes had been made, the study found.
The parts of the datasets meant for these disclosures remained empty as of early May, the researchers said.
In most of the edited datasets – 106 of them – the word “gender” was swapped out for “sex”.
US agencies have not confirmed or explained the changes, but Freilich and Kesselheim suspect they were made to comply with a Trump order to remove “messages that promote or otherwise inculcate gender ideology”.
The executive order, issued on Trump’s first day in office, said it aimed to restore “biological truth” to US agencies, and set a national policy “to recognise two sexes, male and female,” they said.
It was part of a flurry of activity to clamp down on LGBTQ protections. Some data was initially removed from federal websites, but was largely restored due to a court order.
The Lancet analysis indicates that not all changes were reversed. It is not clear if the actual numbers were altered, or only the labels about gender or sex.
But Freilich and Kesselheim said the distinction is important because some people will respond to surveys differently based on each term. That can undermine the quality of the data and lead researchers to draw inaccurate conclusions, they said.
They described the undisclosed changes as a “crisis” that could make US data “untrustworthy and unusable”.
They called on governments elsewhere in the world to invest in other data sources.
