Researchers warn fabricated citations are increasingly appearing in scientific papers and influencing clinical research

A worrying trend is emerging in the world of medical research: fabricated references in scientific papers are rising sharply, according to a new research letter published in The Lancet.

The study, led by Maxim Topaz from Columbia University, found that thousands of medical papers now contain references to studies that simply do not exist.

Researchers developed an automated system to verify references in medical papers published between January 2023 and February 2026. The team scanned more than 2.4 million research papers and reviewed over 125 million references from PubMed Central’s Open Access database.

Out of the 97.1 million references that could be checked using PubMed identifiers (PMIDs), researchers discovered 4,046 fabricated references spread across 2,810 papers.

The findings reveal a sharp increase in the problem over just a few years.

In 2023, roughly one in every 2,828 papers contained at least one fake reference. By 2025, that number jumped to one in every 458 papers. In the first few weeks of 2026, the figure worsened further to one in every 277 papers.

Researchers say the growing trend could seriously damage trust in medical science and potentially affect patient care.

“A medical professional or clinical guideline developer has no way of knowing that the evidence they are relying on does not exist,” Topaz said in a statement.

In one particularly alarming example, researchers found a paper in which 18 out of 30 references were fake.

Even more concerning, some fabricated citations are already being used in other studies and systematic reviews — research that often helps shape treatment guidelines and clinical decisions.

The study highlights increasing concerns about research integrity in the era of artificial intelligence and automated writing tools, which experts fear may be contributing to the rise in false or invented citations.

Researchers say stronger verification systems and stricter editorial checks are urgently needed to protect the credibility of scientific publishing.