|
Abstract
|
Last month, Mandeep Mehra, Amit N. Patel, and Sapan Desai were riding high, with shared co-authorships on major new papers in The Lancet and The New England Journal of Medicine (NEJM) and an influential preprint. Drawing on what appeared to be a vast patient data trove from hospitals around the world, the papers delivered seemingly definitive news about whether already approved drugs were safe for COVID-19 patients, or effective against the disease.
Now, the two journal papers have been retracted, the preprint taken down, and Patel's academic affiliation severed. The three physician-scientists are under the microscope as a shocked scientific community evaluates what may be the first major episode of research fraud in the pandemic. The journals are receiving withering criticism for what some call a failure of editorial processes and peer review. The retractions are "unnerving and disturbing," says Leigh Turner, a bioethicist at the University of Minnesota, Twin Cities. The rush to publish on COVID-19 has exposed a lack of rigor that has reached "elite journals at the top of the academic pyramid," he says.
The retracted NEJM paper "had external peer review and statistical review, as well as scientific and manuscript editing," an NEJM spokesperson says. The Lancet did not comment on its review process. Neither journal notes submission or acceptance dates for papers, but a spokesperson for Mehra says reviews for each paper took about 1 month.
By publishing only author retraction statements, the journals "didn't show any self-reflection, any introspection," Turner says. To him, the case also raises a bigger question about how much access to key data each journal should require--and whether all co-authors should have full access to a data set. "The less access they have, the greater the chances that there will be errors, data fabrication, or outright fraud."
Publication of the Lancet paper abruptly halted many trials of hydroxychloroquine, the antimalarial touted by President Donald Trump, because of its finding that COVID-19 patients receiving the drug had a greater death rate than a control group (see p. 1166). The NEJM paper exonerated blood pressure drugs that some thought might worsen COVID-19, and the preprint found that mortality was dramatically reduced in COVID-19 patients receiving the parasite drug ivermectin, which drove huge demand for the medicine in Latin America (Science, 5 June, p. 1041).
Mehra, Patel, and Desai were the only scientists on more than one of the three papers, and all of the other co-authors are linked to at least one of the trio. After critics discovered anomalies in the data and wondered how Surgisphere, Desai's small company, could have amassed and analyzed tens of thousands of hospital records from around the world, the core authors promised independent data audits. But Surgisphere declined to make the firm's database and hospital agreements available, prompting the journal retractions. "We can no longer vouch for the veracity of the primary data sources," Mehra, Patel, and a third author wrote in the Lancet retraction.
The ivermectin study quietly vanished from the SSRN preprint server. "There's no retraction letter. But its ghost lives on in Latin America," says tropical disease physician Carlos Chaccour of the Barcelona Institute for Global Health, who, with colleagues, raised questions about the preprint. (African physicians who developed a COVID-19 severity rating system with Surgisphere's help withdrew the tool last week.)
Desai, Mehra, and Patel had never before published together, and that should have been a red flag to any journal, says Jerome Kassirer, editor-in-chief of NEJM during the 1990s. Co-authors of high-profile papers normally share subject area expertise or have clear professional ties, he says, calling the collaboration of the apparently disparate individuals "completely bizarre."
Prior to the retractions, Desai, a science-fiction writer, entrepreneur, and vascular surgeon, had defended Surgisphere and its database. Neither Mehra, a highly respected scientist at Harvard University and Brigham & Women's Hospital, nor Patel, a little known cardiac surgeon who recently resigned from an unpaid adjunct position at the University of Utah, has talked to the press. But Mehra apologized in a statement. "I did not do enough to ensure that the data source was appropriate for this use. For that, and for all the disruptions--both directly and indirectly--I am truly sorry."
As CEO of Surgisphere, Desai has received the most scrutiny. He started the company in 2007 as a medical resident at Duke University. It initially produced medical guides. In 2010, under the firm's auspices, he founded the Journal of Surgical Radiology, which folded in 2013. Its articles have been cited only 29 times, according to Scimago, a journal rating service. Yet an undated Surgisphere web page, no longer accessible, said the online-only publication had 50,000 subscribers and nearly 1 million page views monthly.
Surgisphere also claimed to have gathered and analyzed data on nearly 100,000 patients at some 700 hospitals worldwide. But no hospitals have acknowledged giving data to the firm. National Health Service Scotland, noted in a case study on the company's website, tells Science that none of its hospitals worked with Surgisphere. It will ask the firm to remove a website image of a Glasgow hospital...
Language: en |