The founder of Surgisphere, the little-known health data analytics company blamed for the retraction of two prominent scientific papers on COVID-19, is in more trouble.
On Friday evening, Elisabeth Bik, a consultant who specializes in analyzing scientific papers for signs of data manipulation, spotted multiple duplications in images from a paper published by Sapan Desai in 2004, four years before he founded Surgisphere.
Manipulating images to change their scientific meaning, sometimes involving subtle duplications using the clone tools in Photoshop or similar software, is a major cause of scientific misconduct.
BuzzFeed News asked two other independent experts in scientific data manipulation to review the images. Both confirmed Bik’s findings, and one said it was one of the most egregious examples he had seen.
“It’s like the guy went crazy with Photoshop,” Daniel Acuna, a computer scientist at Syracuse University in New York, who has developed software to spot image duplications in scientific images, told BuzzFeed News.
Desai did not immediately return requests for comment.
The paper in question was published by the Journal of Neurophysiology in 2004, as part of Desai’s graduate studies at the University of Illinois at Chicago. The new concerns about its validity add to the growing number of questions about Desai and Surgisphere.
Desai was one of four researchers behind a massive study published on May 22 that linked hydroxychloroquine, the malaria drug hyped by President Donald Trump as a treatment for COVID-19, to an increased risk of death. The paper said the study was based on data on more than 96,000 COVID-19 patients from 671 hospitals across six contents provided by Surgisphere.
Other studies have also concluded that the drug isn’t effective against the coronavirus. But after the Lancet published the study about the drug’s potentially lethal effects, the World Health Organization paused its human trial of hydroxychloroquine.
The paper was retracted on June 4 after scientists and journalists raised questions about inconsistencies and demanded access to the study’s raw data. In the retraction notice, the three other members of the research team admitted that they could "no longer vouch for the veracity of the primary data sources.”
They wrote that they wanted outside experts to independently audit the data, but Surgisphere would not hand it over due to “client agreements and confidentiality agreements.”
Another study using Surgisphere data to study the coronavirus and cardiovascular disease, involving several of the same researchers, was also retracted on June 4 by the New England Journal of Medicine. This time Desai was an author of the retraction notice, which similarly explained that the scientists were “unable to validate the primary data sources.”
Surgisphere, which Desai founded in 2008 while a surgical resident at Duke University, is based in Palatine, a suburb of Chicago. Scientists have wondered how this little-known company has been able to obtain all the data it claims to have. The retractions followed stories about Surgisphere in the Guardian and the Scientist, including that the company only appeared to have a handful of employees and that raised serious questions about the legitimacy of its data sources.
Desai has previously stood by his company’s data. “Our strong privacy standards are a major reason that hospitals trust Surgisphere and we have been able to collect data from over 1,200 institutions across 46 countries,” he told BuzzFeed News by email last week, before the two studies were retracted.
The 2004 paper was on the anatomy of the vestibular system, part of the inner ear important for the sense of balance, in various rodents including rats, mice, and squirrels. The main panel of images in the paper shows images of sections of tissue from a structure called the crista ampullaris, which detects rotational movement.
Bik looked at Desai’s past papers after learning of the concerns about Surgisphere. “I immediately saw that there seemed to be very repetitive areas in these images,” Bik told BuzzFeed News. After spending a couple of hours finding multiple duplications within the images, she posted her concerns on PubPeer, a website where scientists critique one another’s work. Bik also published a blog post about her concerns on Saturday.
“This is highly unusual,” Bik said. “It’s a very, very severe case of duplication.”
Acuna, the computer scientist from Syracuse University, ran the images through his software, confirming many of the problems flagged by Bik. “I’ve never seen something like this, It’s outrageous,” he said.
“I concur with the allegation that there appear to be numerous small duplicated regions in the photographs,” Mike Rossner, a former managing editor of the Journal of Cell Biology who now runs a consultancy firm called Image Data Integrity, told BuzzFeed News by email.
The Journal of Neurophysiology said it would look into the matter, after BuzzFeed News reached out about the paper. “This has been referred to the journal’s ethics officer for investigation,” Bill Yates, a neuroscientist at the University of Pittsburgh and editor-in-chief of the Journal of Neurophysiology, said by email.
Anna Lysakowski, the professor who supervised Desai’s graduate work at the University of Illinois and was named as the senior author on the paper, did not immediately respond to a request for comment.