An article appeared in The New York Times a couple of weeks ago regarding a Dr. Sam Yoon, a surgical colleague who was found to have falsified data in publications going back 15 years. These papers were retracted. The research was conducted when he was at another institution and, at least in part, he was recruited to a senior leadership position at Columbia on the basis of the prestige these results generated. I have not heard his name mentioned since then so I suspect he will not reappear. It was suggested that he could be subject to criminal charges as well since he received and utilized $5 million in federal funds in the conduct of these fraudulent studies. There have been a spate of such reports concerning fraud in research recently, including allegations about 31 papers from the Dana-Farber Cancer Institute by four leading professors including Laurie Glimcher, DFCI’s president. (I myself have collaborated with one of the four.) In almost all the cases, including Yoon, they involve errors with gels or scans or other images that are being manipulated or replicated in multiple publications inappropriately.
It is no surprise to learn that cheating goes on. At least as far back as Gregor Mendel (a monk no less), fudging of data in scientific experiments has occasionally occurred. And is the rest of the world pure? We are inundated with tales of the misconduct of politicians, businessmen, real estate moguls, physicians and those in virtually every walk of life. The books of Jeremiah, Amos, Hoshea and others remind us that miscreants were not born yesterday.
Has anything really changed lately? I am skeptical about it. Let us consider some data. Most studies suggest that over 50% and perhaps as many as 70% of high school students cheat. This is good preparation for their college careers, where similar figures are prevalent.
It might surprise you to learn that over 1.8 million papers are published annually in biology/medicine. That is difficult to comprehend. Of these, about 1 million find their way into Pubmed, the online catalog of the National Institutes of Health’s National Library of Medicine. If they do not achieve this, they are unlikely to gain much traction with the scientific community. Something over half are published in English and the remainder in the native tongue of the scientists, again leading to great effects on the visibility that a paper will achieve. About 40% are never cited, meaning that no one reads them—maybe the author and his/her mother. But the median number of citations for papers overall is about 25, meaning that most papers do get read and considered fairly widely. Thus, cheating and falsification in a publication could theoretically carry some consequences.
The chief mechanism intended to prevent such errors is peer-review whereby other scientists review each manuscript submitted to journals before acceptance for publication. This is generally a pretty good system and does catch significant flaws in methodology and writing. However, it cannot possibly discern the submission of false or fake data, or errors in statistical analysis of many types. Thus, peer-review can only have a limited impact on cheating.
Overall, approximately 10,000 papers are retracted annually. This includes papers that had honest errors (about 40%), 17% due to redundant publication, 15% for other reasons, and 28% for research misconduct.
Another variable in this is the number of co-authors on a paper. The median number has been rising over the past few years and now stands around seven to eight. The import of this for our discussion is that it increases the number of individuals who are available to be cheaters. It is doubtless frequently the case that some postdoc has falsified data or cut a corner or done something untoward of which the senior scientist on the research is totally unaware. This happened to an old friend of mine at an Ivy League medical school, and when it became public, several papers had to be retracted, greatly embarrassing my colleague despite the fact that everyone knew he had no role whatsoever in the deception.
Most papers contain a statement, “The findings of this paper will need to be replicated.” This is a cliché that does not always lead to replication. In circumstances when experiments with positive findings have been repeated, studies show that approximately 70% do not replicate the original findings. Whether this means the first study or the second study was not done well, the second study differed in some way from the first, statistical artifact played a role, or there truly was dishonesty, who can say?
One saving grace in all of this is that clinical trials are conducted under much more stringent rules than other types of experiments. The FDA, the drug company, the institutional review board and others are motivated to ensure that errors and fraud do not occur. Even with all these safeguards, fraud still does occur on occasion, but it is rare. Thus at least human trials are generally safeguarded against the types of cheating we have been discussing.
I end with my recollection of rounds on a patient some years ago. After his presentation, I asked the third-year medical student what the rectal exam showed; he replied that it was normal. Later that day, when I examined the patient myself, there was a tumor the size of a baby’s head growing out of the rectum. I think we have a long way to go with honesty.
Alfred I. Neugut, MD, PhD, is a medical oncologist and cancer epidemiologist at Columbia University Irving Medical Center/New York Presbyterian and Mailman School of Public Health in New York. Email: [email protected].
This article is for educational purposes only and is not intended to be a substitute for professional medical advice, diagnosis, or treatment, and does not constitute medical or other professional advice. Always seek the advice of your qualified health provider with any questions you may have regarding a medical condition or treatment.
