Scientific journals retract more articles, expose fraud and incompetence

Douglas W. Jackson

An integral part of scientific work and publications is based on truths. We have tended too much in the past to rely on the fact that the scientific journal peer review process could protect us from fraudulent and incompetent work.

I first learned this early in my career. I incorporated into my practice the results of an article published in a well-respected peer reviewed orthopedic journal. During my training, this journal had been a good resource and basis for much of our adopted patient treatments. However, this article was the first time that I questioned the accuracy of an article published in a first-rate scientific journal. It was published in the 1970s and described a new reconstructive knee procedure. It was the lead article in the issue, which reported the beneficial results of improving a common knee instability problem.

After reading the article several times and talking to the author in detail, I performed some reconstructions on my patients. Less than a year later, I realized the knees that had this procedure were actually more unstable. I checked with colleagues around the United States, and they were also experiencing the same thing with the procedure. Here was a lead article in a major journal and the procedure was obsolete in the hands of qualified surgeons less than 2 years after publication. The article was never retracted.

I vowed from that point that I would be more skeptical of incorporating new procedures based on an article in any journal. In retrospect, I should have questioned the surgical premises in that article that weakened remaining intact structures, while not addressing the primary ACL deficiency itself. The bad experience made me not only question printed articles more in terms of their controls, documentation and adequacy of follow-up, but also look more carefully for author bias. In addition, if I tried a new procedure described by others surgeons, I would observe the surgeon do the procedure and examine the surgeon’s patients and make sure other surgeons duplicated the published results.

Extent of the problem

We have all read reports recently in the lay press about conflicts of interest and the potential impact on published articles. More than ever before, authors and academicians have been challenged on the findings that support a specific drug or treatment regime, device or prosthesis. In particular, investigators with personal and financial interests are specifically questioned and must sign for or against any potential conflicts of interest. The relevant financial interests are the most obvious and easily ascertained. However, other conflicts of interest identified in some retracted or discredited articles are more subtle. Sometimes, it is simply incompetence or involves willful fraud on the part of the researchers who are publishing to advance their careers or prestige.

The number of retractions of published studies in leading scientific journals has increased significantly during the past 10 years. The Wall Street Journal by Thomson Reuters Web of Science, an index of 11,600 peer reviewed journals, recently reported that since 2001, the number of papers published in research journals has increased 44% and the number of published retracted articles also increased more than 15-fold. Twenty-two retraction notices appeared in 2001, but there were 139 notices in 2006 and 339 notices last year. In the first 7 months of 2011, 210 retraction notices were posted. As such, Retraction Watch, a blog to track discredited articles, has evolved.

Highly publicized article retractions

The Lancet retracted one study 7 years after publication based on serious challenges to the validity of the reported findings. The authors reported that two widely used high blood pressure medications were better in combination than either used alone. This study changed many physicians’ prescribing habits. Its proposed dual therapy was used on more than 100,000 patients and exposed some patients to threatening consequences.

In another alarming case, The Lancet retracted a study linking autism and the measles-mumps-rubella vaccine. That article was retracted 12 years after publication; however, it had already changed medical care by reinforcing beliefs and suspicions.

Another example is the recent partial retraction of a chronic fatigue syndrome study. This 2009 article reported a possible link between the retrovirus XMRV and chronic fatigue syndrome. Partially based upon this paper, which the results have not been reproduced, the American Red Cross prohibits blood donations from people who report they have chronic fatigue syndrome. Several investigators have seriously challenged the work, and the paper has been partially retracted online in Science. All 13 authors of the original work signed the partial retraction, and two authors took the additional steps of attempting to have their names removed from the original paper. Science is still determining if further retraction is warranted.

Lesson learned

While retractions of tainted papers are absolutely necessary, once they have been published, it is a waste of precious resources. More significantly, it can negatively impact patient care. Once a study is published and then retracted years later, it is difficult to reverse its effects as many individuals do not see the retraction and persist based upon the findings.

Even after the laborious efforts of proving a study invalid, almost routinely, the investigators stand by their results. When the accusations are eventually proven true, the retraction discredits the investigators and often ends their credibility. A major problem in the retraction process is how long it takes from the time of publication until a public retraction is made. The initial challenges are made, and the formal review process is long and arduous in itself. Studies that are formally questioned often will be referred back for departmental, institutional, journal and impartial outside reviews. As a consequence, it involves evaluating the quality of the initial institutional and departmental supervision of the research. The review process for the research often questions the local IRB as well as the institutional supervision of the research. Many egos and reputations are challenged along the way. It is often difficult to prove something is forged or is the result of incompetency.

As a result, all good journals work harder to prevent publishing articles that may need retraction later. While prevention is the obvious solution, it is more difficult than it sounds.

Reasons for the problem

Scientific journals feel they deserve credit for raising the issue and identifying some of the retracted work. This is, in part, because software and reviewers are much more thorough in investigating suspicious data. However, in the retracted cases, it is usually other investigators, individual physicians and scientific organizations that contact the journal and insist and persist in demanding an investigation and even a retraction.

The stakes are high for all scientific researchers as they compete for funding. Publishing articles in prestigious journals is a necessary ingredient to advance careers and peer recognition. Not only can having an outstanding article published in a prestigious journal be a path to success, it also can be the difference to becoming a tenured professor, chairman or other academic appointments.

There are some potential conflicts of interest among journals themselves. They have pressure to publish current and innovative research, and often compete for top articles. The journals feel the pressure to publish what may become classic articles that are widely referenced. Good articles help maintain readership. Journals often send press releases on cutting edge articles to garner recognition and wide-spread media coverage as an authoritative source. As a result of recent retractions, all journals are becoming more defensive.

However, it is my opinion, that journals have made much progress. The institutions where the research is performed still do not have adequate systems in place to properly investigate, evaluate and discipline fraud, incompetence and misconduct. While journals do a reasonable job in prevention, we would benefit from more efforts and resources to obtain a rapid assessment of when an article is formally challenged.

• Douglas W. Jackson, MD, is Chief Medical Editor of Orthopedics Today. He can be reached at Orthopedics Today, 6900 Grove Road, Thorofare, NJ 08086; email: OT@slackinc.com.