Blaming “data coding errors,” the Journal of Occupational and Environmental Health has pulled an article linking shift work, age and sleeping problems.
The study was published four months ago, but managed in its brief lifespan to garner significant attention in the mainstream media and the blogosphere, although it has not been cited by any other papers. It comes alongside growing interest in the potential lnks between shift work and various health conditions including irritable bowel syndrome and breast cancer. Denmark even awards damages to shift workers who have developed the latter.
Ironically, the researchers, led by Philip Tucker, of Swansea University in Wales, U.K., had hoped to demonstrate the toll of shift work that previous studies were unable to show conclusively because of “methodological difficulties”:
We are … able to examine sleep complaints at a given point in time and to compare them with data previously recorded concerning both the workers’ initial sleep quality and their experience of shift work between baseline and follow-up. As far as we are aware no previous study has examined sleep complaints using longitudinal data from different age cohorts and has been able to distinguish between former shiftworkers, current shiftworkers, and workers who have never worked on shift. The primary aim of this study was to determine whether poor sleep quality is a consequence of shift work or a cause of intolerance to shift work.
Or maybe not.
Tucker’s group claimed that their results, generated through various comparisons of shift worker and employees on more regular schedules to themselves and each other
suggest that sleep problems are worse for shiftworkers, particularly in the early and middle years of working life. In particular, shiftworkers are more likely to suffer from premature awakening. Giving up shift work off-sets the age-related increase in sleep problems, with a net effect that sleep problems stabilize following the cessation of shift work. In particular, it appears that the problems of premature awakening are reduced as a result of giving up shift work. If giving up shift work results in a concomitant amelioration of associated sleep problems, this suggests that sleep problems are a reversible consequence of shift work rather than a cause of shift work intolerance.
Shift workers, especially those in their 30s and 40s, were several times more likely to report trouble falling asleep, staying asleep, using sleeping medication and other sleep-related problems.
Tucker tells Retraction Watch that he discovered the problem while querying his data set for a subsequent analysis:
It is a very large dataset that covers a very broad range of health-related issues. I identified inconsistencies between a some of the variables that were used to categorize participants. This led to my uncovering the coding errors associated with one of these variables.
Tucker says he agrees with the view that the glitch was fatal to the paper as written, but it did not affect the integrity of the data, or future papers that emerge from the set. His group has resubmitted their article for publication with the corrected analysis, he adds.
The journal may have acted quickly to retract the paper, but not before significant damage was done. The American College of Occupational and Environmental Medicine, which publishes JOEM, touted the study in an April press release (which is still available on the society’s web site).
The college also accredited the article for continuing medical education credits, or CME, which physicians must accumulate to maintain their licenses to practice medicine. Does the retraction mean doctors who obtained CME credits for reading the article have to forfeit them?
According to Murray Kopelow, chief executive of the Accreditation Council for Continuing Medical Education, which oversees CME in the United States:
The ACCME does not have jurisdiction over credit. Credit systems have been developed by physician membership organizations (for example, the American Osteopathic Association, the American Medical Association, the American Academy of Family Physicians). It is not normally recommended that credit should be withdrawn as that penalizes the learner who participated, in good faith, in the activity.
However, the ACCME has recently deliberated about whether providers should have a responsibility to inform learners, should an educational activity be noncompliant with our accreditation requirements.
Representatives for the occupational medicine society have yet to tell us whether they intend to honor the CME credits. [See updates below.]
Meanwhile, there’s another wrinkle: The pharmaceutical industry has been keen on targeting sleeping problems tied to shift work — an increasingly common labor pattern worldwide. Shortly after the JOEM article appeared, the drug maker Cephalon launched an ad campaign promoting awareness of shift work sleep disorder, or SWSD. (The company also has a web site that discusses the “condition” — quotation marks added for skepticism.) Cephalon makes the stimulants Provigil and Nuvigil, neither of which are approved for sleeping problems but are often taken to combat fatigue.
Update, noon Eastern, 8/13/10: Marianne Dreger, ACOEM’s director of publications, says her group has pulled the online CME test associated with the article and is trying to figure out how many people received credit for taking the exam. ACOEM has yet to decide if it will honor the credits, Dreger says, adding:
This has never happened to us before.
Update, 2:30 p.m. Eastern, 8/13/10: Dreger said ACOEM is revoking the credit because the questions were based on “faulty conclusions.” The good news, however, is that fewer than 10 people took the test, she says. The society will be notifying them shortly.