.jpg){kind=link}
Irish eyes most definitely were not smiling on three papers that purported to contain data from a national repository from the Emerald Isle.
The articles, which appeared in a trio of journals from Dove Medical Press — part of Taylor & Francis — were written by various researchers at Nanchang University, in China.
Two of the articles have been retracted. “Serum Human Epididymal Protein 4 is Associated with Depressive Symptoms in Patients with Chronic Obstructive Pulmonary Disease,” from 2020, was published in the International Journal of Chronic Obstructive Pulmonary Disease. Per the abstract:
The data on 219 participants with COPD from The Irish Longitudinal Study on Ageing (TILDA) were analyzed for the association between serum HE4 levels and depressive symptoms, accounting for relevant confounding factors. All the COPD participants were prospectively followed up for a median period of 48 months. Cox proportional hazards analysis was used to evaluate the predictive value of serum HE4 for predicting depression events in these COPD patients.
Trouble is, the TILDA contains no such information.
As the retraction notice explains:
The Editor-in-chief and Publisher of International Journal of Chronic Obstructive Pulmonary Disease wish to retract the published article. Concerns were raised to us by the The Irish Longitudinal Study on Ageing (TILDA) investigation team. TILDA did not measure the biomarker human epididymis protein 4 (HE4) or disaggregate data on lung disease sufficiently to identify COPD as reported in the article. Thus, the observations and findings reported in the study were not supported by the TILDA dataset.
Despite several attempts, the authors could not be contacted for an explanation and the decision was made to retract the article.
Also from 2020 was “Increased Blood Lipid Level is Associated with Cancer-Specific Mortality and All-Cause Mortality in Patients with Colorectal Cancer (≥65 Years): A Population-Based Prospective Cohort Study,” which appeared in Risk Management and Healthcare Policy. Its retraction notice reads:
The Editor-in-chief and Publisher of Risk Management and Healthcare Policy wish to retract the published article. Following a review of the article by The Irish Longitudinal Study on Ageing (TILDA) investigation team, it was found the numbers reported in the study with colorectal cancer to be inaccurate, with a sample far higher than that reported with colorectal cancer in the TILDA public archive. Thus, the observations and findings reported in the study were not supported by the TILDA dataset.
A third paper, “Increased Levels of Serum Glycosylated Hemoglobin are Associated with Depressive Symptoms in a Population with Cancer (≥49 Years): An Antidepressant-Stratified Analysis,” published this year in Clinical Interventions in Aging, has been slated for retraction, too, according to Paul O’Mahoney, the operations manager for TILDA, based at Trinity College in Dublin.
O’Mahoney told us that of the three retractions, the COPD paper:
was the most flagrantly fraudulent in that it fabricated data (it reported on a biomarker which TILDA does not collect), and it was the main flag and catalyst to check the others; establishing that this and the other articles were in fact fraudulent was possible only for an internal team here with extensive knowledge of the dataset and processes for accessing it – the TILDA R&D Manager, Biobank Manager and Data Manager, and a postdoc who took the time to check the papers thoroughly and try to reproduce them. The other papers retracted this week did not fabricate data but misreported them in a variety of ways, e.g. inflation of underpowered sample sizes, unreproducible flow charts.
These are the first instances we’ve seen of this kind of misconduct. All articles flagged for review are published within the past 2 years. My inkling is that publicly available longitudinal datasets may be the most recently identified target of ‘paper mills’, offering the potential to write up articles to a template which outwardly are methodologically sound, but with adjustments made to present significant associations etc., where editors and peer reviewers could not possibly be expected to catch the malpractice.
O’Mahoney said that he also has notified other journals about several other questionable studies using TILDA data:
I initially flagged a set of articles for checking (the 3 retracted, 2 others in another, higher-ranked journal, the editors of which have been contacted but who have not responded yet, and a further 2 which may be legitimate but require review). It was an onerous process, as the articles use the public TILDA dataset, which is open access and accessible through two different channels in Dublin and Michigan, meaning publications based on it may not come on our radar for some time.
Like Retraction Watch? You can make a one-time tax-deductible contribution or a monthly tax-deductible donation to support our work, follow us on Twitter, like us on Facebook, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].