Erasmus Medical Center in the Netherlands has released the final report on Don Poldermans, finding that the institution’s once-star cardiology researcher committed misconduct.
As Larry Husten of CardioBrief reports, Erasmus MC’s investigation included that Poldermans — perhaps best known for his work on beta-blockers — has admitted to acts of misconduct but not to fraud.
That’s not merely a distinction without a difference. Poldermans, for instance, claims that the suspect gaps in his data records, which Erasmus MC calls “serious shortcomings” in fact reflect losses due to water damage and files the university itself trashed.
We may never know the truth here, but one thing on which both sides agree is that no patients were harmed by the misconduct.
It seems unusual to us, given the findings — perhaps worthy of our own Expression of Concern? — that none of Poldermans’ studies seem to have been retracted, nor even subject to Expressions of Concern. As we noted at in an earlier post:
Sixteen of his papers have been cited at least 100 times, according to Thomson Scientific’s Web of Knoweldge, and one, in the European Heart Journal, has been cited more than 700.
Adam– thanks for the mention. The full report does say that several editors should be notified of the committee’s findings. I hope to elaborate on this in my followup piece. You should also know that coverage of this story may become a whole lot easier with the publication of an English translation of the Erasmus MC report.
A very brief summary:
The committee looked at 5 projects, and specifically at 7 papers and a PhD dissertation chapter.
For three of the projects (Decrease IV, V and VI), the committee recommended notifying the journals involved, on the basis that it “cannot guarantee the reliability of the findings in the publication and the validity of the conclusions”. This is based on the following criteria:
– in- and exclusion criteria of patients as reported in the papers were not applied in reality (Decrease IV, V)
– patients were included without their being informed, there are no informed consent forms, or patients have been informed insufficiently or incorrectly when asking informed consent (Decrease V)
– absent or inadequate documentation of essential study parameters (IV, V)
– inadequate documentation of treatment, including deviation from the treatment reported in the papers, and patients being left out of groups to which they had been assigned in an ‘intention-to-treat’ analysis (IV, V)
– results are unreliable, in absence, inadequate or constant changing of definitions (IV, V, VI)
– source documentation (electronic patient records), study case record form, study database and paper do not agree on essential points (IV, V, VI)
This amounts to five papers, and to me it basically sounds like a recommendation that the board of Board of Directors request a retraction.
The contents of the PhD dissertation chapter have not yet been published in a journal. No irregularities were found, but the committee recommends additional cross-checking with the electronic patient records before submission of the manuscript.
Regarding the final project, the committee called into question the integrity of a database on which the two final papers were based. The committee concluded that it could establish neither correctness or incorrectness of data on causes of death in the database, in the absence of general practitioner’s declarations. It called this absence “regrettable but not an infraction”, in light of the applicable laws and the fact that the database was built over 10 years ago. The committee refrained from making any recommendation.
Overall, the committee is not sure whether further follow-up is worth the effort without additional evidence, given the large number of contradictory statements provided to them by the scientists involved.
Okay, maybe not such a very brief summary…
Let me guess…none of his papers said anything controversial. Especially the 700er.
The full reaction of Poldermans can be found over here (.doc, in Dutch):
http://www.artsennet.nl/web/file?uuid=017961f5-269b-441d-97eb-67b6136aef5f&owner=06a113d8-63a3-
4936-9f1a-a1fdb2a33334
Hope this link works better:
http://www.artsennet.nl/web/file?uuid=017961f5-269b-441d-97eb-67b6136aef5f&owner=06a113d8-63a3-4936-9f1a-a1fdb2a33334
Thanks to @Edwin van Wijk. I hope that, one day, we will see a translation of Dr. Poldermans’ response, as well. Fortunately, Google Translate does a good enough job. The original Erasmus report sounded like so much bureaucratic babble, and the final report isn’t much better. By contrast, Dr. Poldermans’ response is well written and fairly convincing. Someone who knows the field should comment, and perhaps I should not. Still, the whole business has the ugly feel of small-minded nit-picking.
The reason why non-specialists should care about Poldermans, and might find this incident disturbing, is that we have to find the right balance between (a) encouraging clinical studies which have transparent data and methods and (b) avoiding defensive, complex, over-papering designed to avoid procedural criticism while making scientific analysis more difficult. The PI for a clinical study should not routinely have to hire a lawyer — neither before the study, to walk him through some Byzantine labyrinth of paperwork, nor afterwards, to spin the results for arbitrarily-invoked quasi-judicial tribunals. There’s no net gain for science if we simply replace every bad scientist with two good lawyers.