Retraction Watch

Tracking retractions as a window into the scientific process

Scientists investigated for misconduct lose appeal in suit against Harvard. Lawyers explain what it means.

with one comment

Paul S. Thaler

Paul S. Thaler

Richard Goldstein

Richard Goldstein

Retraction Watch readers may recall the case of Piero Anversa and Annarosa Leri, both formerly of Harvard and the Brigham & Women’s Hospital in Boston. The pair — which has had their work subjected to a retraction, expression of concern, and correction — sued their former employers in 2014 for costing them job offers after the institutions notified journals, triggering notices. A judge dismissed the case a year ago, saying that Anversa and Leri had to try other administrative remedies before bringing suit.

But Anversa and Leri appealed, and last week, a court denied that appeal. (See the judge’s decision — which begins by quoting Ecclesiastes and includes the delicious word “gallimaufry” — here.) We spoke by email to two attorneys — Richard Goldstein, who represented the scientist in Bois v. HHS, the first case to overturn a funding ban by the Office of Research Integrity (ORI), and Paul Thaler, who has represented scientists involved in misconduct proceedings for more than 25 years — about the case, and what it could mean for similar lawsuits.

Retraction Watch: The decision seems to stop Anversa and Leri from continuing their suit against Harvard and the Brigham, but also acknowledges some of the scientists’ concerns as legitimate. How would you summarize the findings and their implications?

Richard Goldstein: The federal appeals court indicated it would not allow lawsuits for money damages to go forward while the institutional/administrative process is on-going. In effect, the court said a respondent can sue for money, just not while the matter is still pending before the institution. It’s unclear whether they merely have to wait for the institution to finish its process or whether they also have to wait until ORI is done. If it’s the latter, that could add more years to the process. Because the Court accepted, for present purposes, violations may have occurred and the plaintiff shouldn’t be denied their ‘day in court,’ the respondent’s lawsuit was ‘stayed’ (i.e., put on hold) and was not dismissed.  However, given the slow pace at which the administrative proceedings in this case are moving and number of levels of agency review yet to transpire, it could be several more years before the lawsuit gets going again.

As a practical matter, because respondents (or others) won’t be able to pursue a lawsuit until the institutional/ORI process is over, scientists will be limited to complaining directly to ORI during the course of the misconduct hearings if they feel aggrieved by the institution.

Paul Thaler: The decision largely focuses on the process of the interaction between research misconduct inquiries and investigations by institutions and resulting complaints filed in court by aggrieved respondents.

The plaintiffs in their suit against Harvard complained about a number of issues resulting from the investigation by Harvard, primarily concerning the significant amount of time the proceedings were taking and the resulting impact on the respondents’ careers. This is a common complaint among respondents, and understandably so. Where allegations pop up against a scientist and his energies are directed to defending his career and reputation, time is extraordinarily important. Although the applicable federal regulations prescribe a fairly modest period of time in which the institutions may conduct their inquiries and, when needed, investigations, practitioners in this field know that ORI routinely grants extensions of time in which these matters may be looked into. This is needed largely because of the complexity of the field itself as well as the allowance for committees to broaden the scope of their investigations should the evidence lead to additional areas of concern.

While the Court of Appeals affirmed the district court’s decision to require that the plaintiffs exhaust their administrative remedies, the appellate court specifically instructed the lower court to modify its order to stay the proceedings, as opposed to dismissing the proceedings. Under the trial court’s original order dismissing the action, some of the plaintiffs’ claims would have been time-barred through no fault of their own. Thus, the plaintiffs’ claims are now preserved pending final administrative review. They will be able to pursue them in court – they just have to wait for the final decision in the administrative process.

RW: Did the court make the right decision, in your opinion?

Goldstein: As a legal matter, the court followed well established precedents in the area of administrative law; allowing a lawsuit of this type to proceed is the exception rather than the rule. The Courts try to protect an administrative agency from being dragged in to court at the drop of a hat. Therefore, it is hard to say the decision was clearly ‘right’ or ’wrong’ as a matter of law.

However, the court said the decision was “intensely practical” and was based on the characteristics of the particular administrative procedure.  The research misconduct process is an unusual one because it combines an initial investigation and finding by a non-government research institution with a subsequent investigation and finding by a government agency (ORI) followed by a formal hearing by an HHS ALJ. The court seemed to place a lot of trust in ORI’s ability and willingness to “probe an institution’s compliance with the regulations and take appropriate enforcement action.”  For people involved at the institutional level, regardless of whether they are the accused, the accuser, or a witness, the stakes can be very high. The institutional hearings and final decisions are very important to how the case is handled by ORI.  As a lawyer who represents scientists in institutional research misconduct processes all across the nation, it has been my experience these processes vary widely in terms of an institution’s adherence to ORI rules. The court’s decision contains an expectation ORI will monitor and, if necessary, ‘police’ the institutional process as it is happening.  The ‘correctness’ of the decision will depend on whether, going forward, research institutions and/or ORI exercise their administrative and oversight powers fairly and expeditiously.

Thaler: While I understand the logic of requiring an exhaustion of administrative remedies prior to filing suit in federal court, my practice tells me that for many respondents, the old adage of “Justice delayed is justice denied” rings true and accurate. If the system could be counted on to be a pristine model of adherence by all stakeholders, the appellate court would be correct – that the respondents in misconduct proceedings have nothing to lose by first going through that process before filing suit. Their claims for money damages would still be recoverable.

The problem arises when certain elements of the process get compromised. The most common mistake (intentional or not) is breaching the requirement of confidentiality throughout the process. The regulations provide that only those who “need to know” about the proceedings should be included in learning about the misconduct proceedings. That is for good reason – there are plenty of respondents who are eventually exonerated. If the misconduct proceedings were completely transparent, the reputations of good, honest scientists who have done nothing wrong would be unnecessarily sullied. While the regulations allow for a restoration of reputation requirement at the end of an investigation in which a scientist is exonerated, it is not a magic pill. Once word has leaked out that a scientist is under investigation, the resulting damage can be difficult to repair. Thus, it remains incumbent upon institutions conducting the inquiries and investigations to strictly protect the process.

Even more immediately damaging are the instances in which institutions begin taking negative employment actions against scientists employees when the matter has not been finally resolved by ORI. In those cases, I would view the Anversa case as troubling. Once a scientist is terminated, especially if it is done improperly, it is extraordinarily difficult for him to get another position – especially when the matter is still pending at ORI. And, as we know, ORI has been overburdened and cases are taking a long time to get resolved. A terminated scientist waiting for the administrative process to complete could be unemployed for years thereby making him potentially unemploy-able. The practical affect is potential financial devastation. Possibly by the time the administrative process is concluded and the scientist is procedurally able to bring a lawsuit, the scientist no longer has the financial means to do so.

RW: The court also acknowledges that the Harvard/Brigham investigation is taking a long time, but seems to suggest that is inevitable given the complicated nature of biomedical research. Do you agree?

Goldstein: Not necessarily. Many administrative processes are technical and specialized, so dealing with complicated issues is not unique to academic misconduct inquiries. Also, the scientists who comprise the panels should be familiar with research procedures and the current state of scientific knowledge. So, it shouldn’t take a long time just because it involves biomedical research. I’ve been involved in cases where the decisions are issued quickly. However, sometimes a case deserves a longer period of review than an institution is willing to allow.

Delay was just one of the problems alleged by the respondents.  The court’s decision suggests the hearings were delayed because the scope of investigation kept expanding but a lot of a factors may have contributed to the delay. There also seem to have been problems with the composition of the panel, the nature of the interim decisions they reached, and the lack of confidentiality.

Thaler: This particular case is not taking any longer than many cases going through the misconduct system. Unfortunately, they routinely take many months (sometimes years) longer than the regulations anticipate. The appellate court’s decision in the Anversa case supports letting the process run its course, no matter how long it takes, because it is complicated and deserving of careful analysis.

As the court stated in its opinion, “”Virtually by definition, research misconduct is a complicated area….” Indeed, the Court of Appeals opines that the scientific expertise which ORI brings to bear during the course of its oversight review could effect the “shape of the plaintiffs’ claims” in the subsequent litigation.

RW: The ruling does not seem to comment on one of Anversa and Leri’s allegations, that by notifying journals of the ongoing investigation, Harvard and the Brigham were violating Federal confidentiality statutes. Do you agree with this allegation, and if so, does that outweigh the need for journals to keep the literature as up-to-date as possible?

Goldstein: I can’t comment on whether notification of the journal in this case violated federal regulations.

How to tell the journal and the wider community about possible errors in a published paper is one of the many challenges facing institutions and authors in these matters. There is a definite tension between notifying journals and protecting the confidentiality of the misconduct investigation but, in my experience, it’s not always a zero sum game. There is no single answer to this challenge; each situation is different and must be handled on a case by case basis. What all parties must keep in mind is that federal regulations prohibit disclosure of the pendency of a misconduct investigation or the parties to the proceeding. I believe disclosure to a journal can be done in a way that preserves a scientist’s right to confidentiality and protects them from the indelible stain on their reputation that can occur when news of a research misconduct investigation gets out. Don’t get me wrong, publications like Retraction Watch perform a very valuable service to the scientific community. But, in some cases, the academic journals (and the broader scientific community) must refrain from asking the question about possible misconduct. Authors and institutions must also take steps to prevent inadvertent disclosure of that information. Once there is an identified need to contact a journal, my suggestion is for all parties to focus, in the near term, on correcting the scientific record; the details of an institutional process (i.e., who is responsible for an error, the level of culpability, and the actions have been taken, if any, against an author), should be deferred to a later date.

Thaler: The regulations provide that only those who have a “need to know” should be given information about a pending investigation. There have been differing opinions about whether publishers of journals fall within that “need to know” category. Having represented many dozens of scientists through the years, I can attest that if there is a balancing test in place (between the public’s “need to know” and the scientist’s career), institutions have routinely favored notifying publications to the detriment of their scientists’ careers. This does open the institutions up to criticism and possible litigation. I have taken the position many times that the individual scientist’s career faces possible irreparable injury with a public announcement of the proceedings. As I have written, many of the researchers are ultimately exonerated so that damage is entirely unwarranted.

Of course, the counter argument that science itself has an interest in proper revisions is important to remember. That is typically addressed when the institution and the scientist both submit position letters to the publisher concerning retractions or corrections. It probably should be managed without reference to ongoing misconduct proceedings, however.

Hat tip: Virginia Hughes

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, sign up on our homepage for an email every time there’s a new post, or subscribe to our daily digest. Click here to review our Comments Policy. For a sneak peek at what we’re working on, click here.

Written by Ivan Oransky

September 7th, 2016 at 9:35 am

  • Giacinto Scoles January 19, 2017 at 12:43 pm

    one does not need to FAKE data to be “miscoonducting”, but if you write the paper first and you give to the lab for validation AND, in case the latter is not forthcoming, you give it back without any changes enough times till the data fit the ideas of the paper you commit whatI would call intellectual misconduct that in many cases is worse than faking data because if nothing else is more dificult to prove.

  • Post a comment

    Threaded commenting powered by interconnect/it code.