Looking back at the Baltimore case

I mentioned the Baltimore affair in the previous post about scientific fraud. Here’s a review of Daniel Kevles’ definitive book about the case. I don’t usually mention where these things appeared, but it’s relevant here as its publication, in Nature, drew a comment from the author. He says I miss a point about the historical significance of the case. I’m not sure, at this distance, that I agree: nor was I then. But I reproduce the letter below, for completeness. I suspect he was piqued by a lukewarm review of what (I found to be) a very long, detailed account of just one alleged case of scientific misconduct. Fair enough: it must have been a tremendous amount of work…

 

Few of us, I dare say, had the stamina to follow the Baltimore affair properly. A rich stew of supposedly fishy results, personal vendettas, enormous egos and the murky politics of leaks, smears and intimidation, it simmered for ten years. True, it had spice. An American Nobel laureate in biology was hauled before pugnacious Congressmen to rebut charges of scientific fraud in a paper he had co-authored. The US secret service pored over laboratory notebooks from no less an institution than the Massachusetts Institute of Technology (MIT). But the charges turned on the interpretation of fairly arcane aspects of the expression of immunity in transgenic mice. And, by the time the multiple investigations, and investigations of the investigators, came to a halt, it was hard to feel any the wiser.

Few of us, I dare say, had the stamina to follow the Baltimore affair properly. A rich stew of supposedly fishy results, personal vendettas, enormous egos and the murky politics of leaks, smears and intimidation, it simmered for ten years. True, it had spice. An American Nobel laureate in biology was hauled before pugnacious Congressmen to rebut charges of scientific fraud in a paper he had co-authored. The US secret service pored over laboratory notebooks from no less an institution than the Massachusetts Institute of Technology (MIT). But the charges turned on the interpretation of fairly arcane aspects of the expression of immunity in transgenic mice. And, by the time the multiple investigations, and investigations of the investigators, came to a halt, it was hard to feel any the wiser.

So Daniel Kevles performs a signal service by applying his historian’s acumen to the mountain of documents accumulated by all those investigations, and to his own detailed interviews with the participants, to give us an orderly narrative of the whole affair. He did so, he says, because it seemed to him likely to “throw some light on science in late-twentieth-century American society”.

But he clearly got caught up, as historians of contemporary affairs will, in the detail of the events, to the extent that he published a lengthy article inThe New Yorker exonerating the accused scientists some weeks before a panel of the US National Institutes of Health came to the same conclusion. The result is a book that is likely to be judged, as one of the blurb writers suggests, the definitive account of the affair, but not, perhaps, a complete consideration of its wider significance.

Trials and errorsImanishi-Kari: faced accusations from within her own laboratory.

So what was it all about? The first thing is that it should more properly be known as the Imanishi-Kari affair. David Baltimore’s co-worker, Thereza Imanishi-Kari, was an expert in cellular immunology, and was accused of faking results published in Cellin a paper co-authored with Baltimore and four other biologists. In fact, the accusations grew stronger over time, but that was the final allegation. The results — and the scientists — were then subject to an extraordinary series of investigations: by Tufts University, Imanishi-Kari’s new employers; by MIT; by the self-appointed ‘fraud-busters’ at the National Institutes of Health, Walter Stewart and Ned Feder; by a congressional committee; and by a succession of NIH panels and reviews.

The first two found no case to answer. The others convicted Imanishi-Kari and Baltimore variously of inaccuracy, scientific misjudgement, error, arrogance and fraud. After the results of a draft report by the NIH’s newly fledged Office of Scientific Integrity were leaked in 1991, Baltimore was obliged to retract the paper (and, soon after, to resign his presidency of Rockefeller University in New York). After further thought, he retracted the retraction. Finally, an NIH appeals board reviewed all the charges — now numbering 19 — that had been laid against Imanishi-Kari by NIH investigators, and rejected every one.

Kevles is clear throughout that the Cell authors were victims of a miscarriage of justice, only put right because of their refusal to be cowed by weighty indictments based on shoddy evidence. He explains in detail how the final review showed that the NIH’s supposedly exhaustive investigation had supplied no plausible motive for fraud. And its case rested almost entirely on forensic evidence (mostly from analysis of printer inks recording radiation counts and of the sequences of entries in lab notebooks) which itself failed to withstand serious scientific scrutiny. The data offered against Imanishi-Kari, in other words, fell below the standards her paper was being held to. Before recording the two biologists’ vindication, though, Kevles has the more taxing task of explaining how the affair acquired such extraordinary momentum.

Imanishi-Kari’s original accusers were from her own laboratory, her main antagonist being a postdoctoral researcher, Margaret O’toole — later cast as a heroic whistle-blower by almost all the reporters covering the story. Somehow, difficulty with some of the techniques used in the paper, construction of different interpretations of the results, and long contemplation of fragments of the data, coalesced to generate first suspicion, then certainty, that all was not as it seemed.

Kevles’s account shows these suspicions being nurtured in an institutional hothouse, rife with thwarted ambition, personal animosity, jealousy, nepotism, double-dealing, secrecy and paranoia. He does not say whether these currents were stronger in this particular lab, at MIT, or are more or less those found in any corner of a modern biomedical research enterprise, where everyone is frantically chasing publications, tenure, grants, or just enough data for a PhD.

What he does show is how these suspicions were picked up and amplified by critics pressing the US scientific community to take fraud and misconduct more seriously, mistrustful of self-regulation, and determined that researchers must be accountable for all the NIH’s billions.

Although he is critical of how most of these people conducted themselves, Kevles seems to feel that their reasons were mostly valid. It was just that in this case they got the wrong people.

It is a sorry saga, the more so because Kevles is bound to tell us a story whose end we already know. But in hindsight it is not, I think, of enormous significance, except for the victims. The outcome helped to stem the over-zealous pursuit of scientific misconduct, but there were enough genuine cases to fuel continuing concern. The NIH’s final verdict upheld the Cell authors’ scientific right to publish what they judged to be the best interpretation of their results. But this did not overturn the broader arguments for researchers to submit to outside scrutiny.

What is the moral of the story? The most useful lesson is one that Kevles only hints at at the very end of the book. He quotes approvingly from Baltimore to the effect that deciding what counts as a worthwhile scientific paper is a matter of professional judgement, specific to a particular field at a particular time.

This fact of research life, it seems, was denied by the congressional inquirers, by many journalists, and by some of the accused workers’ scientific peers. But, if all of these people appeared to believe that every research publication must contain certified and irrevocable Truth about how the world is, perhaps that is because that image of science has been found advantageous by most scientists most of the time.

If any good came of the unsavoury affair, it may be the extended public exposure of what scientific practice is actually like — full of ambiguous results, unexplained anomalies, imprecise assays, and of the craft skill which turns all these things into a story fit for journal referees. As Baltimore said, however a paper makes it look, researchers know that “no study is ever complete”. (1998)

The Baltimore Case: A Trial of Politics, Science and Character by Daniel J. Kevles

Norton. 509pp

—————

and the author’s comment…
Star chambers will result in injustice

Daniel J. Kevles1

  1. California Institute of Technology (228-77), Pasadena, California 91125, USA
    e-mail: Email: kevles_d@caltech.edu

Sir

Jon Turney’s generally favourable review of my book The Baltimore Case offers the judgement that the affair was “not of enormous significance, except for the victims” (Nature 395, 30–31; 1998). That is to miss the point of historical importance of the case, and a major point of the book. The case had great significance for the civil rights of scientists charged with fraud.

Thereza Imanishi-Kari was investigated relentlessly by a congressional subcommittee and by an entity in the US Department of Health and Human Services now named the Office of Research Integrity (ORI). In both venues, she was denied elementary protections of due process, especially the right to see and evaluate evidence and the right to confront and cross-examine witnesses against her. So she could not defend herself effectively. The miscarriage of justice was put right because Imanishi-Kari’s tenacious refusal to concede ultimately gained her access to an appeals board that afforded her full due-process rights.

I think it perfectly legitimate — indeed, obligatory — for government authorities to call to account publicly supported scientists who cheat. My quarrel with the investigation of Imanishi-Kari is not primarily, as Turney says, that the authorities went after “the wrong people”. It is that the ORI went after Imanishi-Kari in the manner of a Star Chamber. The denial of due-process rights to anyone, innocent or guilty, is indefensible. The ORI investigators defended their procedures on the grounds that they were not conducting a legal inquiry but engaging in a dialogue between scientists. But, if a criminal penalty was not necessarily at stake, careers and reputations were.

The case has already had a major impact on how such charges in biomedical research are handled in the United States — by, for example, establishing the right of all accused scientists to use the appeals process. The case is deeply instructive for scientists and policymakers in the United States, where the procedures for handling fraud cases are still evolving, and in other countries, where questions of how charges of scientific fraud should be adjudicated are commanding increasing attention. The ORI’s original procedures threatened the civil rights of every scientist who might be charged with fraud.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s