Today we see yet another story on the emerging epidemic of bad science, this one from the former head of Amgen’s global cancer research. Of 53 “landmark” publications in top journals, Amgen could not replicate 47 of the claimed results. Many of the papers were from universities. This is nothing new, following on studies by John Ioannidis, Jonah Lehrer and others that show that many science articles published in top journals are simply wrong. Other studies indicate that an alarming proportion of university scientists are willing to fabricate or selectively publish data.
In today’s report, scientists point out that the pressure to publish in academia is so great, and the protections on research that would seek to replicate so poorly represented, that one wonders if university science is, generally and simply put, unreliable in its published form. If so, then peer review is failing, not just in highly politicized areas like climate science, but in core areas of research, such as cancer.
This bodes very badly for the continuing university technology transfer emphasis on biotech. The tech transfer community is part of this problem, and has to take steps–now–to address it openly and decisively. Some suggestions:
- Every description of a “technology available for licensing” must indicate if the claimed invention has been independently replicated by another lab.
- Grant a standard, no-cost, no-formalities license for anyone to make and use an invention for the purposes of evaluating published claims
- Require inventors to disclose all their findings, not just those that claim a distinctive or apparently valuable result
- Make all backing data available for inspection for anyone seeking to evaluate a technology available for license–whether for commercialization or research use–on the same standard as for publications relied upon for public rule-making (see 2 CFR 215.36(d))
However, these suggestions are just for the conventional patent-marketing technology transfer office. What we really need, for technology transfer to be really a leader, is for university IP management to get out from behind the patent desk, away from the marketing counter, and return to what Bayh-Dole, and Vannevar Bush, expected–back to collaboration as the primary goal. And by collaboration, I’m not meaning a patent license or hitching up to a state-run economic development resource (though at times these can be quite good)–but rather by collaboration I mean honest, forthright engagement, so that others can visit the lab, see the data, review the practices–all outside and before any peer-reviewed publication. That’s where technology transfer starts, not with an invention disclosure and a patent pay-wall to be “marketed”.
If we are getting bad science, and technology transfer plays the clueless shill to try to hawk it to industry, and the papers in the top journals are being shown to be unreliable, then tech transfer practiced in the conventional manner is part of the problem. Time for tech transfer to be part of the quality assurance program, not the make-money while you can program. Getting a patent doesn’t mean the science is valid–it may just mean that the reputation of the university has been spent on overmatching a patent examiner, and when this happens, it is the work of the technology transfer office.
Things can be different. There can be innovation in innovation management. Universities and their technology transfer functions have a tremendous opportunity to be leaders here. It would be a shame if they stayed quiet, left the present crisis in science as a problem for industry and entrepreneurs, and continued braying about a funding gap. It would be a shame if they didn’t ask whether the road they are on is not only not a decent local maximum, but perhaps not much of a maximum at all. This would be a great time for leadership in technology transfer. This is a great time for technology transfer to get on a new road, before it is tarred with the problems that are catching up with university science.