Panel discussion on the Future of Translation, moderated by M. Ristkes, with P. Pound, J. Ioannidis, F. Miedema, W. Beumer, and @Meijboom75.
Miedema: we used to review our own, and our colleagues' papers prior to publication. Now we find students submitting papers that were not checked by their supervisors.
Ritskes: Ioannidis stated the translation pipeline is leaking, where is it leaking and can we do to fix the pipeline?
Ioannidis: If we fix just one little piece, it will leak somewhere else. We must take multiple shots and some of them might hit the target!
Ioannidis: If we fix just one little piece, it will leak somewhere else. We must take multiple shots and some of them might hit the target!
Who should lead reproducibility/translation improvement?
Ioannidis: We can't name a Science King to fix things. Some fields are taking the lead, some branches of science more healthy than others, and will likely influence others. Funders have major power and can set standards.
Ioannidis: We can't name a Science King to fix things. Some fields are taking the lead, some branches of science more healthy than others, and will likely influence others. Funders have major power and can set standards.
Ritskes: Pound mentioned psychological/institutional lock-ins for accepting
Pound: animal research is a very closed institution, so it should involve and be open to input from people from other fields, funders, clinicians, public, animal protection groups.
Pound: animal research is a very closed institution, so it should involve and be open to input from people from other fields, funders, clinicians, public, animal protection groups.
From audience: there is very good experience in the Netherlands from allowing animal protection groups into the discussion on animal use and their welfare, even inviting them into the labs. It's good PR, of course, but scientists are genuinely committed to doing things better.
If it is true that we have succeeded in treating a number of conditions in mice, but not in humans, shouldn't we just stop using animals?
Meijboom: if it doesn't deliver, of course.
Ioannidis: not ready to throw the towel. If we improve methods, there is room to improve outcome.
Meijboom: if it doesn't deliver, of course.
Ioannidis: not ready to throw the towel. If we improve methods, there is room to improve outcome.
[My view: we have not succeeded in curing cancer in mice. That is a misconception. Reducing tumor growth/number in comparison to controls, yes, but not more than that. As in humans, only surgery is truly curative. If anyone "cures cancer in mice", a Nobel prize is in order.]
Ioannidis: also plenty of room to dramatically reduce waste from repeating unnecessary, redundant, experiments to obtain knowledge we already know. [From audience: maybe true for regulatory tests, not basic and applied science. Ristkes: we've seen repetition in out sys. reviews]
From audience: is comparison of same outcomes the best approach to evaluate the best model? Shouldn't we be comparing systems, going from animals to humans?
Ioannidis: observations prone to bias from current narratives, same for selection of references (eg. positive vs negative)
Ioannidis: observations prone to bias from current narratives, same for selection of references (eg. positive vs negative)