Education: Arriving with a Consort

Muhlhausler BS, Bloomfield FH, Gillman MW (2013) Whole Animal Experiments Should Be More Like Human Randomized Controlled Trials. PLoS Biol 11(2): e1001481. doi:10.1371/journal.pbio.1001481

The ARRIVE (Animal Research Reporting In Vivo Experiments) guidelines are laudable and something I think will help make animal experiments more transparent. These follow-on from the production of the CONSORT guidelines, which is were introduced to make clinical trials more transparent and robust.

This paper suggests that animals studies should be more like randomised trials and that the ARRIVE guidelines do not go far enough. Well before doing this, I think people need to adopt the guidelines or at least the central core elements of the ARRIVE guidelines and this is not being done. This is because of word-limited papers and other things that may be a bar to publication. I will write about this in the future.

However, lets face it clinical trial reports are often as dull as dishwater. Science reports are about discovery and this may mean following your nose and where you start the science journey is not where it actually ends up. This is not about one experiment but a series of experiments that evolves over months and years Importantly not all science papers are done with the treatment of humans in mind, and so it is not about drug testing.

Interestingly this paper suggests "We also believe that a registry of animal experiments would reduce publication bias, as do sites such as www.clinicaltrials.gov for human randomzed controlled trials. Such steps are integral to improving the usefulness of whole animal experiments". 
I have heard this idea bantered around a few times.This is where this needs thinking through, as there are implications. There is the cost and logistics of registration, but there is the issue of confidentiality until the work is published, as it can take years to do work. You exposure yourself to your competitions who may have resource to do the work before you can do it, unless registering stops others do the work..this ain't going to happen. In clinical trials drugs are usually protected by patents and if they are not protected then they often go nowhere, which is why we need the BPA (see above)


Importantly are the pharmaceutical industry going to do this?  surely the answer is No!!!

Their work will often be done before patents are filed, as this work provides the information for patent filing. Registration would be a form of public disclosure, which makes the subsequent patent useless. Furthermore, do you think a company from one country could trust a registry in another country controlled by a foreign Government, who could favour their own companies (Conspiracy theory off we go). There is a publication bias that only positive results tend to be published, but were the experiments designed so that they could detect a real failure rather than a hint of a positive result? However this is not going to be changed by having a registry. 

Clinical trials are registered. This does not influence where the positive trials are reported and often the negative trials are not reported or it trickles out only years after the event.  This certainly happens in experimental studies as it takes a lot more work to try and break dogma once established. I remember trying to repeat some stuff from a paper and on speaking to a few people found at least eight other groups around the world who had tried and failed to repeat it. A registry of these studies and reporting their outcomes would have shown lack of reproducibility. However, this could also be manipulated by pharma to rubbish others studies.

However, having just read a few patents today (no deviations on the EAE graphs)  if the ARRIVE guidelines were applied to all patents then this would be of value, because this work is done with a view of commercialisation and drug development for human use. Therefore, it would be appropriate that it is reported properly. At present the value of the patent is judged by Lawyers and my personal experience has found that they take the literal word and the actual value of the science is not always important. They therefore accept stuff that is clearly weak scientifically such as comparing data from a knockout mouse on C57BL/6 mouse background with data from a wild-type sabra mouse, that is not C57BL/6 mouse, but a completely different strain. 

Patents lack the peer review process of science papers and so incorporation of reporting guidelines here would be of real value  Food for though......What do you think?..Maybe you can't be bothered?

Labels: