I am in New York City attending a New York Academy of Sciences’ conference titled: “Animal Models and Their Value in Predicting Drug Efficacy and Toxicity.” I presented a poster saying what I usually harp on here. Namely, that animal and humans are complex systems thus extrapolating the results of perturbations that occur at levels of organization where disease and drug response occurs is problematic.
I was optimistic that this conference might actually discuss the problems with using animals to predict efficacy and toxicity. After all, the name does rather imply that this would be discussed. Instead I heard pretty much the same things I have heard many times before. For example:
1. My animal model of X predicts human response really well, but I admit that almost all of the others do not.
That speaks for itself.
2. Animal model Y correlated with human response in response to drug Z therefore, despite the fact that thousands of other animal model models did not correlate with other drugs and despite the fact that this animal model never correlated with human response again, this one correlation proves that animal models can predict human response to drugs.
Someone needs a course in simple statistics.
3. If we can find the mechanism of a disease or drug response, it will be the same across species.
And maybe a course in evolutionary biology.
4. I realize that a lot of failures in clinical trials are secondary to the fact that the animal models were wrong, but all in all animal models work really well for drug development.
And, while they are in the learning mode, a course in critical thinking / logic.
5. The animal model got it right but the clinicians used the drug the wrong way.
Darn those stupid MDs! Next time I am sick, I going straight to a PhD.
6. Yes, animal models have misled but the times they got it right contributed to medical advancements therefore all the misses are justified.
Tell that to the people that died or were permanently injured from those misses.
7. Animal models are all we have.
Actually, there are a lot of other ways to test drugs. Moreover, the problem is not lack of ways to test drugs the problem is that animal models are not predictive. If a test does not accomplish it’s purpose then the test should be abandoned. That is what adults do when tests do not work.
What I did not hear was word one about prediction. The breakout sessions I attended were the same old same old. Scientists got up and discussed the research they are doing in their labs and why their animal models are so wonderful. They did not use statistics to show how the animal model was predictive for disease X or Y. They did not discuss how their animal model could be improved or what problems it presented when compared with human data. There are a few hours left in the conference so maybe all this will be cleared up during the last sessions, but I doubt it.
There have been some interesting quotes, however. William Stokes, DVM, DACLAM, of the National Institute of Environmental Health Science and the US National Toxicology Program Interagency Center for the Evaluation of Alternative Toxicological Methods stated: “As a veterinarian, we are used to taking into consideration all the species differences.” That is amazing! Especially in consideration of the fact that most of the species differences, for example gene expression and regulation, have not even been discovered yet.
Dominic Wells, MA, VETMB, PhD of The Royal Veterinary College stated that surgery residents should be allowed to use pigs to practice technical skills “before they are turned loose to operate on grannie’s innards.” Yeah, I hate it when those surgery residents are turned loose on innocent patients. Just for the record: surgery residents are never “turned loose.” They are always supervised by someone with enough experience and training to adequately supervise their actions and they “act” only after watching said action many, many times. The first time they “act” be it to suture or make the skin incision there are multiple sets of eyes watching closely and the attending almost always interrupts and correct something (even if the poor resident was doing it properly and was not in need of correction). But my experience of interacting with surgery residents in the OR at four medical schools has been confined to the US. Wells is from the UK. Maybe over there they just let naïve students experiment on patients. You might want to contact Dr Wells and ask him about that.
The above quotes illustrate the abject ignorance of the both science (Stokes) and real life (Wells) that were represented so well in this conference. I spoke with many people at the conference and found very few that even understood what a complex system is much less what complexity implies for extrapolation. But everybody was positive that animal models either currently, or would in the future, predict human response for efficacy and toxicity and not just human response in general but the response for every genotype.
The organization for the conference was very good. I have attended a lot of medical and science conferences and this one was well planned and well run. The New York Academy of Sciences deserves recognition for their efforts. The New York Academy of Sciences was not the problem. The problem was that the conference was asking people with a vested interest in animal models to criticize those models publicly and a vast majority of the speakers just could not bring themselves to do it. I acknowledge that many in the animal model community have criticized their models; it’s just that either none of them were at this meeting or they were staying very quite. At least in the sessions I attended.
Perhaps the quote that best sums up the conference titled “Animal Models and Their Value in Predicting Drug Efficacy and Toxicity” came from Alan Daugherty, PhD, DSc of the University of Kentucky. The moderator of a session asked panel members to answer the question: “What makes and animal model good?” Daughtery replied: “One that gets me a grant.” You have to respect his honesty.