3 ms·
There have been public complaints about this too, such as that "cancer findings don't reproduce" editorial (where "reproduce" is defined as "generalize") that g
by epistasis 9y ago
There have been public complaints about this too, such as that "cancer findings don't reproduce" editorial (where "reproduce" is defined as "generalize") that got lots of press 5 years ago. My best understanding of the situation is that big pharma and research are going at opposite purposes. Big pharma wants something that generalizes and can lead to drugs that hit almost all of the target population. Research emphasizes novelty, and finding new stuff, which most often won't generalize.
Chasing the findings in "leading journals" is going to escalate this disconnect. Leading journals go for novelty and surprise alone. These are exactly the findings that are least likely to generalize.
I'd love to see how and where VC is funding life science research. It seems unlikely that the types of projects that VC funds would overlap the type of research that NIH funds, but I would love to be wrong on that!
- aaavl2821 9y agoIts definitely true that big pharma and academic research are def going at divergent purposes, which isnt always a bad thing, but i agree that when quality is sacrificed for "publishability" then i dont think anyone wins. A more somber issue with reproducibility is that you simply cant replicate a study's results independently even if you do it the exact same way. Not only is it not generalizable, it isnt even robust in a specific case. A lot of this is just due to chance, but in the worst instances, this happens because researchers are able to bias experiments to get a result. There are so many tiny details of experimental design and execution that influence results, and a lot of these dont end up in the "methods" section of a publication Some big pharma companies and startups are actually starting to look at specific rather than general populations (the whole "precision medicine" thing), which is good. This lowers revenue potential, but decreases risk of clinical trial failure, bc you are only targeting a group of patients where you have high-ish confidence it will work. VC is increasingly funding stuff that would be done in academia. Biotech startups have been so successful in recent years that investors are taking on more risk, which means earlier stage, bolder science, and this overlaps with some academic stuff. Most VC funded stuff is translational in nature though, ie projects in between academia and big pharma. however only a few "profitable" diseases get funded, like cancer and orphan disease. very little going in in psychiatric disease, addiction, heart disease, diabetes