Does it take too long to publish research?

This is the question posed by Nature this week. The article is full of stories of research papers submitted to Science and then finally accepted two years later in PLoS One. Certainly I've had experiences pretty close to that, and for big stories from my lab about a year between submission and acceptance is normal. At Nature, the median time between acceptance and publication is 150 days (up from 85 days a decade ago). Even more striking is the amount of data needed to get into Nature - a 10-fold increase in data panels (and each panel has a lot more information too!).

The big problem is not really the dozens of experiments needed to reply to reviewers though. Rather, I think the hardest part is the roulette of getting editors/reviewers that like the paper. The article is rather dismissive of "journal shopping", but the simple fact is that submitting a paper is a lot like rolling the dice. 95% of articles that I have published have ended up in a journal of similar rank to the initial submission (the other 5% cause most of the heart-ache). But this doesn't mean that there is a smooth ride. Rather, you can spend a year at review at Cell, doing all the experiments those reviewers want, then you still get rejected. The paper gets rejected at Immunity without review, then Nature Medicine sends it out but gives you new reviewers who want an entirely different set of experiments. No matter how much you have done, the big journals will always ask for more - and you can't predict in advance what they will ask.

All of this takes a lot of time, however having published in the social sciences as well, they are even slower. The difference is in how much effort and energy the publication process takes in medicine. At a top journal, it is not unusual for the revision to require €100,000 in salary and reagents to get those last experiments done for the reviewer. To me, the more important question is whether this cost is worth it.