Small Data: The huge cost of developing drugs
Everything about the US pharmaceutical company Pfizer's attempts to buy its Anglo-Swedish rival AstraZeneca is big. Let me try to explain why, writes Anthony Reuben.
The Association of the British Pharmaceutical Industry estimates that it costs on average £1.2bn to bring a new drug to market - and takes 12 years.
And if that seems like a lot, Forbes Magazine reckons the biggest companies pay way more than that.
- A series on curious numbers cropping up in the news, by stats watcher Anthony Reuben
Its reporter Matthew Herper took the slightly crude but basically valid step of taking the amount spent on research and development by pharmaceutical companies over 10 years and dividing it by the number of drugs each company brought to market over that time.
On that measure, he found that AstraZeneca was spending the third most at $9.6bn (£5.7bn) per drug while Pfizer came in at number five spending $7.8bn (£4.6bn) per drug.
Looked at in that context, you can see why the £63bn that Pfizer is offering for AstraZeneca (and which AstraZeneca's board has turned down) would not have to give Pfizer access to very many new drugs to be worthwhile.
That's even clearer when you look at the rewards available for a blockbuster drug.
The world's bestselling drug last year, according to IMS Health, was Humira, a treatment for rheumatoid arthritis among other things, which had global sales of $9.9bn (£5.8bn).
That one drug alone would almost have been enough to put a pharmaceutical company in the global top 20 firms.
So that's why there are such a lot of big numbers circulating linked to this potential takeover and why companies keep merging so they can shoulder the costs of research.
Some scientists reckon that the costs linked to bringing out new drugs are just too high, making them unaffordable for patients, and that a new approach is needed if the new drugs the world needs are to be discovered.
But until there is a new approach, expect to see more such mega-mergers proposed in the future.