Which are the most efficient drug developers in biopharma?

Few industries suffer from the kind of extraordinary R&D failure rates that afflict biopharma. When 95 out of 100 projects that go into the clinic fail, the odds against success are immense. And the cost for a single regulatory win can be extraordinary.

In breaking down the numbers on 98 publicly traded pharma and biotech companies, Forbes' Matthew Herper found that Big Pharma with its multibillion-dollar research budgets bears the lion's share of the cost. Adding up their total research costs, Herper came up with a whopping $5.5 billion median cost for each new drug developed over the past decade by the largest companies with 8 to 13 new drug approvals on the record. For companies with 4 to 6 approvals to their credit, the median cost was $5.1 billion, while the two-or-three-approval class came in at $1.8 billion and companies with a single approval recorded a thrifty $351 million median cost.

Herper worked with figures from Bernard Munos' Innothink Center for Research in Biomedical Innovation.

Size is a factor. Companies that spent more than $20 billion on R&D in the past decade spent $6.3 billion for a new drug compared to $2.8 billion for companies spending $5 billion to $10 billion over the same period. As J&J's Paul Stoffels notes, the company factors in a big expense just tracking the safety of approved meds. So the more new drug approvals you have, the greater the follow-on costs.

Herper also singles out a few clear winners on cost efficiency. Bristol-Myers Squibb ($BMY) had 9 new approvals at a cost of $3.4 billion per drug approved. That's half of what Eli Lilly ($LLY) and Pfizer ($PFE)--two companies which have grappled with R&D setbacks in different ways--have paid. Other standouts: J&J ($JNJ), Novo Nordisk ($NVO), and Amgen ($AMGN) in the Big Pharma group and Regeneron ($REGN), Gilead ($GILD), and Biogen Idec ($BIIB) among the biotechs.

Merck R&D chief Roger Perlmutter

Merck ($MRK) R&D chief Roger Perlmutter tells Forbes that two bad habits contribute to inefficiency. One is that companies are often reluctant to kill programs that look increasingly unlikely to succeed in the clinic. The other is that CEOs can't just order up another big blockbuster. Companies have to follow the science.

"Great drugs build great franchises, but great franchises don't necessarily build great drugs," Perlmutter tells Forbes. "If you are too prescriptive with your R&D, you can spend an awful lot of money and not be terribly productive because there may actually not be any new mechanisms that you can get to right now that will help you in a particular disease area."

- check out the feature for more ideas

Special Reports: Which Big Biotechs are hitting the gas pedal on R&D spending? | The Biggest R&D Spenders In Biopharma | FDA Approvals of 2012