My post last Friday on Brazil and drug patents inspired some readers to take issue with the one instance in which I came to Big Pharma's defense: "It costs an awful lot of money to successfully develop a new drug. It costs very little to copy an already developed drug. If intellectual property protections didn't exist, why would an Abbott or a Merck bother to create new drugs?"

The critique focused on the phrase "awful lot of money." And many good points were made: The pharmaceutical industry is extraordinarily profitable, benefits from research paid for by the public sector, and devotes much of its R&D efforts to developing new drugs that are only minor modifications of old drugs. But it all led me to wonder: What are the actual numbers?

The canonical figure cited by the drug industry in appearances before Congress is that taking a new drug from start to finish costs $802 million, as of 2001. These figures derive from a study conducted by researchers at the Tufts Center for the Study of Drug Development. But that study has come under sustained attack from Big Pharma critics.

Some of the points made: The Tufts Center relies heavily on pharmaceutical industry funding. The figures incorporate a so-called opportunity cost -- that is, it takes so long to develop a drug that if the money spent on R&D was invested elsewhere, such as in bonds or the stock market, the return on investment would be significant. The money that could have been made but isn't, is thus counted as a cost. Actual "out-of-pocket" expenses are some 50 percent less, or $403 million. And then there's the problem of the real data: We don't get to see it, as the numbers are closely held by the industry.

Futhermore, as other eloquent critics of the "802 million dollar myth" point out, R&D expenses may add up to only about 20 percent of the drug industry dollar. At least that, and perhaps twice as much, goes to marketing. Then there's overhead, sales expenses and, of course, profit. Which all adds up to the not entirely newsworthy revelation that drug companies make a lot of money, and the high prices of many new drugs are not likely justifiable, on a pure dollars-and-cents basis. There's clearly room for some belt-tightening.

None of this, I still argue, changes the fact that copying something that has already been invented is a lot cheaper than developing something new; thus the friction that makes the global struggle over intellectual property so heated. But when the issue at hand is public health, better accounting would be nice, for a start.