Merrill,

A number of comments on the piece:

Do you know how many people must be treated for one year to eliminate one death by heart attack with a cholesterol-lowering drug? 750. No wonder the clinical trials to approve a new medicine in this category must enroll 6,000-8,000 patients to have statistical validity

There is some truth that chasing minimally incremental benefits for new drugs will require larger trials, but I'd suggest regulatory hurdles for approval are the bigger culprit. Your example, is actually my point. The need to prove impact on mortality rather than just cholesterol-lowering was a regulatory requirement, because FDA was not willing to just believe that lowering cholesterol had clinical benefit. The earliest cholesterol lowering drugs, including the first statins, did not need to show mortality benefit. The large trials necessary to show a mortality benefit would be similarly large for the first cholesterol lowering drugs as they currently are for newest drugs. Meaning, it was the higher efficacy hurdle that impacted size not that they were me-too drugs.

If your previous cholesterol-lowering drug that generates billions of dollars in sales is coming off patent, you might need one. But does the public?

Some do, some don't. Depends on how they've responded to current drugs (population medicine doesn't function at the individual level). Only those who need it should pay for it. You can argue for more restrictions on DTC, but the problem isn't me-too drugs, its people willing to pay too much for incremental differences. Everyone wants the best drug, the other guy can take the generic. Pharma is profiting from people's unwillingness to compromise at all between quality and cost. In addition, multiple me-too drugs on patent is the most effective way of bringing down drug costs-- in fact, when people talk about "government negotiating prices"-- the details are actually about forcing price competition among me-too competitors. No me-toos, no price competition. (In terms of how proposals in the US have thought about price negotiation.)

Now let's take a truly innovative drug to cure a rare form of leukemia that might strike 15,000 people a year.

Comparing a cardiovascular drug to a cancer drug isn't a fair comparison. Regulatory hurdles for cancer drugs are extremely different than other drugs, the standards are much, much lower. I don't disagree that there is some truth to your point that incrementally better drugs have a tougher burden of proof, but your comparison dramatically overstates the reality and it isn't a fair comparison.

On your general point on distorted market incentives, I'd suggest that your general conclusion are increasingly outdated. If you read public comments by senior executives of pharma companies, they recognize that the model described above (pushing me-toos via DTC, Viagra, etc.) is a model that is broken. The sector as a whole, has been underperforming compared to the 90's, where innovation in cholesterol lowering drugs, blood pressure drugs, new antibiotics, depression medications, etc. did drive real innovation and success. The industry got lazy and its paying the price in terms of performance. Companies spend more time talking about the pharma innovation conundrum than anyone else, and are heavily reinvesting in basic research and discovery as well as strengthening ties to academia. There's a long way to go, but your characterization of a broken industry with perverted incentives is increasingly outdated.

I was responding to the arguments in the Atlantic blog by McArdle, who started with the premise that it cost $500 million to develop a new drug largely because of the cost of clinical trials. My point in comparing a cholesterol-lowering drug to a cancer drug was that her blended, inflated number was meaningless.

As to your point about higher regulatory hurdles, proving that a drug affects mortality and improves health (efficacy) IS the regulatory requirement, not improving a biomarker like lowering cholesterol. Look at the Vytorin trial. The Zetia component of that combo pill lowers cholesterol, but there is ZERO evidence to date that it saves lives, and the early evidence is that it doesn't. We won't know a full answer until 2011 -- nine years (!) after the FDA approved the drug. We have lower regulatory standards that we had a generation ago, not higher regulatory standards. For a full discussion of that issue, see my review of University of Chicago Richard Epstein's book "Overdose: How Excessive Government Regulation Stifles Pharmaceutical Innovation" in Vol. 10, Number 4 of the DePaul Journal of Health Care Law, published last year.

I am familiar with all the latest trends that have made the old arguments about "the high cost of developing new drugs" irrelevant, which I discuss in detail in the essay "The Pharmaceutical Innovation Conundrum" on this website. Indeed, that essay was written in response to an academic who wanted me to address all the old questions for a book he was putting together. After I wrote it, he rejected it. I would be very interested in knowing your response to that essay, since you're obviously quite knowledgeble about the field of drug innovation and research.

It seems to me that somewhere along the way drug companies and doctors for that matter, forgot that the real goal should be to improve lives, not just sell a drug or perform a surgery.

It seems like many of the drugs that are currently produced while marketable may not really improve health outcomes

It's ironic for McArdle's "military research is pretty useless" comment to be published on the Internet, the result of some of that "one of everything/useless" research (DARPA's total outlays on all areas of research, since its inception in the late 50s, come to only about $60 billion --- an investment that repays itself annually in taxes paid by people whose incomes depend on the internet).

Re: Wisewon's comment on the demand for non-generics:

I think you're missing the fact that health-care consumers are faced with a massive information asymmetry compared to drug developers and advertisers. Drug ads don't say "buy our product, as it will be marginally better than our competitors'" - they imply that their drug is the only cure for what ails you. More importantly, generics aren't advertized, so consumers are less aware of them as alternatives. Your comment seems to infer exogenous consumer preferences from a dynamic system - i.e., you assume that people prefer to overpay (irrationally) for namebrands than buy the (far) cheaper and (assumedly) slightly less effective generic, rather than this being an indication of a lack of information attributable to the perverse incentives of advertising healthcare as if it were a commodity.

Seriously, it's like wisewon forgot that drug companies spend more on marketing than R&D. How else to explain "people willing to pay too much for incremental differences." Pharma is not advertising those differences as incremental, and often the doctors themselves don't know any better either.

Fascinating stuff, thanks :) Spike makes a point that got me thinking: "Pharma is not advertising those differences as incremental, and often the doctors themselves don't know any better either."

Another factor here is probably that given marginal or negligible differences in efficacy of competing meds, doctors themselves are more susceptible to advertising. We know from experiments that doctors' prescribing patterns are indeed effected by pharmaceutical companies' gifts, even though the doctors claim not to be effected. An ethical doctor deciding between an effective drug and snake oil will prescribe the effective drug no matter how many lunches the snake oil company has bought him. But if the differences in efficacy are so marginal or unclear that the prescribing decision is essentially arbitrary there's a lot more room for drug company gifts to nudge the doctor one way or another.

There are basically three kinds of advertising working in tandem. First, direct advertising to patients and doctors to increase the number of people who believe they have a problem that should be treated pharmacologically. This increases the size of the pie. Second, direct advertising to doctors in the form of gifts and meals influences which drug the doctor will choose once he or she has decided to prescribe a drug. This is the fixed-pie competition between drug companies. Third is that a big part of R&D is actually advertising. When a drug is about to be elligible for manufacture as a generic, the drug company needs to drive people away from the generic, and simple brand-loyalty isn't good enough. By developing a new version of the drug which is arguably marginally better, they can maintain a monopoly on production and sale of the "best" version of the drug and keep doctors prescirbing the "latest and greates" version of the brand name rather than the generic. My understanding is that these "improvements" are often tiny and have almost no impact on efficacy--the perceived marginal value to the consumer is much higher than the actual marginal value.

Merrill,

Your point on mortality being the regulatory requirement for cholesterol-lowering drugs was my point-- it wasn't the case previously with the first cholesterol loweing drugs, including first generation statins.

In either case, in reading your other piece you suggested, I think we are in general agreement on this issue. The timeline you provided on key drug safety legislation and seminal events (e.g. HIV) that changed the "speed to market vs safety" conundrum up and down over time is spot on. As you suggest, the post-Vioxx period does seem to be a swing back towards greater safety, meaning that regulation this decade is more stringent than the 90's. I would suugest that the recent increase in regulatory hurdles has had an impact on cost of trials.

On the other piece, generally, I think your conclusions raise some provocative policy questions that do need to be addressed. A few differences from my perspective, both on facts and perspective:

-- the industry is getting away from the blockbuster model. Not to the other extreme you suggested in personalized medicine, per se, but are more willing to explore drugs in the $300 million to $1 billion, that previously would have been neglected.

-- you made a point on the "breakthroughs" made in science but a problem being translating that into new drugs. I think the breakthroughs part is pretty suspect-- there are a lot of intriguing findings, but little on defining how disease works biologically. In other words, finding new markers or targets associated with cancer is different than understanding how cancer works. I think that the breakthroughs have really yet to come, which limits the treatments that derive from them.

-- I'd think about the problem around diseases around smaller populations and pharma's reluctance to pursue them. This is a general policy problem. A science problem is going to require a certain amount of effort, and hence dollars (whether funded publicly via NIH or in industry) and in an environment of constrained dollars that can be spent on health care, its a legitimate policy question how much we societally should be pursuing things that impact fewer people. My personal view is that it depends on the degree of impact for a given disease, i.e. a life-saving treatment that affect infants is more soceitally worthwhile than one only impacts morbidity later in life. Market demand, in Western countries that can afford to pay drugs, can roughly correspond to this level of impact. The key is paying the industry for the value that it provides in terms of degree and breadth of impact. Unfortunately, this is easier said than done, because when you look at drugs that are lifesaving today that truly are lifesaving such as Cerezyme, which cost $100K per year, the politics are too easy to bring down those prices. But when you look at the impact and few patients in serves, its a reasonable cost compared to many treatments considered commonplace in our health care system. In short, aligning profit incentives with unmet need will gives industry the right incentives to pursue the societally advantageous treatments. If pharma still doesn't pursue them, then its probably the right call-- the same on the NIH would make if they allocated all of the dollars. We simply can't afford to do everything in health care.

A major factor in the cost per drug developed in recent years is that projected scientific advances have not materialized. The human genome project didn't produce the low-hanging fruit pharma companies thought it would, so their chemistry based approaches aren't very useful. They need to make money somehow, so they churn out the silly me-toos. (And, yes, they charge absurd prices for these marginal improvements. But not because the wool has been pulled over the consumer's eyes, but because most consumers don't care what the drug costs in the first place. After all, it's all just a $10 copay.)

That said, if academic biomedical research had been generating blockbuster ideas, biotechs would be eclipsing pharmas. There is no lack of venture capital for good ideas and no lack of MBAs roaming the halls of academia in search of the next big thing. There is a lack of good ideas.