Category: Medicine

A new paper, "Sexually Transmitted Infections, Sexual Behavior and the HIV/AIDS Epidemic"
by Harvard economics graduate student Emily Oster, asks why prevalence
rates for HIV/AIDS are ten to fifteen times higher in Africa than in
the United States. Using a simple model that decomposes infection
levels into differences in sexual behavior and differences in
transmission rates, she attributes the entire difference in HIV
prevalence between the United States and Sub-Saharan Africa to
differences in transmission rates. The intuition, as she writes, is
that "Higher transmission rates produce more infections this period,
and each new infected person can infect people next period, so the
result of a higher transmission rate is multiplied many times over."

One
of the implications of her findings is that lowering transmission rates
by targeting STDs is more cost effective than trying to reduce HIV
prevalence using expensive antiretrovirals or education programs aimed
at changing behavior.

Read more here.

(I am quoting from Randall’s email here, Typepad won’t let me indent beneath the fold…)

"A) For anyone who is dying the lack of FDA approval should not prevent the use of a treatment.
   While I would favor complete revocation of FDA ability to keep drugs off the market that is not going to fly politically. But selling a more limited change where those with, say, a projected 12 month or 24 month life expectancy get a basic "I’m free of the FDA tyranny" card would accelerate drug development. Phase I and Phase II drugs would be more widely available….

D) The coercive power of the state should force a percentage of all income to go into medical spending accounts.
   The goal here is two fold:
      – Decrease the number of people who need state medical funding.
      – Also, increase the amount of medical treatment purchases that are paid directly by patients. That will increase market forces in medicine.

E) Self-employed people should be able to buy medical insurance pre-tax.

F) People should be able to make tax deductable donations into medical spending accounts for their future children before the children are conceived. Then the money should be usable to buy medical catastrophe insurance against the possibility of birth defects and also for more mundane medical care.

G) People should be able to bring their own medical insurance policies with them to a job and have their employers pay on those policies rather than on the employer’s group policy.
   That way a person could move around between jobs and never reach a state where their COBRA runs out and a pre-existing condition makes them uninsurable (assuming they can even afford to make COBRA payments while unemployed). The way things stand now the tax law forces people to go uninsured between jobs. When you have no income coming in you suddenly have to try to get coverage. That problem must be fixed.

H) Medical records should be made electronic and more widely shareable by researchers that most medical patients effectively become enrolled in "virtual" medical trials."

Most of all, the surgeons:

Surgeons who play video games three hours a week have 37 percent fewer errors and accomplish tasks 27 percent faster, he [James Rosser] says, basing his observation on results of tests using the video game Super Monkey Ball.

But the patients too.  Perhaps they could play against each other, no?

Read more here.

Replace the tax deductibility of health insurance for the wealthy and middle class with tax credits for low-income individuals to buy health insurance.

Andrew Samwick explains here and here, try this excerpt:

Who benefits from this deductibility [of insurance]? …the average family with $100,000 or more in income receives a benefit of $2,780. Compare this to an average benefit of $1,231 for a family with $30,000 – $39,999 in income. Because tax rates are higher at higher income levels, and those with higher incomes are more likely to have coverage, the benefit goes up with income…

The portion of this disparity that is due to the progressivity of the tax system is ridiculous. Subject it all to tax, and take some portion of the $100 – $200 billion saved and use it to provide refundable tax credits to purchase health insurance, whether through an employer or an individual policy. The credits should phase out at higher income levels. 

Brad DeLong signs on as well.  R. Glenn Hubbard and co-authors endorse a related but not identical idea in Wednesday’s Wall Street Journal.

The core rationale is simple.  Employers would substitute toward plans where beneficiaries care about costs at the margin.  At the same time, the uninsured would have a greater chance of receiving coverage.

My take:  Whether I press the "yes" button depends on what we compare it to.  It sounds like a good package deal, especially if done in revenue-neutral fashion.  The distributional change is compelling.  It looks less good when broken into constituent parts.

People at the upper end will face higher insurance costs; I suspect at the margin many will be pushed out of group plans into higher-premium private plans.  (I’ve never been convinced that the tax subsidy for insurance, while costly, is the fundamental problem in the health care market.)  Why should we favor this, except as a revenue-raising measure?  On this part of the plan I vote "no."

What about the newly created benefits at the lower end of the income distribution?  Well, I can think of many worse ways for the government to spend its money.  But for me the proposal is neither a first nor a second best.

Many of the uninsured are linked to recent immigration.  But if we are willing to spend money in this fashion (I am), I would rather spend the money taking in more immigrants.  That helps the people who need it most and brings significant health benefits to many, including the recipients of remittances.  (Admittedly emergency room budgets will take a whacking.)  I also would prefer a more targeted program to address the behavioral obstacles that keep many of the poor, insured or not, from seeking preventive care.  So on this second part of the plan I vote "maybe."  It is not my favored way of spending government money, though it is an above-average way of spending the money.

Nobody has a really good health care plan.  This idea remains in the running, and might make sense as a "best politically feasible package plan."  But it is not my first or second best choice, and as a blogger I will demand license to dream.

Today I tie up some loose ends (here is post 1 and post 2 in the series) beginning with an argument in favor of me-too drugs that I do not like and then moving on to an attack on Marcia Angell’s book The Truth About Drug Companies.

Many people argue that price competition is a benefit of me-too drugs.  But recall the point I made yesterday, me-too policy is patent policy.  If you think lower prices are a good idea then you really think that weaker patents are a good idea.  Indeed, as far as price competition is concerned the ultimate me-too drug is a generic so the logic suggests we should get rid of patents.  For obvious reasons, that is not a good idea.  (Our current policy is actually quite good – strong patents for about 12-15 years of effective life followed by very sharp price competition from generics.   Note that since profits are discounted the future competition doesn’t reduce R&D very much.)

I dislike Marcia Angell’s The Truth About Drug Companies not because of her conclusions but because it isn’t serious research.  She has lots of citations to newspapers, for example, but not a single citation to Frank Lichtenberg the premier researcher on the value of new drugs.  (I don’t agree with everything in Jerry Avorn’s Powerful Medicines, and he doesn’t cite Lichtenberg either, but I learned something from his book.  If you are interested in these issues I recommend it highly.)

Concerning me-too drugs, on page 90 Angell says "there is little evidence to support the notion that…if one drug causes side effects, another one won’t."  That’s odd because on page 81 when discussing the me-too statin’s (Zocor, Lipitor, Pravachol etc.) she notes "Bayer’s Baycol had to be removed from the market because, at the approved dose, it caused a deadly side effect."  Hmmm.  Similarly, early tests suggest that Celebrex may not have the same side-effects as Vioxx, despite the fact that both are Cox-2 inhibitors.

Angell is also skeptical that me-too drugs can have different effects in different people.  Frankly, I was shocked at this argument. Every clinical trial that has ever been run demonstrates that the same drugs have different effects in different people – it’s hardly a surprise that different drugs have different effects.  And me-toos are different – different enough not to violate the patent on the innovator drug almost certainly means different enough to have different effects in some people.  My local supermarket carries at least a dozen different styles of peanut butter, a fact of which I approve, but Angell thinks two angiotensin-converting-enzyme (ACE) inhibitors may be one too many (p.90).  Give me a break. 

Finally, it’s important to recognize that small changes can actually make for important improvements.  What could be more me-too than a once-a-day pill replacing a twice-a-day pill?  Yet, to dismiss this change is to overlook the people factor.  A once-a-day regime that people stick to is much better than a twice-a-day regime that people fail to follow.  Forget the chemical structure the economics says a drug that people actually take is a better drug. 

Yesterday I introduced the gold-mine model.  Today, I want to look at some solutions but also ask whether the model fits the pharmaceutical market.

Recall, the problem is that I discover a gold-mine, you undermine my profits by digging on nearby land.  Society loses because instead of searching for another gold-mine you spend resources trying to exploit what has already been discovered.  Applied to me-too drugs the idea is that firm A innovates and earns big profits, firms B,C,D try to imitate and grab some of the profits rather than search for innovations of their own. 

In the gold-mine model one solution is to give the miner who first makes the discovery the mining rights on all nearby land.  The miner won’t exploit these rights but will prevent others from wasting resources through undermining.  How does this apply to me-too drugs?  Critics of the pharmaceutical
industry will probably be upset to find that the analogous solution is to grant stronger patent rights.  In particular, the problem with me-too drugs is companies investing resources in R&D that will end up producing a drug with similar effects through somewhat different means.  If patent rights were broader then the costs of undermining would be higher and the me-too problem reduced.

Thus, me-too policy is patent policy and now we begin to see why the problem is complex.  Broader patents, for example, have costs as well as benefits.  Selden’s auto patent, for example, was originally held to be so broad that it almost finished Henry Ford.  (For more examples see my paper Patent Theory versus Patent Law (email me if you can’t get access) or the new book Innovation and Its Discontents.)

Moreover, we have been assuming that the innovating firm strikes gold and then other firms rush onto nearby land.  But that’s not the way most pharmaceutical innovation works.  More often, there is some basic research, often done in a university lab, which suggests a possible drug target or mechanism.  The research is public knowledge so a number of pharmaceutical firms begin the long slog of trying to turn an idea into a drug.  Think of the original research as a prospector shouting "there’s gold in them thar hills," – the firms then rush into the hills to start researching/digging.  One of them may strike gold first but the others are close behind.

The key point is that the R&D used to develop the me-too drugs was not spent to undermine the innovator it was spent in an effort to become the innovator.  Think about it this way.  Ten people are in a race to deliver a letter.  Critics of me-too drugs complain that the runner coming in second is wasting society’s
time.

Now it is possible to have too many firms racing to be the innovator – perhaps we should only have 8 firms in the race not 10.  But critics of me-too drugs don’t argue that there is too much R&D, which at least would be consistent, they argue that there is too little. 

Although it is possible to have too much R&D, I find the argument especially difficult to believe in the pharmaceutical industry.  First, even in the best scenario the returns to the innovating firm are less than the social returns so "too much" R&D may simply make up for this defect.  Second, there are many positive externalities to drug research.  A substantial fraction of the increase in life expectancy over the past thirty years has been due to pharmaceuticals and the value of this reduction in mortality is in the trillions.  Third, research indicates that the R&D efforts of different firms is in fact complementary – when you drain your mine of water my costs of mining fall.   

Tomorrow I will wrap up with some final comments, Why me-too for you may not be me-too for me.

At Tyler’s prompting I will have several posts this week on me-too drugs.  It’s a difficult area but one that I have been thinking about.  Why are me-too drugs bad, even though competition is ordinarily good?  The reason is that when an industry is characterized by market power and price exceeds marginal cost the private returns to innovation may exceed the social returns.

Here’s a simple model.  I own a profitable gold mine.  You buy some land nearby and begin to dig but instead of finding your own gold you tunnel underneath my land and steal my gold – you undermine my profits.   Now this may be very profitable to you but as far as society is concerned your digging is wasted effort.  You expended resources not to increase production but simply to transfer profit from me to you.  The prospect of stealing my gold attracts too much entry (and the threat of this may in turn cause me to invest less in the first place).

A patent on a blockbuster drug is like a gold mine.  Me-too drugs undermine the profit from the blockbuster.  The R&D that goes into the me-toos is thus a social waste.

The gold-mine problem is the economic case against me-too drugs.

The gold-mine model leads to some surprising insights.  Me-too drugs are bad because resources are used to undermine someone else’s profits.  But a firm won’t undermine it’s own profits.  Thus, a firm’s own "me-too" drugs are not an example of the gold-mine problem.

AstraZeneca introduced Nexium just as their drug Prilosec was going off patent.  Nexium is very similar to Prilosec and for almost all patients it offers few additional benefits – it is widely cited as a me-too drug.  The Nexium problem, however, is quite different from the gold-mine problem.  The Nexium problem is, Why do people buy the expensive brand when the cheap generic would be just as good?  The Nexium problem is that customers think, or act is if they think, that Nexium is not a me-too drug. 

In contrast, in the statin market there is Zocor, Lipitor, Pravachol, Lescol and Crestor.  Even if consumers thought these drugs were me-toos, each of them would still have a market.  Here the problem is, or appears to be, the gold-mine problem.

Solutions to the Nexium problem, such as better informed patients or getting consumers to pay a larger share of their drug purchases, do not necessarily solve the gold-mine problem.    

In a future post, I will look at some solutions to the gold-mine problem.  I will also look more closely at whether the pharmaceutical market fits the assumptions of the model.

Today’s Op-Ed in The Financial Times defends me-too drugs:

There is a tendency among the medical fraternity to tut-tut about the proliferation of drugs as unproductive and unnecessary. Marcia Angell, a former editor of the New England Journal of Medicine, argues in her book The Truth About Drug Companies that me-too drugs are symptomatic of Big Pharma’s intellectual bankruptcy.

Dr Angell says there is "almost no evidence of price competition in the me-too business" and trials "almost never compare me-too drugs with one another for the same condition at similar doses". One or two drugs would do: "I know of no rationale for, say, the seven brand-name Ace inhibitors that are sold to treat high blood pressure and heart failure."

There is something to her critique. One of the attractions of me-too drugs has been that pharmaceuticals companies have been able to market them heavily in the US – spending $3.3bn (£1.7bn) on direct-to-consumer advertising in 2003, and giving $16bn of free samples to doctors. This makes them look more like branded consumer goods companies than research-based scientific organisations.

But there are already signs of drug proliferation leading to falling prices. In Europe, the existence of alternative treatments has helped governments to cap prices. Pfizer is fighting a German proposal to pay the same for all cholesterol drugs, cutting the sum it gets for Lipitor by a third. Even in the US, drug companies increasingly have to compete to get their drugs on to approved prescription lists.

That means direct comparisons on both price and effectiveness. Despite the industry’s traditional resistance to "head-to-head" trials, they are becoming more commonplace. Bristol-Myers Squibb financed a head-to-head study of Pravachol and Lipitor in an effort to displace the market leader, only to find this year that Lipitor worked better.

…This may all sound terribly wasteful to a doctor, but it is the same thing that is good for consumers in other markets: open competition. The problem until now has not been too much of it, but too little. The last thing that a patient should want is a choice of only one drug. As the Vioxx withdrawal shows, me-too pills may inspire little affection, but they would be missed if they were not there.

Why have me-too drugs been so prevalent?  Are they an inefficient form of product mimicry?  An artifact of bad patent or FDA policies?  The long-run path to affordable medical care?  I don’t have the expertise to answer these questions, so in my spare time I bug Alex to write more about them.

1 out of 3 New Yorkers is on Medicaid:

…the states set the level of Medicaid spending, but the Feds match the states dollar for dollar. New York State decided that a good way to soak up extra Federal money was to require the local governments to match the state, dollar for dollar. Since the Feds match all state and local spending, this had the effect of doubling Medicare spending in the state of New York, at no cost to the state, other than the psychic anguish of its highly taxed citizens.

By the end of 2004 Medicaid spending will cost $1650 per person in New York City.

My Arnold Kling-like question: If every distortion of this kind were eliminated, how much would the price of medical care fall?

Thanks to Jane Galt for the fact (and here is a rare photo)…

One way to structure a vaccine comitment would be to guarantee a price of, say, $15-20 per person for the first 200-250 million people immunized, in exchange for a commitment from the developer to subsequently drop the price in the poorest countries to a modest markup over manufacturing cost.  A commitment of this size would offer firms an opportunity for sales comparable to those available in commercial markets.  It would be extremely cost-effective, saving more lives than virtually any imaginable health expenditure.

That is from Strong Medicine, by Michael Kremer and Rachel Glennerster.  The authors have an excellent book and a noteworthy idea, but I have some worries.

Some poor countries, such as Ghana, have quasi-functional government.  But other governments won’t allow this to proceed unhindered.  Remember when some Nigerian states banned the polio vaccine for (supposedly) spreading sterility and AIDS?  That is an extreme example, but how about this?

In Africa, for example, it is estimated that only between 2-15% of children slept under bed-nets in 2001-a simple, effective and proven method to prevent malaria.   

If the cure for AIDS were a single glass of clean water, millions of the infected still would die.

This is why economic development is so hard and so resists formulaic treatment.  Correcting any single screwed up incentive won’t bring as big a payoff as you might think, given how many other things are screwed up.  We have to go one step at a time, but every step brings both short-term costs and political opposition, while not showing much in the way of immediate benefits.

Prizes work best when the prize-giver is aiming at a well-defined end, where success is easy to measure.  This fits "inventing a malaria vaccine" better than "distributing a malaria vaccine."  I would be willing to try this scheme, given the high upside returns.  But it is quite possible we could go ten years or more without seeing much in the way of tangible results, even once something is invented.