(If you haven't, read Part 1 here.)

Though I divide the solution into a "Doctor Side" and "Pharma Side," it is imperative that both solutions be executed simultaneously.  Doing only one will absolutely fail.

Doctor side:

Pharma does not make meds for patients, it makes them for doctors-- they represent the demand.  Read that again, that's Axiom #1.

Doctors have no current incentive or pressure to consider the cost (effectiveness) of the meds.  
But without the pressure, there is no incentive from Pharma to create products that are cost-worthy.  If doctors don't have to consider cost, then Pharma can effectively get doctors to add on, say Nexium.  It works, so why not?  If doctors must explicitly consider cost, then not only can't Pharma successfully market Nexium, it will not even bother to create it.  Pharma will work on something that's really worth the money.

Consider also that unlike other consumer products, price has no relationship to relative value.  Nexium and Lipitor are the same price, but (arguably) Lipitor is more important.

So doctors need to consider cost, which in turn will force Pharma to consider cost.  So doctors-- not patients, President Bush-- must be given a healthcare budget, specifically a pharmacy budget.  $20 per patient per day.  Go.

That changes the market.  If you do that, prices come down, especailly for "luxury goods" (e.g. Nexium.)  And Pharma will create wonderful things (not that they haven't already.)



Pharma Side

Axiom #1 is: Pharma makes meds for doctors, not patients.

Corollary:  They don't need to make a drug that is useful, or is awesome; only a drug that doctors will prescribe.  Sometimes the two are the same, but that isn't by design.

Pharma gets no points, no credit, for creating drugs that work, only drugs that sell.  No one I know has hugged a Lilly rep, thanking them for having a drug that works, even if imperfectly.  Their only thanks is the money.

The single problem from the Pharma side is the blockbuster drug model. 

The common criticism against the blockbuster model is that it entices other Pharma companies to invent "me too" drugs-- another SSRI, another statin, another Nexium.

But there is a much greater, critical, consequence of the blockbuster model: it makes doctors think that the mechanism of action of the blockbuster is the only, or most, important one-- it creates a paradigm that is hard to think outside of.  In other words, the blockbuster model confuses science.  It may be that lowering cholesterol is itself a red herring, and that the actual benefit is something else-- consider Vytorin lowered cholesterol more than simvastatin alone, yet was not better at preventing intima thickening.  But because cholesterol drugs-- nay, HMG co-A reductase inhibitors specifically-- are the big drugs, that's all doctors think about.

From 1980-1998, SSRI were all psychiatrists thought about.  So obsessed were they with SSRIs that they tried to explain nearly all psychic phenomena by serotonin.  Depakote was such a blockbuster that people couldn't even comprehend a "mood stabilizer" that wasn't an antiepileptic. 

You can't get a novel mechanism passed an NIH grant reviewer; Pharma isn't interested either.  How is it that 6 atypical drugs all have prominent serotonergic activity, but no one investigated glutamate?  You have a  working paradigm, you have stuff that is well established, it's hard to abandon it and try something new.  Not when you operate on the blockbuster model.

And meanwhile, Pharma loses out on new opportunities.  Pharma may have already invented a novel mechanism drug  that reduces heart attack risk-- but not only does it have to bring it to market, it has to retrain doctors that they trained to be statin obsessed-- that cholesterol, after all, isn't everything.  Strattera and Cymbalta-- both invented at the same time as Prozac-- languished in Lilly's basement because the world (of doctors) was not ready to hear about drugs that weren't "selective" or serotonergic.

But as long as doctors don't have to consider cost, blockbusters are the best way to make money.  You want to make a drug that doctors will simply add on to everything-- and they will, if they're not paying.   

The Incentive Model:

Broadly, there are three categories of people working from different incentives.

1. For Pharma researchers, the incentive is to get FDA approval for an indication.
2. For reps, it is to get market share. Many reps compete against other reps in their own company.  (e.g. if they come up with a great sales pitch, or have an awesome speaker, etc,  they don't want to share them with other territories.)
3. For managers, it is indirectly market share, more directly certain rep based metrics (reps made growth targets, reps conducted the right number of programs, all expense reports were done on time, etc.)
4. For the company itself, the current incentive is to create what doctors will prescribe.
   

It's obvious that the incentives are different, and none are actually in alignment with the company's goal of increased revenue.

1. Bringing a drug to market should not be incentivized, even from the basic profit perspective.  Just because it gets an indication, doesn't mean it will generate any money.  But since researchers are incentivized precisely on that, you see a lot of obvious "me-too" drugs and indications. The fix here is to incentivize researchers based on the future success of the drug, not the indication.  Coupled with a pharmacy budget for doctors, the incentive will be to invent a drug of value-- whether in a disease that has few treatments; or a significantly better/safer alternative.  While more difficult, it will be more profitable to the researcher than another SSRI (which doctors won't want to spend their budget on.)

2. Market share is also a bad metric, yet it's the one everyone uses.  It does not matter at all that Seroquel has more market share than Zyprexa because they may not be used for the same things.  It's not Coke vs. Pepsi.  As an example, for Lilly, Zyprexa competes against Geodon, but not against Depakote-- because it doesn't share the same FDA indication-- even though in the treatment of bipolar, Depakote is a much bigger competitor than any antipsychotic.  If Zyprexa wins all of Geodon's market share, it gains very little in real dollars.  But convert all the Depakote to Zyprexa, and Lilly wins.

3. Managers.  I am not really sure if there is any value to managers.  I am not being glib or insulting, and I'm open to information.  But my read is that Pharma could easily cut the number of managers in half, reducing expenses but also freeing reps to focus on selling.  Or, it could use a mentor model where good or senior reps are paid extra to mentor newer reps.

4. Next, you have to incentivize Pharma to make valuable drugs; at least don't de-incentivize.  There is a gigantic likelihood that any really useful drug will be commandeered by the NEJM's Marcia Angell and her band of merry socialists.  Why would any company would want to spend any money on a cure for HIV, when there is a real chance it will be stolen by the UN even before it gets approve?.  And we're supposed to accept that, because society "needs" it.  "Sanction of the victim" is what Ayn Rand called it (and I'm confident I just lost readers by invoking her name.  Bite me.)   So instead, you know what Glaxo's next big drug is?  Treximet: Imitrex + naprosyn.  Commandeer that, Dr. Angell.


The Research Model:

The current model is completely broken.  Allowing "thought leaders" to peer review grant proposals is as good an idea as allowing senior Senators to choose your next President.   Letting Pharma do it is like letting Coca-Cola decide your breakfast cereal.

I recognize that it can't be fixed all at once, so I propose an incremental solution.

Pharma puts money into a pot to fund independent research, no strings attached.

They can still fund their own stuff, of course, but we need a pool of non-taxpayer capital that private research can use. I recognize this is an expense for Pharma-- but not huge, NIMH budget is $1.5B-- but it earns considerable respect, and will likely lead to new ideas and directions-- and someone at Pharma will ultimately benefit. 

The problem is who will manage this pot: putting it in the hands of "thought leaders" and academics is stupefyingly obviously a bad idea.  You may as well let the Politburo decide.  (Yeah, I said it.)  Want to know what that would look like?  The NIH.

The solution is a "Digg" style voting of research projects.  Any doctor-- not just psychiatrist, because you need people with different mindsets to make good evaluations-- can vote up or down a research idea/protocol.  If you want to get fancy, they can also vote up the amount the study gets (as opposed to simply approving an amount.)  It's possible that allowing other scientists to also vote could be of benefit, but there are some problems with it that I haven't worked out.

Again, Pharma and the NIH can continue the same biased, barely readable quasi-research they have always done, this is just another funding source.