I don't believe public funds pay for any large-scale, randomized Phase 3 trials in oncology -- the cost can rise above $100 million per trial! The overhead in terms of patient registration and follow up require the staffing, technology and expertise of a large pharmaceutical industry. And once that step is mandatory (again, in most large Phase 3 trials, exceptions made for pointed therapies like Gleevec), "the market" is necessary, functional or otherwise.

At this point I risk conflating not just two but three arms of the government: the NIH (funding), FDA (approval), and Medicare (payment).

> Why not just cut out the middleman?

Medicare has hinted at doing this, using its purchasing power to drive down the cost of drugs and refusing to pay for treatments approved to work but who have a low cost-to-efficacy ratio. A recent example is the fight over Hepatitis C medication[0]

[0] http://www.npr.org/blogs/health/2014/05/12/307747798/boomers...

Sometimes Walmart goes to a manufacture and says "We want to buy all of your output for the next 3 years, here's exactly how we want you to modify the product, and here's the quality assurance tests you need to do. Here's what we are going to pay." Walmart pays them some percentage up front and then over time as product ships and sells.

Is Walmart paying for the QA testing?

If I understand your analogy correctly, Walmart is Medicare, and the manufacturer is a drug company? If so, then said QA testing is not a guarantee; you cannot design a drug that will extend the lives of patients with end-stage cancer like you design a grill. It'd be like there was a government agency that compared the manufacturer's new grill to all other grills on the market, and only let the manufacturer sell theirs if it had a clearly defined benefit over the others.

I guess I didn't make my point clearly enough either time.

The government pays for the early research, sets all the ground rules for the drug trials, and -- as the drug companies' biggest customer -- ends up ultimately shouldering much of the cost of those trials whether successful or not. The government certainly has the financial wherewithal to fund stage 3 trials, $100M a pop is a rounding error on federal spending or even on just federal medical spending. So why not just go ahead and do so? Why the farce with pretending there's a healthy drug market when the government is driving so much of the process? Let's just admit this is a market failure of the public good variety, like firefighting and roads, and have the government provision it.