Over at Marginal Revolution, Alex calls for moving the FDA to a “safety-only” system, which means eliminating efficacy testing in Phases II and III as a condition of marketing approval. He cites both his own work (gated) and a recent WSJ op-ed by Boldrin & Swamidass (gated). Essentially, this is calling for fundamental changes to the 1962 FDA Amendments.
As Alex and others point out, Phase III studies are not required for off-label use. Companies can bring a molecule to market if they can prove efficacy for anything, even a quite narrow indication. In the wake of Sorrell v. IMS Health (my amicus brief here), we can expect the FDA to relax rules against off-label promotion. The Supreme Court just radically adjusted the regulatory balance between the companies and the FDA, indirectly weakening the power of Phase III. Perhaps we should evaluate these changes before proposing more fundamental reforms.
If Phase II and III trials were no longer required, many more drugs would reach the market. The FDA approves about two dozen NMEs a year, while the number of drugs abandoned in Phases II and III are more than an order of magnitude greater. In a system without an efficacy hurdle before marketing approval, the FDA might approve many more drugs in the next few years. In the future, that number could increase, as companies rush to bring molecules to market in a much less expensive regulatory environment. The industry would solve the “patent cliff” problem almost overnight.
But would this flood help population health? Only if the new drugs are both safe and effective (leaving the cost-effectiveness question for another day). So the key to any proposal in this area is creating, disseminating and using unbiased efficacy data, with some public goods characteristics.
In the current system, the drug sponsor privately produces most of this data as a condition of marketing approval. One strength of this system is the private provision of the public good, with the best-informed and most-motivated company in charge. The weaknesses include conflicts of interest (bias in company-funded studies) and the government’s role in setting the height of the efficacy hurdle and the minimum acceptable methodology for clinical studies. The left focuses on the former, while the right the latter.
One alternative is private provision without regulatory hurdles. But drug companies will not invest hundreds of millions of dollars in studies that might force their drug off the market. Merck learned this lesson with its post-approval Vioxx studies. Competitors might sponsor comparative studies to build market share, but they can do that now and few do.
The government could offer alternative regulatory carrots and sticks for creating these public goods. Existing models include paying companies for conducting pediatric studies by granting six additional months of marketing exclusivity. Just this month, Merck won an additional six months for Lipitor, worth more than a billion dollars in profits, by studying atorvastatin in children. What is missing from this program and similar ones is any sense of cost-effectiveness. The government delayed generic approval of Lipitor, at a huge cost to public and private payers, in order to “purchase” pediatric studies that might have cost a small fraction of that amount. Direct funding of pediatric studies by the NIH would make more sense, and we’d get studies on drugs that kids really need instead of drugs that happen to sell well to adults.
The proposal by Boldrin & Swamidass doesn’t suffer from this particular cost-effectiveness problem. They suggest mandatory generic pricing at market approval, followed by market pricing once the companies provide efficacy studies that meet FDA standards. Essentially, this is a two-step process: Phase I and II trials are required for marketing approval and Phase III trials are required before the company earns any profits. Companies will certainly conduct efficacy studies under these conditions, at a price no worse than the existing system. The primary danger to public health will be the safety risk with widespread use of drugs before Phase III trials are completed. The public health payoff would come if safe and effective drugs make it to the market more quickly. For the companies, this pricing system may be a marketing disaster. Mandatory generic pricing at marketing approval will make it difficult to charge full market prices a few years later when the Phase III trials are approved.
Another alternative is public provision of efficacy studies. Alex supports this idea (gated), as do many on all sides of the political spectrum, but the recent politics of government-funded comparative effectiveness studies suggests practical difficulties in funding and political interference. Efficacy studies are currently funded largely through an off-budget process in the sense that companies recoup clinical trial costs through insurance reimbursement from Medicare, Medicaid and private insurance. Funding efficacy studies directly from the NIH could eventually save money in health care drug budgets, but the CBO is unlikely to score it that way in the short run. Nor can we be confident that companies wouldn’t use their political power to skew the direction, scope and magnitude of these studies. I see public provision as a supplement to the current system, not a full-scale replacement.
Finally, the US could simply free ride off the studies produced to satisfy Europe’s Phase III approval process. That would work only so long as the EU didn’t make the same changes.
Bill Gardner shares his thoughts on Alex’s post here.