When Dr. von Eschenbach's oped, Medical Innovation: How the U.S. Can Retain It's Lead, ran last week there was a lot of positive feedback - and plenty of good questions about how his proposal for new pilot programs for breakthrough therapies would, in fact, operate.
Several commentators mistakenly thought that Dr. von Eschenbach was calling for the wholesale repeal of the 1962 Kefauver Amendments, which require the FDA to test new drugs for efficacy. They argue that this move would unleash a wave of placebos on the country - expensive drugs that would be useless, or, worse, dangerous.
Take this Forbes article, by former Pfizer scientist John LaMattina:
Dr. von Eschenbach's position is untenable in a variety of ways. First of all, for a patient, physician and payer to accept a new drug, they all have to be convinced that the risk in taking the drug is worth the benefit. Imagine a doctor telling her patients: "I have a new drug to treat your migraine. I have no idea if it works. But it seems to be pretty safe." The doctor might just as well prescribe a placebo.
This complaint was echoed by Pharmalot:
In other words, this approach envisions substituting epidemiological studies for double-blind, randomized clinical trials. This is ironic, because Andy laments that the FDA is losing its status as the gold standard among regulators, which was the rationale for writing his missive. ...
Of course, the pharmaceutical industry would benefit, since fewer trials would be required prior to approval, which would lower development costs. Meanwhile, payers - both private and public - would provide coverage for medicines, even though efficacy remains unknown until post-marketing study results are reported.
Derek Lowe, who actually labors in the trenches in drug development, was more open-minded, but still skeptical:
One is to translate [Dr. von Eschenbach's proposal] into less editorial language and propose that "Patients (and their insurance companies) should be able to pay to try therapies before they're proven to have worked, as long as that proof is forthcoming". That's not prima facie a crazy idea, but it's subject to the same sorts of objections as [Andy] Grove's earlier proposal. The post-marketing data will likely be of lower quality than a properly run clinical trial, and it will be harder to use it to establish efficacy. On the other hand, useful therapies would get into the hands of patients faster than happens now, and the expense of drug development would (presumably) go down. But useless therapies would also get into the hands of patients faster than happens now, too, and that's something that we're not currently equipped to deal with. ...
These are all fair questions following an 800 word op-ed that introduced an idea that requires explanation.
Let's take a stab at a few responses here to the key objections:
Efficacy: Dr. von Eschenbach is not saying that we should junk the FDA's efficacy requirements. What he is saying is that science is advancing to the point where we can, for some breakthrough therapies like regenerative medicine, begin to characterize medical products to such a degree that the traditional Phase I, Phase I, Phase III paradigm (including large, placebo controlled randomized trials) is outmoded.
For these products "proof of concept" can be established in preclinical studies that define mechanism of action in a way that is predictive of both effectiveness and risk. Subsequent small first-in-human (phase 1 clinical trial) can confirm not only the predicted safety but also the expected benefit.
This is the first step in providing regulators and companies sufficient confidence that they are safe and effective enough to be marketed in a controlled environment for unmet medical needs. The second step is to have in place a rigorous postmarket surveillance mechanism that validates the expected benefits and risk of the therapy. In fact, such a mechanism of prospective validation is far superior to the current system of passive post market safety reporting.
Creating an innovative regulatory pathway for selected products is not without precedent. Twenty years ago, in the midst of the AIDS crisis, regulators and patient groups came together to create an Accelerated Approval pathway based on surrogate markers that is vulnerable to many of the same objections raised to Dr. von Eschenbach's current proposal. The Accelerated Approval pathway did not result in reckless regulation but rather saved more lives and spurred tremendous innovation in AIDS medicines, and, later, cancer.
Accelerated Approval (and later Fast Track) changed the criteria for regulatory review and approval based in a rational way that saved lives while maintaining quality. Standard setting for the next generation of novel therapeutics for unmet medical needs should proceed with that same spirit of innovation and collaboration. For these therapies - as for targeted cancer therapies - evidence of efficacy even in proof of concept trials (or however we want to describe them) should be significant and readily apparent. With this new pathway, the FDA should also gain additional authorities to remove a product if efficacy or safety issues later emerge that sharply shifts the risk/benefit balance of the therapy.
Registry/postmarket-surveillance: Critics complain that the industry does not complete some postmarket trials now, with the perception that the FDA doesn't have the authority to enforce postmarket requirements - or doesn't want to anger patient groups by removing treatments from the market it considers ineffective. Registries would strengthen the FDA's position be collecting and disseminating evidence on efficacy (and safety), and make patient's groups more comfortable with drug withdrawal when necessary. Information collected through registries could also help identify sub-groups of patients who might benefit from the therapy, allowing for a more nuanced regulatory approach. Later, follow on diagnostic tests - as have been developed for Tysabri and PML - would further strengthen the ability of patients and clinicians to judge for themselves the risk/benefit balance of the therapy on a case by case basis.
Payment: Who pays? Who's going to pay for drugs that don't work? The short answer is no one. But that is not what we're proposing, and companies that couldn't produce innovative therapies with clear patient benefits wouldn't utilize this pathway anyway. Of course, in our current system, some therapies (or procedures, like back surgery) and diagnostic tests are provided to thousands of patients who don't benefit from them - or worse, are harmed by them.
The paradigm that we're proposing would arguably change this for the better because it would target the highest need patients, and the best cutting edge science. We'll be taking risks in these areas, with rapidly evolving science, but these are the areas that we should be taking risks in.
Some insurers are already using their own internal registries that collect safety and efficacy information and use that information to create care guidelines and limit (inappropriate) off label use. This is simply good medicine as well as smart cost containment. On the other hand, some biomedical companies might balk at sharp restrictions on marketing and may worry that registries could detect false signals of problems that lead to nuisance lawsuits. These are legitimate concerns.
This pathway would have to be carefully crafted to not only be a registry but have a scientific rigor in data analysis to separate signal from "noise", much like sophisticated algorithms monitor credit card fraud. Furthermore, patient registries could be strengthened by patients groups and industry working together to utilize mobile health devices to collect real time data and patient reported outcomes to improve product utilization and safety.
The pathway should also be crafted to ensure fully informed consent for patients and tort protections for participating companies - both of whom are, in effect, submitting to extensive data collection in exchange for early market access to life saving and health-enhancing therapies.
Dr. von Eschenbach's proposal has stirred up some controversy, but we welcome debate.
It's easy to forget how conservative the old FDA became in the wake of the thalidomide tragedy - when new drugs were frequently introduced first in Europe, and reviews of drug applications in the pre-PDUFA era could take two and half years or more. It was said that speeding up drug reviews through PDUFA would lead to a safety catastrophe - but an in depth study by the Milken Institute of safety before and after the introduction of the Prescription Drug User Fees (which substanially accelerated reviews) demonstrated that just the opposite occurred--the benefits to patients swamped any potential problems from accelerated review.
We should also recall that thalidomide was later brought back to market - with the encouragement of the FDA - to treat complications from AIDS, leprosy, and later, cancer. The drug itself is not bad; it simply inhibits angiogenesis. Once you know how it works then you know how to use it - and when to avoid it. Better science, in effect, taught us how to save the baby, save the drug, and save the cancer patient.
Developing a novel pathway for innovative therapies and finding ways to sharply reduce the time and costs required for drug development (particularly for the most innovative therapies) will be neither easy nor quick. But we should not allow ourselves to be distracted by caricatures that such a pathway will "flood" the market with placebos that insurers would be required to pay for.
Companies know that they cannot market their drugs without sound science behind them, and that the public, insurers, and policymakers will never accept sham science. The question we are asking is simple: Are we using the best science - at the FDA, in industry, at the NIH - to advance innovation as quickly, safely, and effectively as possible? We believe that we can do better, and will continue to advance this argument in future opeds, meetings, and white papers.