Bad News On December 14, 2012, the US Food and Drug Administration granted an accelerated approval to the drug Ponatinib, which is used to treat patients with serious and life threatening forms of leukemia. Unfortunately, the risks associated with the use of the drug have proved to be far greater than anticipated. Thus on October 31, 2013, the FDA switched course and asked the manufacturer of ponatinib, Ariad Pharmaceuticals of Cambridge Massachusetts, to suspend marketing and sales of the drug pending the FDA’s further evaluation of the potential costs and benefits of ponatinib’s use. The FDA wants to conduct that study so that it can better determine which patients are the ideal targets for the drug, and which patients are most likely to succumb to its grave negative side effects, “the risk of life-threatening blood clots and severe narrowing of blood vessels.”
Accordingly, the ever-prudent FDA immediately recommends the adoption of a three part program. First, no new patients should begin treatment with the drug. Second, those who are using the drug who are not responding to it should immediately cease use. And third, those patients who are responding favorably to the drug should be allowed to continue its use if their health care professionals believe that the potential benefits from the drug use justifies their filing for a “a single-patient Investigational New Drug (IND) application or expanded access registry program while FDA’s safety investigation continues.” Not surprisingly a recent New York Times Editorial takes the view that the FDA has set up the right process for dealing with this problem by allowing physicians and other health care professionals to go through a “streamlined” process to continue drug use where they think that the anticipated benefits exceed the anticipated costs.
Whose Cost Benefit Analysis? Both the FDA and the Times have done the wrong cost benefit analysis. For starters, all the information about the adverse effects of ponatinib are in the public domain, available to all professionals who prescribe the drug. If the FDA had done nothing other than republish the conclusion of the recent studies, we can be confident that the serious risks associated with its use would quickly work themselves through the system. The amount of the drug prescribed for established and new patients would drop, given the increased revelation of the risk. Higher cost means lower use.
The question therefore is what approach will accomplish two ends. The first of these is to decide which patients should continue or initiate a course of treatment with ponatinib. The second is what is the quickest way to target those subpopulations that are most likely to get a positive experience from use of the drug. The correct answer to both questions is to leave the drug on the market. The first point here is that any limited IND program takes time and costs money. There will therefore be some cases in which doctors will be reluctant to go through the proceedings, at which point their best clinical judgment will be derailed by an unneeded administrative process. That is a serious risk because in all cases where the variance in response is critical, the downstream information about individual patients is likely to be far more reliable that the generalized upstream information that is all that is available to regulators. There is no reason to displace an efficient decentralized sorting mechanism with an ineffective centralized sorting mechanism that is necessarily one step further removed from the patient.
Next, if we know that some patients have had positive experiences with a potentially deadly drug, there is every reason to believe that at least some new patients would be suitable candidates for use. The FDA, by preempting the process, blocks individual physicians from making those determinations. The doctors in question need not rely only on their own prescribing experience, but can marshal data collected from all sorts of public sources to increase the confidence of their judgment. The likelihood here is that there are higher levels of social utility with informed downstream judgment than with the FDA request for a blanket cessation. Once again the structural mistake of the FDA is to assume that only its devices for drug safety matter in the world. In fact cooperative actions by physicians and hospitals and HMO groups all contribute data that is of value in running the process, which is one reason why off-label use is so high with cancer drugs.
The benefits from allowing the drug to stay on the market are not only valuable in the short run, but also in the long run as well. Any effort to isolate good from bad cases will depend at least in part on a close examination of the patients who have used the drugs to figure out the markers that correspond to successful and unsuccessful treatment. In general, the larger the population sample, and the longer the individual use, the more valuable this pool of information in helping with the assessment. It should not be assumed therefore that the FDA or manufacturer studies will be as good without that new information as it is with it. The point here is especially true with ponatinib, which is used to treat rare a rare form of leukemia, so that thus far only 2000 people have tried the drug in question. Those numbers are already too small to get reliable information. There is no reason to shrink them further by government fiat.
Conclusion: FDA Go Slow In making all these judgments, I of course have no knowledge of whether the additional work will salvage the market position of ponatinib as a viable drug. The prospects do not look rosy. But there is no reason to make the suspension of the drug the quid pro quo for the accelerated approval. The basic truth remains is that the FDA treads on dangerous ground whenever it uses its gatekeeper power of new entry. I think that so long as drugs are prescription only, those barriers should be torn down. But if they are kept up, it does very little good to enforce the back end of the approval bargain if it makes all present and future sufferers of this leukemia worse off than they would otherwise be. Collective ignorance is not a good reason for collective government action. It is a good reason for the administrative state to hold its hand.