What is the Right Number of Unsafe, Ineffective Drugs for the FDA to Approve?

By Rachel Sachs

Later today, the Senate will begin voting on the 21st Century Cures Act, which passed the House overwhelmingly last week. I’ve blogged repeatedly about the Act (most recently here), and many academics and commentators are rightfully worried about the Act’s efforts to lower FDA approval standards in different ways. I write here to put some of these concerns more plainly (and more bluntly), by asking a simple question: what is the right number of unsafe or ineffective drugs for the FDA to approve? I would like to hear the Act’s supporters answer this question. Below, I offer some thoughts of my own on how we should think about and evaluate this question.

More generally, when we think about FDA approval of new pharmaceuticals, we have to consider how the FDA should balance Type I and Type II errors. You may think the FDA ought to focus on minimizing the number of unsafe or ineffective drugs that it approves (minimizing Type I errors). After all, we don’t want the FDA putting its stamp of approval on drugs that harm patients or that don’t work. Over time, this would lead to an erosion of public trust in the FDA as a tool for consumer protection. More generally, this is the entire reason we’ve given the FDA its powers to begin with. Scandals involving unsafe or ineffective drugs prompted Congress to give the FDA more, greater powers over the years, in large part to prevent such products coming to market in the first instance.

Instead, you may think that the FDA should focus on minimizing the number of safe, effective drugs it fails to approve (minimizing Type II errors). In other words, it is worse for the FDA to deny patients access to a drug that is safe and effective than it is for the FDA to approve a drug that later turns out to be unsafe or ineffective. On this view, the FDA should still perform some screening against drugs with significant safety signals or against drugs with no plausible mechanism of action, and perhaps should require post-market surveillance studies, but the FDA ought to be enabling sick patients to access drugs more quickly. This view of the FDA’s role places greater responsibility on insurers, physicians, and patients to gather, process, and act on information about a drug’s safety and efficacy.

Currently, our system is set up in general to minimize the number of unsafe, ineffective drugs the FDA approves (minimize Type I errors). Importantly, on this view the right number of approved unsafe, ineffective drugs is still not zero. The FDA, for all its small-c conservatism, certainly makes mistakes, and so although the “right” number in our current system is something small, it is not zero. The 21st Century Cures Act envisions a system in which the FDA approves many more drugs whose efficacy has not yet been tested in the real world – and thus envisions a system in which the right number of approved unsafe, ineffective drugs is much higher than it is right now, and certainly far higher than zero.

This is a defensible position. It’s one I haven’t often heard advocates for the bill take explicitly, though, and I’d very much like to. If you want to make it easier for drugs to come to market through the Cures Act’s mechanisms, you need to own the consequences: the FDA will approve more unsafe, ineffective drugs this way. More generally, I worry greatly about a system that would change the FDA’s balance of risks without changing the other parts of the system. What happens if you reduce FDA approval standards but don’t empower insurers to reject drugs for a lack of efficacy, or don’t provide additional tools for patients and physicians to gather and process information about new drugs? What happens if the public loses trust in the stamp of FDA approval? The public deserves more answers than Cures backers have given them.

[NB Of course I recognize that this is not truly an either-or debate. It is consistent to require vaccines or other preventive interventions to undergo quite strict safety testing, as they are administered to asymptomatic people, and at the same time speed drugs to market for deadly conditions where patients have no other treatment options. There are already accelerated approval systems in place to help accomplish this latter goal today, systems which may account for a larger percentage of the unsafe, ineffective drugs approved now. However, the 21st Century Cures Act seeks to reduce approval standards more generally and most of its provisions do not clearly distinguish between types of conditions in the way that our current systems do.]

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.