The approval of new drugs and medical devices is a process fraught with scientific, political, and ethical landmines. Inherent in any such process is an unavoidable conflict between rigorous science and safety on the one side, which tend to slow the process down by requiring large randomized clinical trials that can take years, versus forces that demand faster approval. For example, patients suffering from deadly diseases demand faster approval of drugs that might give them the hope of surviving their disease, or at least of surviving considerably longer. This is a powerful force for reform, as evidenced by HIV/AIDS activism in the 1980s and 1990s that led to the development of fast-track approval mechanisms for drugs for life-threatening conditions, a change whose effects have been mixed. It’s also a powerful force potentially for ill, as I’ve documented in my posts about the understandable but misguided “right-to-try” movement. After all, what politician can say no to a constituency representing desperately ill people who only want a shot at survival? It’s not all desperate patients, however. Also wanting more rapid drug approval are powerful business interests in the form of the pharmaceutical and medical device industries, for whom the time and expense of prolonged clinical trials eat into profits and make some drugs not worth developing from a business standpoint.
In 1962, after Frances O. Kelsey, MD, PhD (who unfortunately died a week ago at the age of 101) successfully prevented the approval of the drug thalidomide in the US, a drug found to cause serious birth defects, Congress passed the Kefauver-Harris Drug Amendments to the Federal Food, Drug, and Cosmetic Act. These amendments required that drug companies not just show safety before their drugs could be FDA-approved, as had been the case prior to the amendments, but also to provide substantial evidence of effectiveness for the product’s intended use. That evidence had to be in the form of adequate and well-controlled clinical trials, which at the time was considered a revolutionary requirement. (Believe it or not, no requirement for high quality clinical trials existed before 1962.) This led to the current system of phase I, II, III, and IV clinical trials in force in the United States today. The amendments also included a requirement for informed consent of study subjects and codified good manufacturing processes, as well as the requirement that adverse events be reported. This has been, with some tweaking over the years, the law of the land regarding how the FDA approves drugs for specific indications
Medicine is a lot more complex now than it was in the 1960s however, and there has been a growing sentiment that the system is, if not broken, at least functioning in a way that is behind the times, a manner that was acceptable and appropriate 40 years ago but is no longer so in this era of genomics, precision medicine (formerly known as “personalized medicine”), and targeted therapies. The new drug approval process, which can take up to a decade and cost a billion dollars, it is argued, is too rigid, cumbersome, and slow for the 21st century. (Why it wasn’t too rigid, cumbersome, and slow in the 20th century, no one seems to say. I guess that “21st century” sounds way cooler.)
Into this ongoing controversy have marched Rep. Fred Upton (R-Mich.) and Rep. Diana DeGette (D-Colo.), who have sponsored a bill passed by the House of Representatives in a rare display of bipartisanship in July. The bill, H.R.6, is entitled the “21st Century Cures Act“. Given how it passed the House by a vote of 344-77, one would think that it should glide through the Senate easily. Certainly, its sponsors and supporters have mounted a mighty PR effort. That might not be the case, given that in the Senate a single senator can hold up or even kill a bill through a filibuster, and to shut down a filibuster or prevent a threatened filibuster requires 60 votes. Be that as it may, I’m not so much interested in the politics of this bill, which, if it survives the Senate, will almost certainly be significantly amended, but rather what the bill does.
I’ve been meaning to write about the 21st Century Cures Act since I first read about it in a news report a few months ago. As with most bills, there’s some good, there’s some bad, and there’s some ugly. Do the bad and the ugly outweigh the good? Sadly, they do. The 21st Century Cures act is nothing more than the old, vinegary wine of promises that deregulation will deliver us cures to diseases made for years by libertarians looking to weaken the FDA in a sexy-looking new bottle wrapped in “precision medicine.” Let’s take a look.
21st century cures: The (somewhat) good
Perhaps the best thing about this bill is that it increases the budget of the National Institutes of Health (NIH). Achieving agreement on this was not easy, as many Republicans oppose any sort of “expansion” of government, and the overall zeitgeist in Congress is to slash, not increase funding (except, of course, for the military). First, the bill mandates a 3% real increase in the NIH budget per year for the next three years. Second, it provides an additional $1.86 billion a year in the form of an “innovation” fund that would primarily support young scientists and precision medicine. Those of us who run laboratories competing for NIH research funding know that the NIH budget has been, in essence, frozen for nearly a decade now. The tiny increases that have been approved have not been enough to keep up with inflation, leading to a 20% decline in the NIH budget in inflation-adjusted dollars since 2003 with a concomitant decline in the number of the R01-equivalent grants (the R01 is the gold standard, multi-year grant that every NIH-funded researcher strives for) of 30%. In 2013, the sequestration added insult to injury with a disastrous additional effect on the NIH budget.
This decline in funding and opportunity for R01 grants, which give labs stability, has led to great concern that the leadership of the US in biomedical sciences will be eroded irreparably. In the “trenches” (where I am), the mood is very, very gloomy, with scientists expressing increasing degrees of despair and pessimism and NIH director Francis Collins warning that budget cuts put US science at risk and that China and other countries will soon surpass us. In particular, it is definitely not a good time to be a researcher in medical academia, particularly if you are a young investigator whose application for tenure depends upon being awarded NIH or NIH-equivalent funding. The addition of more money to the NIH budget and a fund, part of which will be targeted to young investigators, can only help that.
As for the fund for precision medicine research, I’m of a mixed mind about this. There is no doubt that this is arguably the most exciting time in history to be in biomedical research, which is why the dearth of NIH funding makes it simultaneously the most frustrating time in history to be in biomedical research. We finally have the tools to delve the depths of the mysteries of the genome and use the information to design better treatments for cancer and a wide variety of other diseases, but we lack the resources to use them. It’s like owning a brand new Maserati but there’s no gasoline to be found to drive it, or maybe having the biggest, baddest car of all in the world of Mad Max but having to fight for precious gasoline to run it.
That being said, I also think that precision medicine has been hyped far beyond what it could ever possibly deliver. Definitely, with President Obama’s new precision medicine initiative, precision medicine is the buzzword everyone likes to use. It’s often sold in near-miraculous terms about what it will do, in much the same way the human genome project was sold in similar terms in the 1990s, just as the “war on cancer” was sold in the early 1970s. Results didn’t match the hype, which is why on the tenth anniversary of the publication of the human genome and the 40th anniversary of the “war on cancer,” we saw articles declaring both “failures”. I’m guessing that every five or ten year anniversary after this, we’ll see similar articles. I can’t wait to see the disappointed and pessimistic articles about the “war on cancer” in 2021, its 50th anniversary. I’m being facetious, but for a serious purpose: Precision medicine can’t possibly live up to the hype. That doesn’t mean I don’t consider it a promising avenue, but I have no reason to expect that progress in precision medicine will be anything other than incremental. Maybe it will be faster than in pre-precision medicine days, but it will be incremental nonetheless. Besides, all the real hype has moved on to immunologic therapies of cancer, which, if you believe Fred Hutchinson Cancer Research Center president Dr. Gary Gilliland, who declared that cancer is “running scared” and that, thanks to new immunologic therapies, “it is actually plausible that in 10 years we will have cures and therapies for most, if not all, human cancers.”
Yes, I’m being a bit sarcastic, but if you want to see skeptical takes on precision medicine, I recommend a recent op-eds by Michael Joyner in the New York Times about “moonshot medicine,” Rita Rubin in JAMA declaring precision medicine to be more about politics, Cynthia Graber in The New Yorker, and Ronald Bayer and Sandro Galea in the most recent New England Journal of Medicine. Basically, the number of conditions whose outcome can be greatly affected by targeting specific mutations is relatively small, far smaller than the impact likely would be from duller, less “sexy” interventions, such as figuring out how to get people to lose weight, exercise more, and drink and smoke less. I remember how in the 1990s Judah Folkman’s work targeting tumor angiogenesis was touted as a “silver bullet” to treat cancer in just the way cardiologist Eric Topol is quoted by Rita Rubin in JAMA as saying about precision medicine:
The science is catapulting forward. You can now come up with so much big data on each individual, and it is changing medicine.
“However, it’s not just about the sequence,” Topol added. “That’s just one layer of the information. If you really want to change medicine, youhave to have all the information on the individual.” That includes their environment, the bacteria in their gut, and other distinguishing characteristics, he said.
The problem is whether focusing in the genetic underpinnings of disease will provide the “most bang for the buck,” given how difficult and expensive targeted drugs are to develop. However, if getting some money for precision medicine is the price for an increase in the overall budget of the NIH and funneling more funding to young researchers, I’m willing to pay it. My preceding passages aside, I’m actually optimistic about the potential of precision medicine; I just don’t think that it will result in Star Trek-like treatments any time soon. (I remember several other “cures for cancer” touted during my professional lifetime that weren’t.) I’m just more cautious in my optimism, as I pointed out when I answered the question, “Why haven’t we cured cancer yet?”
There are other bits of “good” in the bill, such as a provision making deidentified data from NIH-funded clinical trials more easily available to researchers and facilitating collaborative research, speeding up the review of new vaccines by the Advisory Committee on Immunization Practices, requiring a strategic plan every five years to identify research opportunities and strategic focus areas, reducing administrative burdens on researchers, and limiting the term of office of directors of centers and institutes to five years (with reappointment by the NIH director possible). The act also requires the directors of each institute to “establish programs to conduct or support research projects that pursue innovative approaches to major contemporary challenges in biomedical research that involve inherent high risk, but have the potential to lead to breakthroughs” and “set aside a specific percentage of funding, to be determined by the Director of NIH for each national research institute, for such projects.” I’ve been a bit skeptical of such initiatives in the past, and such orders have gone forth before, usually to be lost in the NIH bureaucracy, but I have no objection to their being in this law. Another potentially useful part of the bill is a “sense of Congress” that the “National Institute on Minority Health and Health Disparities (NIMHD) should include within its strategic plan ways to increase representation of underrepresented communities in clinical trials.”
The (mostly) bad and the (very) ugly
The central question regarding my supporting the 21st Century Cures Act is whether the decent—but not spectacularly so—provisions in the bill outweigh the bad and the ugly. To answer this question, let’s take a closer look. In examining this bill, it is important to note the central assumptions behind the bill. The first, as mentioned in the previous section, is that “precision medicine” will save us all. The second is that we have to get “creative” in loosening regulatory requirements. In reality, there are several provisions in this that bill would markedly undermine the scientific basis of drug approval in this country in the name of “flexibility” and speed. Indeed, the scariest part of the grant is the section entitled “Subtitle D—Modern Trial Design And Evidence Development“, which requires the FDA, in essence, to water down its standards for approving drugs.
Some of this is not inherently bad. For instance, the act would require the FDA to incorporate “the use of adaptive trial designs and Bayesian methods in clinical trials, including clinical trials proposed or submitted to help to satisfy the substantial evidence standard under section 505(d) of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 355(d)).” We are, of course, fans of Bayesian methods, and I’ve discussed adaptive clinical trials before, such as the I-SPY 2 trial. Basically, an adaptive clinical trial uses patient outcomes to immediately inform further treatment decisions. They often involve:
- Interim analyses to stop or to adjust patient accrual
- Interim analyses to assess stopping the trial early either for success, futility or harm
- Reversing the hypothesis of non-inferiority to superiority or vice versa
- Dropping arms or doses or adjusting doses based on outcomes
- Modification of the randomization rate to increase the probability that a patient is allocated to the most appropriate arm
Another hot area in adaptive design is the use of biomarkers as the basis for adapting the trial. In fact, the FDA is already moving to allow for this sort of trial design in its regulations and has issued draft guidelines for such designs for clinical trials of drugs and biologics as well as medical devices. Also, as a 2008 review of adaptive trial designs notes:
It is a concern whether the application of adaptive design methods may have introduced some unexpected operational bias, which may have an impact on accuracy and reliability of the obtained statistical inference. It is also a concern whether the application of adaptive design methods is able to control the overall type I error rate at the prespecified level of significance. As a result, regulatory agencies such as the US FDA request strategies or plans for preventing possible operation biases and for controlling the overall type I error rate be developed. In addition, the sponsors are encourage to begin a dialogue about adaptive designs with medical reviewers and/or statistical reviewers from the regulatory agencies as early as possible (preferably at least a year) before the planning of a trial.
A type I error is defined as the rejection of a null hypothesis that is actually true. In other words, it’s a false positive that finds that the intervention tested “works” when in fact the results in the experimental group are indistinguishable from the control. Obviously, adaptive designs are no panacea. They’re designed to get faster results, but they do so at the price of increasing the risk of false positive conclusions and positive study results that are difficult to interpret. However, it might not be an unreasonable trade-off to accept methods that have a higher risk of false positives for the payoff of faster approval or needing fewer clinical trial subjects. That is more a value judgment than anything else and depends on how much the false positive rate is elevated.
Here’s where things get “badder” and even ugly, in Section 505F, entitled, worryingly enough, “Utilizing evidence from clinical experience“:
(a) In General.—The Secretary shall establish a program to evaluate the potential use of evidence from clinical experience—
“(1) to help to support the approval of a new indication for a drug approved under section 505(b); and
“(2) to help to support or satisfy postapproval study requirements.
“(b) Evidence From Clinical Experience Defined.—In this section, the term ‘evidence from clinical experience’ means data regarding the usage, or the potential benefits or risks, of a drug derived from sources other than randomized clinical trials, including from observational studies, registries, and therapeutic use.
A homeopath would love this provision, and, I’m sure, so would drug companies. Why bother with the time, bother, and expense of those pesky clinical trials to get your drug approved for additional indications, when you can rely on clinical experiences based on therapeutic use, uncontrolled observational studies, or registries instead? If I were the CEO of a pharmaceutical company, I’d love it. Indeed, the one thing this provision most definitely does not do is to speed effective treatments to patients. Rather, it smacks of being a payoff to pharmaceutical companies.
Next up is a section that takes a good idea too far, namely the use of biomarkers in clinical trials. Biomarkers are basically measurable indicators of the presence or severity of a disease or condition – it’s anything that can be used to assess or follow the progress of a disease or treatment. One common biomarker is the prostate-specific antigen (PSA), routinely used to follow prostate cancer for response to treatment and recurrence. For purposes of the 21st Century Cures Act, a biomarker is defined as, “a characteristic (such as a physiologic, pathologic, or anatomic characteristic or measurement) that is objectively measured and evaluated as an indicator of normal biologic processes, pathologic processes, or biological responses to a therapeutic intervention; and (B) such term includes a surrogate endpoint.” In this bill, biomarkers can be used as the basis for drug approval:
“(9) SURROGATE ENDPOINT.—The term ‘surrogate endpoint’ means a marker, such as a laboratory measurement, radiographic image, physical sign, or other measure, that is not itself a direct measurement of clinical benefit, and—
“(A) is known to predict clinical benefit and could be used to support traditional approval of a drug or biological product; or
“(B) is reasonably likely to predict clinical benefit and could be used to support the accelerated approval of a drug or biological product in accordance with section 506(c).
Now, this in and of itself is nothing new. Drugs are provisionally approved on the basis of biomarkers frequently, such as drugs that lower cholesterol. However, as Jerry Avorn and Aaron S. Kesselheim noted in the New England Journal of Medicine:
The bill would also encourage the FDA to rely more on biomarkers and other surrogate measures rather than actual clinical end points in assessing the efficacy of both drugs and devices. The FDA already uses surrogate end points in about half of new drug approvals. Some biomarkers are accurate predictors of disease risk and can be useful measures of the efficacy of a new drug (such as low-density lipoprotein cholesterol for statins). But though a drug’s effect on a biomarker can make approval quicker and less costly, especially if the comparator is placebo, it may not always predict the drug’s capacity to improve patient outcomes. Bevacizumab (Avastin) delayed tumor progression in advanced breast cancer but was shown not to benefit patients. Similarly, rosiglitazone (Avandia) lowered glycated hemoglobin levels in patients with diabetes even as it increased their risk of myocardial infarction. In 2013, patients began to receive a new drug for tuberculosis approved on the basis of a randomized trial relying on a surrogate measure of bacterial counts in the sputum — even though patients given the drug in that trial had a death rate four times that in the comparison group, mostly from tuberculosis. These provisions in the legislation would not immediately change FDA approval standards, but they would give the agency greater discretion, backed by congressional support, to approve drugs on the basis of less rigorous data.
However, right now the FDA also requires further post-approval studies to verify that the effect on the biomarker or surrogate endpoint actually results in concrete benefit to the patient. Examples include improved survival in cancer or decreased cardiovascular deaths in patients with hypertension.
Does this worry you? It sure worries me. The bill also explicitly encourages using less rigorous evidence. For instance, in the case of antibiotics needed to treat “serious or life-threatening infection” in patients with an “unmet medical need,” the FDA would be allowed to accept “nontraditional” measures of efficacy, including “preclinical, pharmacologic, or pathophysiologic evidence; nonclinical susceptibility and pharmacokinetic data, data from phase 2 clinical trials; and such other confirmatory evidence as the secretary [of health and human services] determines appropriate to approve the drug.” True, such drugs would carry a disclaimer on the label, but does anyone think this really would make sure that these drugs were only used for the “unmet medical need”? Worse, the bill makes further clinical trials to verify that these “nontraditional measures” actually correlate with improved outcomes optional.
As if these provisions loosening the required standard of evidence were not enough, things get really scary in the realm of medical devices by creating an expedited approval pathway. Take a look at what would be considered adequate evidence for the approval of medical devices under this pathway:
(ii) For purposes of clause (i), valid scientific evidence may include—
“(I) evidence described in well-documented case histories, including registry data, that are collected and monitored under a protocol determined to be acceptable by the Secretary;
“(II) studies published in peer-reviewed journals; and
“(III) data collected in countries other than the United States so long as such data otherwise meet the criteria specified in this subparagraph.
It gets worse, though. If the device manufacturer submits studies published in peer-reviewed trials, the FDA would have to submit a request for the underlying data, but even if it couldn’t get that data it would still have to consider the study. Even more incredibly, as noted in the NEJM, another section of the law allows device makers to pay a third-party organization to determine whether the manufacturer can be relied upon to assess the safety and efficacy of changes it makes to its devices instead of submitting an application to the FDA. It’s hard for me to view this provision as anything other than free market faith run amok and/or a payoff to the device industry.
But that’s still not all.
The act also politicizes the process of drug development. Provisions in Sections 2021/2022 in essence create a parallel process by which standards for approval of drugs or devices are developed, as described by a Yale Law School researcher, Greg Gonsalves, in an article by Julia Belluz:
What industry is asking for here are adjudications on the use of key drug development tools very early on, with the guidance of outside experts now taking the central role rather than that of scientists at FDA. These decisions [then become] political ones in tete-a-tetes between [the experts] and the Secretary of Health and Human Services, and commitments to an approval pathway using these drug development tools are agreed upon in an almost binding fashion and are difficult to step away from.
These outside experts will include experts chosen by the companies being regulated.
There’s other mischief in this act as well. For instance, it would weaken informed consent protections on patients by adding another category when it is acceptable to forego informed consent, namely when “the proposed clinical testing poses no more than minimal risk to the human subject and includes appropriate safeguards to protect the rights, safety, and welfare of the human subject.” As Merrill Goozner put it, even if the risk is minimal, why would the authors of this legislation waive a central tenet of international agreements designed to protect the rights of human subjects in clinical trials? I don’t understand, either, particularly since the act doesn’t define “minimal risk” or specify who determines whether a study is minimal risk. The number of studies that would be facilitated by such a change in the requirement for informed consent must be vanishingly small.
Then there’s a provision that sidesteps the Sunshine Act, adding an exemption for reporting medical education payments, which has the effect of increasing the number of things drug companies can pay doctors for without having to report them to the physician payment Sunshine database.
The FDA doesn’t need the 21st Century Cures Act
When I first learned of the 21st Century Cures Act, I wanted to like it. I really did. I especially wanted to support it because it provides more funding to the NIH. Even though the act does put too much faith in precision medicine as the be-all and end-all of progress in biomedical research, at least it boosts the NIH budget and provides more funding for early career investigators, who are being particularly harshly squeezed right now. Indeed, I don’t know why they do it, given how small their chances are of attaining NIH funding, which is usually a necessary prelude to winning tenure. The problem is the rest of the bill, which, while not explicitly ordering the FDA to use a lower standard of evidence to approve drugs and devices, opens the door—and outright encourages—less rigorous science in the name of speed.
Like misguided right-to-try laws, none of these additional provisions are needed, and much of what else is in the bill is likely to be harmful to the scientific rigor of the FDA drug approval process. As author Joseph Gulfo, a critic of how FDA approves drugs who is actually in favor of less regulation and who might be expected to be in favor of a bill like this, argues, the FDA already has the power to do most of what is in the 21st Century Cures Act. Gregg Gonsalves and Mark Harrington, former members of Act Up/New York, and David A. Kessler, commissioner of the FDA from 1990 to 1997, agree that the FDA already has the tools to do what the bill’s sponsors want.
Again, the underlying assumption behind the 21st Century Cures Act is that the system is so broken that radical reform is urgently needed. It turns out that that’s an exaggeration. For instance:
An underlying premise of the bill is the need to accelerate approval for new products, but this process is already quite efficient. A third of new drugs are currently approved on the basis of a single pivotal trial; the median size for all pivotal trials is just 760 patients. More than two thirds of new drugs are approved on the basis of studies lasting 6 months or less — a potential problem for medications designed to be taken for a lifetime. Once the Food and Drug Administration (FDA) starts its review, it approves new medications about as quickly as any regulatory agency in the world, evaluating nearly all new drug applications within 6 to 10 months, an impressive turnaround for such complex assessments.
Indeed, it’s been pointed out that the FDA now approves drugs faster than most European regulatory agencies. Moreover, as examples such as Avastin for breast cancer suggest, under accelerated approval we might be approving some drugs too quickly.
The bill has also been characterized as a “Christmas tree,” something upon which to hang everything on the wishlist of every healthcare-related interest group that’s been frustrated over the last several years:
According to the Senate’s lobbying database, the number of groups listing the 21st Century Cures Act as a lobbying issue grew from 62 in the second quarter of last year to 222 in the first quarter of this year.
“It became a Christmas tree,” said Billy Wynne, of Thorn Run Partners, who has lobbied the bill. “Many people see it as the moonshot opportunity for lots of stuff,” another lobbyist said.
Basically, the 21st Century Cures Act is a solution in search of a problem. Well, not quite. Increasing funding to the NIH is definitely a laudable goal, given how much NIH funding has plummeted in real dollars since 2003. Increasing funding to the FDA is also a laudable goal, although some have argued that the bill doesn’t increase it enough to cover the new responsibilities placed on the FDA. Everything else in the bill is nothing more than a fantasy solution to a nonexistent problem rooted in the belief that if we decrease the standard of evidence for new treatments somehow, magically, lots of “21st Century Cures” will be approved. The bill in its current form is a danger to patients and should not be passed.