Decelerating the hyperbole over accelerated approval

Achieving rational discussion of accelerated approval will require employing robust science to evaluate its value

In the increasing controversy surrounding FDA’s accelerated approval pathway, it’s far too easy to overlook why this pathway exists in the first place — to define a medicine’s value through its clinical utility, while striking a regulatory balance between quality, speed and value.

The key “red thread” behind much of this manufactured controversy is the prioritization of cost effectiveness over clinical benefit.

Recent articles published by the Institute for Clinical and Economic Review (ICER), The Milbank Quarterly and others strive mightily to show the economic inefficiencies of innovative medicines approved by FDA via its accelerated approval pathway by ignoring the (often substantial) clinical benefits of these new medical technologies, using flawed research designs and, most notably, misrepresenting and ignoring data that do not support their cost-over-care mantra.

Robust science must drive the conversation.

Unfortunately, this is not always the case. A recent JAMA commentary exemplified this trend by making some sweeping generalizations based on non-representative value reference points. What’s missing from this and similar arguments comparing drug approvals across regulatory systems are datasets essential to rational conversations and conclusions.

For example, country-specific comparability of multiple variables, concurrent comparators (vital to the assessment of unmet need), representative value, time relativity, and similar boundary conditions unique to each geography and regulatory regime are critical elements missing from the JAMA essay.

Many of these elements are relevant in a broad sense for European and Canadian conditional approval factors, and they were well-captured in the essay (as they have been elsewhere), given that the value measures for approvals were based solely upon Healthcare Technology Assessment reviews in France, Germany, and Canada.

U.S. value assessment data were entirely overlooked, however.

Without such data, available via sources such as the National Comprehensive Cancer Network (NCCN) for oncology indication endorsements, any conclusions about U.S. applicability are woefully thin. It would have been useful to make some attempt at evaluating whether non-U.S. variables have any comparability to U.S. therapeutic value assessments. Equipoise cannot be ignored when it is inconvenient to one’s conclusions.

A broader look at accelerated approvals in the U.S. would have demonstrated whether unmet need was addressed at the time of initial approval, and what measures of relative value U.S. stakeholders would consider persuasive. Such analyses have also been missing or inadequate in other publications.

For a legitimate transnational comparison, it’s surprising that much of the current literature provides: 

(1)   No attention to available comparative treatments in the United States at the time of regulatory approval nor any suggested justification for comparability of non-U.S. HTA sources.

(2)   No attempt to estimate therapeutic utility values in U.S. terms. The assumption that comparative unmet need, criteria for significant therapeutic value, and related critical indicators are similar across borders and regulatory regimes is absurd. This is not hyperbole.

(3)   No attempt to address or even to explain any differences in the datasets used by European and Canadian HTA bodies and comparative labeling of products (or access limitations) in the United States to adjust for significant differences.

Most glaring, in the case of the JAMA article, the non-U.S. HTAs used to determine value were available for only 61% of FDA’s accelerated approvals. By analogy, that’s like a clinical trial analysis excluding almost 40% of  the dataset without reasonable justification, assuming results would be without bias — a situation regulatory reviewers would likely see as invalidating the trial conclusions. The authors’ explanation was that the excluded FDA accelerated approvals were for drugs approved only in the U.S. and therefore had no HTA data from either European or Canadian panels.

There should be no place for “arbitrary censuring” in this area of analysis.

At FDA, accelerated approvals are often based on surrogates such as biomarkers that are subject to a rigorous qualification process to be viewed as reliable. We suggest similar scientific and analytical rigor be applied to tools such as this to assure applicability and representativeness.

While the JAMA article notes the authors’ value methodology had been previously used as a basis for comparisons, using bad data and analytical methods over and over doesn’t make them legitimate. The earth remains round. Vaccines do not cause autism.

The importance of impact on patients’ lives — quantitatively and qualitatively — in terms of unmet medical need offers more compelling metrics for therapeutic value.

Such data, which are readily available from sources such as Friends of Cancer Research’s web postings (including “Data driven insights-Accelerated Approval) deserve more respect and study in creation of meaningful measurements. Artificial and incomparable conclusions that ignore fair balance expose structural flaws in much of the existing literature that seeks to call into question the value of new therapies licensed by the FDA under accelerated approval protocols.

Highly intelligent, dedicated, and experienced researchers and regulators can look at the same datasets and reach different conclusions. That’s science. And the more advanced the science, the greater the nuance and the more difficult the decision. As FDA Commissioner Robert Califf reminds us, “Rules of engagement must be transparent and developed through a process that builds consensus across the relevant ecosystem and its stakeholders.” 

In the post-pandemic “next normal,” a more progressive expression of how data evolve over time is particularly relevant for those innovative medicines approved via regulatory review arteries such as accelerated approval. The FDA will increasingly consider a far broader range of evidence such as real-world evidence, patient-reported outcomes, and quality-of-life measures. And so too must those who are (appropriately) concerned with healthcare technology assessment.

Timothy Franson is a principal at Faegre Drinker Consulting, past president of the United States Pharmacopeial Convention, and immediate past board chair of the Critical Path Institute. Peter Pitts is a former FDA associate commissioner and member of the United States Senior Executive Service; he is president of the Center for Medicine in the Public Interest and a visiting professor at the University of Paris School of Medicine.

 Signed commentaries do not necessarily reflect the views of BioCentury.