Experts discusses need to modernize evidentiary models in rare diseases
Using the US Food and Drug Administration’s (FDA) plausible mechanism framework to allow for greater use of causal inference may help with evidence generation for rare diseases, according to a recent paper published in The Journal of Clinical Investigation.
In traditional clinical trial models, treatments are evaluated across large populations, whereas trials for rare disease medicine “often number in the hundreds, dozens, or even single digits,” Peter Pitts, president of the Center for Medicine in the Public Interest and former FDA associate commissioner, and colleagues wrote in their paper. “In this context, the question is not whether a treatment works on average but whether it works for a particular patient.”
While enrolling more patients may address uncertainty in drug development for more common diseases, this may become less practical as patients increasingly are categorized into disease subtypes. “Patients, physicians, regulators, and developers need to know not merely whether a treatment appears beneficial for a group of people with a shared disease, but for whom it works, for whom it fails, and for whom it may be dangerous,” the authors said.
Pitts and colleagues highlighted the agency’s plausible mechanism framework as “an important institutional response to the precision medicine era by recognizing that biological understanding itself can carry evidentiary value.” However, FDA should also consider causal inference as a “pathway from plausible mechanisms to causal mechanisms.”
Causal inference ties together two concepts: a law of counterfactuals that asks what would have happened had the circumstances around the patient been different—such as whether they would have benefited from no treatment or a different treatment—and a law of conditional independence that determines whether an assumption about cause-and-effect lines up with the data.
The approach could offer “rigorous methods for integrating randomized and real-world evidence, supporting external controls and natural-history comparisons, identifying patient-level treatment effects, and enabling responsible generalization across related diseases and platform technologies,” the authors said.
Pitts and colleagues argued that scientific understanding should take on a larger evidentiary role in rare disease where patient populations are increasingly divided into genetic subtypes.
“The FDA’s Plausible Mechanism Framework permits regulatory confidence to be supported by biological plausibility, mechanistic understanding, target engagement, biomarker response, and natural-history evidence even before traditional demonstrations of clinical benefit are complete. This reflects an increasingly important reality; biological understanding itself has evidentiary value,” they explained. “The Plausible Mechanism Framework is important not because it creates a pathway around conventional evidence, but because it recognizes that the nature of evidence itself is changing.”
The authors said that adoption of causal inference would not be a lowering of evidentiary standards, and the method would “force investigators to confront questions that traditional analyses leave implicit” by analyzing assumptions, alternative explanations, uncertainty, and creating counterfactuals.
FDA modernization
“In its adoption of the Plausible Mechanism Framework, the FDA will increasingly evaluate therapies based on shared biological mechanisms rather than treating every disease as an entirely separate evidentiary universe,” Pitts and colleagues said. “The central question is no longer whether evidence can be generated for a single therapy. It is whether knowledge generated for one therapy can responsibly inform the next.”
Use of causal inference would allow for better regulatory decision-making because it is a tool that can determine whether observations and patterns identified in real-world evidence are informative, the authors argued. Causal inference is supported with steps in the drug development process, including “demonstrating a therapeutic agent’s target engagement, establishing persuasive natural-history comparators, validating biologically relevant biomarkers, and constructing a rigorous narrative,” they explained.
“Beyond a compelling use case, these steps each reduce uncertainty,” Pitts and colleagues said. “The more predictable the FDA becomes in evaluating causal evidence, external controls, platform approaches, and real-world evidence, the lower that risk premium becomes.”
The upcoming Prescription Drug User Fee Act (PDUFA) negotiations may also offer an opportunity to fund an investment into using causal inference as a method for evaluating therapies with small patient populations, they noted.
“If the FDA is to evaluate causal evidence consistently and credibly, it must develop the knowledge and regulatory culture necessary to assess evolving forms of evidence with the same rigor historically applied to traditional clinical trials,” the authors said.
