The use of external controls in FDA regulatory decision making
A new paper published in Therapeutic Innovation & Regulatory Science reviews FDA decisions between 2000 and 2019 for drug and biologic products to identify pivotal studies that used “external controls.” The authors—which include Berkeley Public Health’s Veronica Miller, PhD, director of the Forum for Collaborative Research; experts from BioMarin Pharmaceuticals; and geneticist Jerry Vockley of the University of Pittsburgh Medical Center Children’s Hospital; among others—say that their review may lead to innovative approaches to clinical trial design; quicker product development for rare diseases; and updates to regulatory guidance.
“In clinical research, we are most familiar with the concept of an ‘internal’ control—a trial design in which a group of patients meeting the inclusion and exclusion criteria is randomized to one of two or more arms in a study; one of these arms could be a placebo arm. It assures us that everyone in the trial is drawn from the same group of potential participants and randomization helps to reduce any additional concerns for bias,” says paper co-author Miller. “An ‘external control’ on the other hand, means that the comparator arm (e.g., patients not receiving the new treatment that is being tested) is drawn from another population, for example, patients from a clinic or a network of clinics that contribute data to a natural history study, or it could be patients from a placebo arm in a previous study—in other words, they are external to the study.”
External controls are not normally considered best practice. “An external control is not the preferred method because 1) the external population could be different from those receiving the treatment in important ways and 2) we can’t rely on randomization to reduce potential biases,” Miller says.
But there are situations in which an external control model is valuable. “Some rare diseases have so few patients available for a study that randomization simply is not possible,” says Miller. “And sometimes, if there is strong evidence that a treatment is likely to work, it might not be ethical to withhold it from people if they are suffering from a rapidly progressing, severely debilitating and/or fatal disease. In such cases, knowledge of what would happen without the treatment can be derived from our years of collective experience in clinical care in the form of an external control.”
Miller hopes that the paper will make both researchers and FDA gatekeepers more comfortable with external controls when they are appropriate–especially for product development for rare diseases.
“As we gain more and more experience with externally controlled trials, especially focusing on the totality of evidence and novel, potentially useful statistical approaches, our hope is that we will increase our comfort level with this approach for situations where the traditional approach simply is not feasible. Most importantly, it will be important to follow these case studies and observe to what extent the new product performs well (or not) in the real-world setting,” says Miller.