The Evidence For Endpoints: PROs And Beyond
By Jean Paty
On October 19, 2011 the FDA hosted a workshop that will change how developers of medical products select, evaluate, and provide evidence to support endpoints for their clinical development programs. Dr. Janet Woodcock, director, Center for Drug Evaluation and Research (CDER), started the day with clear statements regarding the instruments used to measure the efficacy, and sometimes safety, of new medical products.
Woodcock cited FDA regulation 21CFR314.126(b)(6) requirements: “The methods of assessment of subjects’ response are well-defined and reliable. The protocol for the study and the report of results should explain the variables measured, the methods of observation, and the criteria used to assess response.”
What Does This Mean?
In December 2009, the FDA released the final guidance on how patient reported outcomes (PROs) will be evaluated to meet the standard of “well-defined and reliable.” Woodcock made it clear to those in the audience that the requirements for developing evidence to support PROs as outcomes equally apply to other outcomes, including: clinician reported outcomes (ClinROs), which are those collected by a clinician evaluating a patient and recording the results; and other observer reported outcomes (ObsROs), such as parents observing and rating children. In other words, the same “yardstick” for evaluating evidence will be used for all outcomes. Taken together, the FDA is now referring to all types of outcomes in trials as clinical outcomes assessments (COAs).
Why Is The FDA Taking This Position?
The release of the PRO draft guidance in 2006 and the final guidance in 2009 articulated scientifically appropriate standards for PROs and for other outcomes. The workshop was the FDA’s signaling to sponsors that the basic evidence needed to evaluate one type of endpoint applies to all others. FDA staff made a clear distinction between direct measures of patient benefit and indirect measures, which include various types of ClinROs, ObsROs, and biomarkers. The workshop participants also discussed how the appropriate endpoint for a given trial varies depending on context of use. For example, in measuring pain in a mentally competent adult, a direct measure of pain obtained from the patient using a PRO instrument may be the most appropriate. However, in mentally impaired adults or small children who cannot self-report, a ClinRO or ObsRO reporting patient behaviors that reflect pain may be appropriate. Sponsors would have to provide the evidence that shows the indirect measures reflect the direct concepts. The FDA, and specifically Woodcock, charged the workshop leaders to not discuss surrogate endpoints that day, and instead strongly encouraged discussion of the use of other forms of evaluation of efficacy of products. However, these endpoints must be compelling and interpretable — that is, they must be well-defined and reliable, as required by 21CFR314.126(b)(6). As Woodcock commented, the outcomes must have an evidentiary basis to support that they are appropriate for their context of use.
How Does This Affect My Clinical Trial Planning?
Following Woodcock’s initial address, the rest of the workshop day was spent discussing what this meant for clinical trial planning. What emerged was that the implication of the FDA’s position is that any ClinRO, ObsRO, or PRO implemented in a clinical program with the goal of supporting the product labeling must have evidence to support that it is fit for purpose (previously referred to as “validated”) for the context of use in the specific patient population. This kind of evidence includes content validity (demonstrating that an instrument measures what we think it measures), reliability (or showing the instrument will give consistent results if the patient is not changing), sensitivity to change (if treatment works, we can detect it), and that the results from an instrument are interpretable (or, being able to assert that a change is clinically meaningful). These criteria, which are standard in the world of instrument development and evaluation, will need to be met and reported in evidence dossiers for COAs. Such dossiers are separate and distinct documents from those on the efficacy and safety of the product.
Jean Paty, Ph.D., is cofounder and senior VP of scientific, quality, and regulatory affairs for invivodata, inc., and chief scientist and regulatory advisor of PRO Consulting.