In recent years, FDA has sought to increase the patient perspective in its regulatory decision making. One way this is accomplished is through use of patient‑reported outcome (PRO) measures in clinical studies. PROs can assess concepts that are unobservable to clinicians and are only know to the patient (e.g., pain intensity, anxiety, and how bothersome a condition is to a patient). This concept can be particularly important for medical products that treat disease states primarily defined by patient perception/complaint. PROs can be used in a wide range of manners in clinical studies including for example to assess inclusion/exclusion, and/or as primary endpoints, secondary endpoints, and/or exploratory endpoints.
CDRH has, recently, sought to provide additional information regarding use and acceptance of PROs in regulatory decision making, including through publication of PRO case studies and a PRO Compendium outlining instances in which PROs were used in clinical studies in fiscal years 2014 through 2017 (both available here on CDRH’s PRO page). Even with this recent focus on PROs in clinical studies, in our experience, it can still be challenging for sponsors to know whether PRO data may be accepted as a measure of safety and effectiveness.
On August 31, CDRH published a new draft guidance, Principles for Selecting, Developing, Modifying, and Adapting Patient-Reported Outcome Instruments for Use in Medical Device Evaluation, aimed at giving sponsors some additional insight on this topic. Importantly, the guidance notes that “information from well-defined and reliable PRO instruments can provide valuable evidence for benefit-risk assessments and can be used in medical device labeling to communicate the effect of a treatment on patient[s] . . . when the use is consistent with the PRO instrument’s documented and supported measurement properties.”
The draft guidance describes overarching principles to consider when assessing whether a PRO meets these criteria, including recommendations about ensuring the PRO is fit-for-purpose and best practices for ensuring that “relevant, reliable, and sufficiently robust PRO instruments” are used in device clinical studies. In order to demonstrate that a PRO is fit-for-purpose, the draft guidance lays out three factors:
- Is the concept being measured by the PRO instrument meaningful to patients and would a change in the concept of interest be meaningful to patients?
- What role (e.g., primary, key secondary, or exploratory) will the PRO instrument serve in the clinical study protocol and statistical analysis plan?
- Does the evidence support its use in measuring the concept of interest as specified in the clinical study protocol and statistical analysis plan?
These are important points, and while they may seem obvious, it is not always clear whether a PRO instrument will fulfill them to the Center’s satisfaction. The draft guidance does not give any indication as to the level of evidence required for a PRO to be fit-for-purpose. Indeed, in our experience, even when a disease-specific, validated PRO is used as a primary endpoint for its intended purpose, it can be met with resistance during premarket review, particularly when the PRO has not previously been used as a primary effectiveness measure.
The draft guidance notes several times that sponsors should consider discussing novel (in that it has not previously been relied upon in premarket decision making) and modified PROs with the Center in a pre-submission prior to inclusion in a clinical study. In such a pre-submission, sponsors should ensure they address each of the above‑described elements in as much detail as possible to convince the Center that the proposed PRO is fit-for-purpose.
The draft guidance also provides practical considerations for PRO development including ensuring that PRO instruments are written in plain language easily understood by patients, and providing PRO instruments in multiple languages, where appropriate, to capture health outcomes from patients with limited English proficiency. If a PRO instrument is newly developed or modified from an existing instrument, it must be appropriately validated. The draft guidance notes that while a PRO may be validated as part of a sponsor’s clinical study, the results of the PRO from such study cannot then be presented in the sponsor’s labeling. An additional study of the device using the PRO would be necessary. This approach is consistent with the Center’s view that separate development and validation data sets are necessary to ensure reliability of device and the tools used to assess their safety and effectiveness.
Bottom line: the draft guidance provides helpful insight into understanding the Center’s thinking regarding use of PRO in medical device clinical studies. In our view, it would be helpful to elaborate on how much evidence is necessary to establish that a PRO is fit-for-purpose and how the level of evidence changes depending on whether a PRO is being used as a primary endpoint as compared to a secondary or exploratory endpoint. Even without this detail, however, the information in this draft guidance should help sponsors as they craft justifications for use of PRO measures in device clinical studies.