Will data from wearables and sensors be acceptable in regulatory drug submissions?
Although as an industry we understand how to develop and validate clinical measures, the acceptability of data derived from wearables and other mobile sensors remains an area of some uncertainly and likely acts as a barrier to adoption for some researchers. That's why a cross industry working group under the umbrella of the Drug Information Association's (DIA) Study Endpoints Community developed and recently published a suggested approach to the development of an evidence dossier to support clinical endpoints derived from wearable technology and other mobile sensors.
The working party built from earlier work of the Critical Path Institute's ePRO Consortium and the Clinical Trials Transformation Initiative, and also included members of the ePRO Consortium and the Digital Medicine Society (DiMe).
Taking the structure of the PRO dossier, used to support the use of patient-reported outcome measures in regulatory submissions, the working group developed a suggested content and structure for a dossier to support the use of a sensor device and the associated clinical endpoints derived from its data. Fundamental to this dossier is the ability to support device selection through provision of (i) analytical validity data that establishes that the outcomes data generated by the mobile sensor technology firmware and any associated software, have adequate technical performance characteristics - e.g., accuracy, reliability, precision etc., and, (ii) data supporting the clinical endpoint and its validation and properties.
Recommended dossier structure, from Walton et al., Contemporary Clinical Trials 2020; 91: April 2020, 105962 (reproduced under Creative Commons licence CC BY 4.0).
In common with FDA's developing guidance on patient-focused drug development the work acknowledges the importance of the patient voice in the identification of meaningful aspects of health that should be the basis of the measurement approach. Without first having an understanding of what is truly important to patients, the endpoints derived from sensor data may be less relevant. For example, when measuring activity using a wearable - is total steps per day or time in moderate/vigorous physical activity as important or relevant to less active populations as (for example) the amount of times that a short sustained bout of walking can be maintained? While this is an area well understood in PRO development, there has been less work to date in this area to accompany the development of measures derived from wearables and sensors.
The full article, published as open access in Contemporary Clinical Trials can be found here:
Senior Director, Global Regulatory Affairs at GSK
4 年At present, I think it depends on what clinical parameter is being measured and the rigor that a sponsor will put in place to assure data integrity. Obviously, the capabilities of such tools, clinical diagnostics, and medical devices will continue to evolve.
Head of Clinical Operations Oncology, Hematology, Immunology, Infectiology, Antibiotics, Rare Diseases
4 年Fantastic, I also strongly believe that IoT will support our clinical trials and clinical development of our medicines. Algorithms may control AEs and side effects.
Literally The Man Who Wrote the Book(s) on Patient Recruitment
4 年Great stuff - looks like a valuable addition to the guidelines which should be extremely beneficial for the industry as a whole for the future.
CEO
4 年This is fantastic work Bill & co. and goes to show how one thing can build on another. You'll be happy to know that we are already building on the foundation that you and others have built. With a good foundation, a bit of work on the missing bits and pieces and some "cement", we will soon be able to construct something long lasting for the industry. Stay tuned ;)
Life Sciences Technology Fan | Podcast Host | People Leader
4 年Great post Bill Byrom, nice to have some more guidance on this topic.