Table 1: Key steps and considerations to advance biometric monitoring devices as clinical outcome assessments

From: Biometric monitoring devices for assessing end points in clinical trials: developing an ecosystem

Key stepsKey considerationsFurther information
Define COI and COU• Each COI must adhere to a series of steps that include: content validation, construct validation, criterion validation, analytical validation and ultimately clinical validation
• To lead to meaningful regulatory endorsement of BMDs as digital drug development tools, clear COU statements must be developed. Each COU is based on the current state of evidence and a clear rationale for its application in the drug development process, or eventually, in health care management
• Categories of biomarker currently detailed by the FDA–NIH include: susceptibility/risk biomarker; diagnostic biomarker; monitoring biomarker; prognostic biomarker; predictive biomarker; pharmacodynamics/response biomarker and safety biomarker. An end point being evaluated for its ability to predict clinical benefit would be a candidate surrogate end point
• As the impact of the clinical decision for the patient grows, the level of evidence needed to support that COU will increase
• Clinical Outcome Assessment Qualification Program:
• BEST resource:
Defining data standards for actionable databases• Data standards can enable prospective collection of data in a standardized format in both clinical trials and observational studies, allow integration of various data sources to quantify the predictive accuracy, utility and reliability of BMDs and expedite submissions and reviews to regulatory authorities
• In 2014, the FDA issued a guidance document on 'Providing Regulatory Submissions In Electronic Format — Standardized Study Data'. Technical specifications associated with this guidance are updated periodically
• Clinical Data Interchange Standards Consortium foundational standards:
• FDA guidance on electronic data submissions:
Assurance of GCP standards• The intended subject for data collection must be verified. Passwords or dual factor authentication are normally used
• A complete audit trail from device to the clinical trial database must be established
• Well-defined procedures are needed to confidently identify artifacts in data
• BMDs often operate in a complex and heterogeneous computer system environment. For example, a watch measuring activity might itself run software and communicate with a smartphone application, a cloud server and a clinical trial database. Software environments may change (including automatic upgrades). Protocols to distinguish changes in BMD measurements due to patient changes versus software environment changes are essential
• The requirements outlined in FDA 21 CFR part 11 are an important consideration for electronic submission of BMD data to the FDA
• Patient privacy must be protected
• FDA GCP guidance:
Build global ecosystems that enable consensus science• Develop a unified lexicon of terminologies
• Bring together patients, caregivers and advocacy organizations, regulators, software developers, academia, pharma, device developers, payers and providers
• BEST resource:
  1. BEST, Biomarkers, EndpointS, and other Tools; BMD, biometric monitoring device; COI, concept of interest; COU, context of use; GCP, good clinical practice.