Many health care stakeholders are embracing digital health technologies (DHTs), including those enabled by artificial intelligence/machine learning (AI/ML), to transform the way health care is delivered in patients’ homes. However, the full potential of DHTs for the detection of prediabetes and undiagnosed diabetes, especially in diverse populations, particularly racial and ethnic minorities, has yet to be realized.
The U.S. Food and Drug Administration’s (FDA’s) Center for Devices and Radiological Health (CDRH) is uniquely positioned to advance the development of high-quality, safe, and effective medical devices, including DHTs, which can provide significant improvements in health care, quality of life, and wellness to diverse populations. CDRH is seeking to gather more information about how DHTs including AI/ML may help with early detection of risk factors for type 2 diabetes, prediabetes, and type 2 undiagnosed diabetes. This effort is consistent with CDRH’s 2022-2025 Strategic Priority focused on advancing health equity.
Request for Public Comment
In particular, CDRH seeks comment from the public on the following questions:
Community Engagement and Consortia Efforts
- Which patient and community groups, health systems, pre-competitive consortia, or other organizations have prediabetes prevention, detection, or management programs?
- Are there any existing connected cohorts for people living with prediabetes? Connected cohorts are defined as groups of people with a common characteristic engaged longitudinally with digital technologies such as smartphones, apps, consumer wearables, or connected sensors for the purposes of health improvement or research. If so, how many people are members of the cohort and how are they “connected?”
- What existing research consortia are developing risk prediction models for pre-disease states including prediabetes?
Science/Innovation
- What DHTs, including those enabled by AI/ML algorithms, are currently being used outside the clinic to prevent, detect, treat, or reverse prediabetes?
- How are DHTs typically being used to capture various signals amongst people who have risk factors for type 2 diabetes, prediabetes, and undiagnosed type 2 diabetes, and what methods are being used to assess such signals as digitally derived measures of biomarkers?
- Who are key subpopulations of interest that might benefit the most from remote screening and diagnostic tools? Please include clinical and non-clinical considerations.
- What are high-prevalence and high-impact risk factors for prediabetes and undiagnosed type 2 diabetes that are or could be captured by DHTs?
- Are there any existing tools, datasets, or devices used for prediabetes detection? Are there any that are particularly indicated for individuals of a particular race, ethnicity, gender, language, and/or comorbid disability?
- Are there applications of using AI/ML on existing health care datasets (e.g., ECG, radiology datasets) that can be used for the detection of prediabetes and/or transition to type 2 diabetes?
Outcomes
- What patient and community-centered outcomes could be measured for prediabetes and undiagnosed type 2 diabetes (e.g., symptoms, activities of daily living, co-morbidities, hospitalizations, costs, psychosocial factors, etc.)?
- How can digitally derived measures of biomarkers be leveraged to improve the engagement, adherence, and self-management of people diagnosed with or at risk for prediabetes?
Clinical Integration and Implementation
- How can digitally derived measures of biomarkers be integrated into clinical decision support systems or electronic health records to identify undiagnosed type 2 diabetes or prediabetes? What are some of the barriers and proposed solutions to implementation?
- How can EHR data and digitally derived measures be used to understand which patients with prediabetes reverse to a healthy state or transition to type 2 diabetes?
- Are there effective ongoing early screening and prevention efforts for prediabetes and undiagnosed type 2 diabetes that utilize DHTs? If yes, please describe. If no, please describe what a solution would look like to enable such an effort.
Please submit all public comments to the docket (FDA-2023-N-4853), available at Regulations.gov. The public comment period will end on Jan. 31, 2024.
FDA has more information.
