Written and maintained by CASRAI Editorial Board
Last updated
The FDA Digital Health Center of Excellence (DHCoE) is a centralized regulatory policy, scientific research, and advisory hub established in September 2020 within the U.S. Food and Drug Administration’s Center for Devices and Radiological Health (CDRH). Created to coordinate agency-wide oversight of rapidly evolving digital health technologies, the DHCoE develops regulatory frameworks for Software as a Medical Device (SaMD), Software in a Medical Device (SiMD), Artificial Intelligence and Machine Learning (AI/ML)-enabled device software functions, wearable biosensors, prescription digital therapeutics, and medical device cybersecurity.
What Is the Role of the FDA Digital Health Center of Excellence (DHCoE)?
Unlike traditional CDRH Offices of Health Technology (OHT 1 through OHT 8), which are organized by anatomical specialty (such as cardiovascular, orthopedic, or ophthalmic devices) and hold direct statutory sign-off over individual 510(k), De Novo, and Premarket Approval (PMA) submissions, the Digital Health Center of Excellence operates as a cross-cutting center of expertise. Its institutional mandate spans four core operational functions:
- Harmonized Digital Health Regulatory Policy: Drafting and maintaining FDA guidance documents governing FDA artificial intelligence regulation, Clinical Decision Support (CDS) software exclusions under Section 520(o) of the Federal Food, Drug, and Cosmetic Act (FD&C Act), Medical Device Data Systems (MDDS), off-the-shelf software, and premarket software documentation levels.
- Specialty Consultations to CDRH Review Divisions: Providing specialized software engineering, machine learning, human factors, and cybersecurity reviewers to assist anatomical OHT review teams evaluating complex digital submissions.
- International Regulatory Harmonization: Leading FDA collaboration with the International Medical Device Regulators Forum (IMDRF), Health Canada, and the UK Medicines and Healthcare products Regulatory Agency (MHRA) to establish joint standards such as the 10 guiding principles for Good Machine Learning Practice (GMLP) and Transparency for Machine Learning-Enabled Medical Devices.
- Digital Health Focal Point for Developers and Trial Sponsors: Operating the Digital Health Inbox (
[email protected]) to answer regulatory jurisdiction inquiries—helping sponsors determine whether a software tool or digital biomarker in clinical trials is regulated as a medical device, subject to enforcement discretion, or statistically excluded from device status under the 21st Century Cures Act.
What Were the Findings of the FDA Software Precertification (Pre-Cert) Pilot Program and Excellence Appraisal Model?
Prior to the formal launch of the DHCoE, the FDA announced the Digital Health Software Precertification (Pre-Cert) Pilot Program in July 2017 as part of its Digital Health Innovation Action Plan. Recognizing that traditional hardware-centric premarket review timelines were poorly matched to agile software release cycles, the FDA selected nine industry participants (Apple, Fitbit, Johnson & Johnson, Pear Therapeutics, Phosphorus, Roche, Samsung, Tidepool, and Verily) to test a Total Product Lifecycle (TPLC) regulatory model described in the agency’s 2019 Working Model v1.0.
| Pre-Cert Working Model Pillar | Intended Regulatory Mechanism | September 2022 Final Report Finding & Statutory Outcome |
|---|---|---|
| 1. Excellence Appraisal | Evaluate the software developer’s organizational culture across 5 Excellence Principles: Patient Safety, Product Quality, Clinical Responsibility, Cybersecurity Responsibility, and Proactive Culture | Organizational appraisals alone could not reliably substitute for device-specific safety and effectiveness review without creating disproportionate burden for smaller startups |
| 2. Review Determination | Use the IMDRF SaMD risk categorization matrix (State of healthcare situation × Significance of information) plus precertification status to allow lower-risk SaMD to bypass premarket review | Statutory Barrier Identified: Under existing FD&C Act authorities, the FDA cannot legally exempt 510(k) or De Novo devices from premarket submission solely on the basis of a manufacturer’s organizational appraisal |
| 3. Streamlined Premarket Review | Replace redundant static document uploads with interactive review of standardized software architecture, verification, and validation elements | Transitioned into the FDA’s 2023 final guidance on Content of Premarket Submissions for Device Software Functions (Basic vs. Enhanced Documentation Levels) and the eSTAR electronic submission template |
| 4. Real-World Performance (RWP) | Continuously monitor post-market user experience, software defects, and clinical performance metrics in live deployment | Directly informed post-market surveillance frameworks for AI model drift in healthcare and Predetermined Change Control Plans (PCCPs) |
In September 2022, the DHCoE formally concluded the five-year pilot and published its capstone report: “The Software Precertification (Pre-Cert) Pilot Program: Tailored Total Product Lifecycle Approaches and Key Findings.” The FDA concluded that while the pilot demonstrated the value of standardized lifecycle artifacts and real-world performance tracking, a company-level precertification pathway could not be implemented under current law because the FD&C Act ties marketing authorization to the safety and effectiveness of a specific device rather than the general quality culture of its manufacturer.
How Does the DHCoE Support Regulatory Policy for SaMD, AI/ML, and Real-World Performance Monitoring?
Following the closure of the Pre-Cert Pilot in September 2022, the DHCoE channeled its lifecycle findings into device-specific regulatory mechanisms backed by explicit congressional authority:
- Predetermined Change Control Plans (PCCPs) Under FDORA Section 515C: In December 2022, Congress enacted Section 3308 of the Food and Drug Omnibus Reform Act (FDORA), adding Section 515C to the FD&C Act. Implemented through DHCoE guidance, a PCCP allows a manufacturer to specify planned iterative software or AI/ML retraining modifications ( the Description of Modifications), the validation methodology (the Modification Protocol), and the benefit-risk assessment (the Impact Assessment) inside the initial 510(k), De Novo, or PMA submission. Once authorized by the FDA, the manufacturer can deploy those pre-specified updates in the field without filing a new 510(k) for each iteration.
- Basic vs. Enhanced Software Documentation Framework: In June 2023, the DHCoE replaced the legacy 2005 “Level of Concern” (Minor, Moderate, Major) software guidance with a risk-based two-tier framework (Basic Documentation vs. Enhanced Documentation) aligned with IEC 62304 software lifecycle processes.
- Digital Health Advisory Committee (DHAC): Established by the FDA in October 2023 and supported by the DHCoE, the Digital Health Advisory Committee convenes external clinical, bioethics, and machine learning experts to advise the agency on generative AI, large language models in clinical workflows, post-market real-world performance monitoring, and wearables.
Frequently Asked Questions
What is the role of the FDA Digital Health Center of Excellence (DHCoE)?
Established in September 2020 within CDRH, the FDA Digital Health Center of Excellence develops regulatory policy and guidance for Software as a Medical Device (SaMD), AI/ML-enabled medical devices, wearables, and medical device cybersecurity, while providing technical review consultations across FDA device offices and answering developer jurisdiction inquiries.
What were the findings of the FDA Software Precertification (Pre-Cert) Pilot Program and Excellence Appraisal model?
In its September 2022 final report concluding the 2017–2022 Pre-Cert Pilot, the FDA found that while Total Product Lifecycle (TPLC) and real-world performance metrics were highly valuable, the agency lacked statutory authority under the FD&C Act to waive premarket device review based on a company-level Excellence Appraisal, leading the FDA to retire the Pre-Cert Pilot in favor of device-specific lifecycle tools.
How does the DHCoE support regulatory policy for SaMD, AI/ML, and real-world performance monitoring?
The DHCoE operationalizes modern SaMD and AI/ML oversight through Good Machine Learning Practice (GMLP) principles, the 2023 Basic/Enhanced software premarket documentation guidance, the Digital Health Advisory Committee, and Predetermined Change Control Plans (PCCPs) authorized under Section 515C of the FD&C Act.
