Digital health teams should take note: the U.S. Food and Drug Administration has formally withdrawn its guidance titled “Software as a Medical Device (SaMD): Clinical Evaluation,” effective January 6, 2026. Originally issued in December 2017, this document shaped how many organizations approached clinical evidence for standalone medical software. Its removal signals a meaningful shift in how the agency is positioning SaMD oversight going forward.

This article explains what was withdrawn, why it matters, and how software manufacturers and healthcare innovators should respond.

What Was Withdrawn—and Why It Mattered

The withdrawn guidance provided a structured framework for demonstrating the safety and effectiveness of SaMD across three pillars:

• Clinical Association – Establishing a valid link between the software’s output and a clinical condition or concept.
• Analytical Validation – Demonstrating the software processes data accurately and reliably.
• Clinical Validation – Showing that the software achieves its intended clinical purpose in the target population.

The guidance closely aligned with the work of the International Medical Device Regulators Forum (IMDRF) and became a de facto roadmap for regulatory submissions, internal validation strategies, and investor diligence.

With the FDA’s withdrawal, this document no longer represents the agency’s current thinking.

What the Withdrawal Does—and Does Not—Mean

It is important to distinguish policy signals from legal requirements.

What has not changed

• Statutory requirements under the Federal Food, Drug, and Cosmetic Act remain in force.
• Manufacturers are still expected to provide appropriate evidence of safety and effectiveness.
• FDA review pathways (510(k), De Novo, PMA) are unchanged.

What has changed

• The FDA is no longer endorsing a single, prescriptive framework for SaMD clinical evaluation.
• Developers cannot rely on the withdrawn guidance as a safe harbor or citation in submissions.
• Regulatory expectations are becoming more contextual and risk-based.

In practical terms, the FDA is stepping back from a “one-size-fits-all” model for clinical evidence in software.

Why the FDA Likely Made This Move

While the FDA has not issued a detailed rationale, the withdrawal aligns with several broader trends:

1. Rapid evolution of software and AI - Fixed frameworks struggle to keep pace with adaptive algorithms, real-world learning, and iterative deployment models.
2. Shift toward risk-based oversight - Lower-risk digital health tools increasingly fall outside traditional device regulation, while higher-risk tools demand tailored evidence strategies.
3. Emphasis on total product lifecycle approaches - Continuous monitoring, post-market data, and real-world performance are becoming as important as pre-market clinical studies.

Collectively, these trends suggest a regulatory philosophy that favors flexibility over rigid documentation templates.

What to Use Instead: The New SaMD Playbook

With the 2017 guidance withdrawn, organizations should anchor their SaMD strategies around the following:

1. Current FDA Digital Health Guidance

Leverage active guidance documents from the U.S. Food and Drug Administration related to:

• Clinical Decision Support (CDS)
• AI/ML-enabled medical devices
• Cybersecurity and software lifecycle management
• General principles of clinical evidence

2. Risk-Based Clinical Evidence

Align the depth of clinical evaluation with:

• Intended use and claims
• Potential patient harm
• Degree of clinical decision influence
• Availability of mitigations and human oversight

3. Real-World Evidence and Post-Market Controls

Demonstrate maturity by incorporating:

• Post-market surveillance plans
• Performance monitoring metrics
• Update and change management processes

4. Clear Regulatory Narratives

In the absence of a single framework, clarity matters. Submissions should clearly explain:

• Why the chosen evidence is appropriate
• How risks are controlled
• How ongoing safety and effectiveness will be ensured

Strategic Implications for Digital Health Organizations

For startups, this change increases both freedom and responsibility. There is more room to tailor clinical strategies—but less regulatory “cover” from prescriptive guidance.

For established vendors, the withdrawal is an opportunity to modernize evidence approaches and integrate real-world performance data more tightly into compliance programs.

For healthcare organizations deploying SaMD, vendor due diligence should now focus less on checklist compliance and more on evidence quality, governance, and lifecycle controls.

Key Takeaways

• The FDA formally withdrew its SaMD Clinical Evaluation guidance effective January 6, 2026.
• The withdrawal removes a prescriptive framework but does not reduce regulatory expectations.
• SaMD developers must adopt more tailored, risk-based clinical evidence strategies.
• Clear justification, lifecycle governance, and real-world performance data are now critical.
• This change reflects a broader evolution in how digital health technologies are regulated.

As SaMD continues to mature—especially with the rapid integration of AI—regulatory success will increasingly depend on thoughtful strategy rather than rigid adherence to legacy guidance.