US Army Secures FDA 510(k) Clearance for APPRAISE-HRI Smartphone Machine Learning Hemorrhage Detection SaMD

1. Overview

The United States Army developed APPRAISE-HRI, a software as a medical device (SaMD) that uses a machine learning algorithm operating on a smartphone. The device is designed to triage soldiers experiencing internal (unseen) hemorrhaging in the field, enabling more effective and life-saving interventions.

2. The Challenge

The initial version of the SaMD was developed outside a fully compliant quality system. The machine learning algorithm was trained on an extensive dataset of patient monitoring data recorded at multiple clinical sites. After receiving initial FDA feedback through the Pre-Submission process, it became clear that support was needed to address regulatory strategy, quality system requirements, clinical validation, and cybersecurity testing to meet all premarket and postmarket obligations.

3. Regulatory Strategy and Solutions

A comprehensive regulatory strategy was developed, initially targeting the De Novo classification pathway supported by a clinical validation study. All required software documentation, risk management files, and clinical data were identified and prepared. The FDA provided positive feedback on the proposed strategy, software risk classification, and use of existing clinical databases for validation.

A quality management system focused on design control, risk management, and change control was designed and implemented. A clinical validation study protocol, statistical analysis plan, and final report were authored without requiring new patient recruitment, leveraging large existing hospital databases from trauma patient transports. As the project progressed and new cybersecurity guidance was published, full cybersecurity documentation and testing were incorporated.

4. Execution Phases

4.1 Phase 1: Development and Verification

Software design and development followed a formal lifecycle process. Failure Mode Effects Analysis (FMEA) was conducted at each stage, with design reviews and release of the final software version. Verification testing included unit, system, and integration tests, along with human factors studies. Full cybersecurity testing was performed, including threat modeling, static and dynamic scans, fuzz testing, penetration testing, and implementation of appropriate cyber controls.

4.2 Phase 2: Validation and Submission

Independent validation of the algorithm was conducted, followed by statistical analysis and generation of the clinical study report. All verification and validation documentation, risk management files, and benefit-risk discussions were compiled. The submission was initially prepared as a De Novo request and later converted to a Traditional 510(k) based on FDA feedback comparing it to a similar predicate device.

4.3 Phase 3: FDA Review and Negotiation

An Additional Information request was received focusing on hardware-software compatibility and clinical data inclusion. Through the SIR Q-submission process, responses were prepared and negotiated with the FDA without requiring major software changes or additional clinical studies. Minor software updates during review were managed efficiently under the established design control process, with appropriate verification testing completed ahead of deadlines.

5. The Outcome

Regulatory clearance was granted for APPRAISE-HRI under the 510(k) pathway, including assignment of a brand new product code. The project was completed within the estimated budget through meticulous project management using cloud-based tools and coordinated stakeholder input. The success demonstrates effective integration of regulatory strategy, quality systems, clinical validation, and cybersecurity controls for innovative military medical devices.

6. Key Success Factors

• Early engagement with FDA via Pre-Submission to align on requirements.
• Rapid implementation of a targeted quality management system.
• Use of existing clinical databases for efficient validation.
• Comprehensive cybersecurity testing and controls aligned with current guidance.
• Proactive negotiation during FDA review to minimize additional burdens.

This achievement highlights the importance of structured regulatory planning for advanced software medical devices in high-stakes environments.