Learn how InferTrust™ creates decision records tailored to different SaMD risk classifications, from wellness applications through Class III diagnostic devices.
Software as a Medical Device (SaMD) encompasses a broad spectrum of AI applications, from wellness tools that provide general health information to Class III diagnostic devices that directly inform critical clinical decisions. InferTrust™ FDA Enforcement (Patent Pending) creates decision records appropriate to the risk classification and regulatory requirements of each device type.
FDA classifies SaMD based on the significance of the information provided by the software and the seriousness of the healthcare situation it addresses. Higher-risk classifications require more rigorous documentation and oversight. InferTrust™ supports this graduated approach by allowing manufacturers to configure decision record detail, confidence thresholds, and escalation policies based on their device's risk classification. A Class II diagnostic support tool may require decision records with standard detail, while a Class III diagnostic device requires maximum record granularity with lower confidence thresholds for human escalation.
Every cleared SaMD has defined indications for use that specify the clinical context, patient population, and intended use environment. InferTrust™ decision records capture whether each inference falls within the device's cleared indications. If an AI device receives input data that falls outside its validated parameters (patient age range, imaging modality, clinical setting), the decision record flags this deviation and triggers the appropriate escalation policy. This creates verifiable evidence that the manufacturer is monitoring and enforcing the device's indications for use.
Many SaMD products are deployed across multiple clinical sites, each potentially running different software versions during update rollouts. InferTrust™ maintains per-device and per-site decision chains that track exactly which model version is active at each location. During FDA inspections or adverse event investigations, this allows the manufacturer to identify precisely which sites and which patients were served by a specific algorithm version.