Digital Therapeutics and Software as a Medical Device (SaMD)

Abstract

Software as a Medical Device (SaMD) encompasses standalone software intended for medical purposes without being part of a hardware medical device, ranging from AI-powered diagnostic imaging algorithms and clinical decision support systems to prescription digital therapeutics (PDTs) delivering cognitive behavioural therapy through smartphone applications. The regulatory landscape for SaMD has evolved rapidly, with the IMDRF's risk-based SaMD framework, the FDA's Digital Health Software Precertification Programme, Germany's DiGA fast-track pathway, and the EU MDR's expanded scope for device software creating a complex multi-jurisdictional environment. This study presents the SaMD Regulatory and Clinical Assessment Framework (SRCAF), evaluating five SaMD deployment models -- standalone diagnostic algorithms, clinical decision support with practitioner override, prescription digital therapeutics, AI-driven continuous monitoring, and adaptive algorithm SaMD with post-market learning -- across four assessment dimensions: clinical evidence generation, regulatory pathway navigation, software lifecycle management, and real-world performance monitoring. Our SaMD Performance and Compliance Score (SPCS) integrates clinical validity strength, regulatory pathway efficiency, software quality assurance maturity, algorithm performance stability, and post-market surveillance comprehensiveness. Adaptive algorithm SaMD achieved the highest SPCS (0.918) through predetermined change control plans that enable continuous model improvement within regulatory guardrails, while prescription digital therapeutics achieved the highest clinical evidence strength (0.955) through randomised controlled trial programmes equivalent to pharmaceutical Phase III standards.