The CDSCO has published a 76-page draft guidance for medical device software (including software in medical devices and software as a medical device) intended to align India’s regulatory framework with global standards. The industry welcomes the move but urges greater clarity on evaluation protocols for AI/ML-based software.
Glimpse:
The draft guidance outlines definitions, classification, licensing, applicable standards, quality management systems and clinical evaluation requirements for software used in medical devices and in vitro diagnostics. However, the Medical Technology Association of India (MTaI) has called for more detailed protocols for AI/ML-enabled devices, especially around algorithm change management and proportional requirements for low-risk software.
The CDSCO has released a draft “Guidance Document on Medical Device Software” under the Medical Devices Rules 2017, marking a significant step toward formalizing regulatory oversight of software used in medical technologies. The 76-page draft defines software in a medical device (SiMD) and software as a medical device (SaMD), outlines classification criteria, technical documentation, quality management systems, licensing for manufacture or import, and approaches for performance and clinical evaluation, including for in-vitro diagnostic (IVD) software.
The guidance aligns India’s regulatory framework with global standards, referencing international practices and clarifying submission processes, application portals, and regulatory expectations. This provides med-tech firms, both domestic and international, with a clearer roadmap for developing software products for the Indian market, particularly those leveraging AI/ML, cloud platforms, and advanced analytics.
Industry response, especially from MTaI, welcomes the initiative as foundational but highlights gaps requiring attention. Key concerns include clearer protocols for clinical evaluation, algorithm change-management, post-market behavior of AI/ML systems, and simplified rules for low-risk software to ensure proportional regulatory burden.
The draft is open for a 30-day stakeholder feedback window, after which the final version will aim to balance patient safety with innovation. The industry will closely monitor how the finalized guidance addresses AI/ML-specific concerns, including algorithm updates, transparency, auditability, and real-world performance monitoring for software-based medical tools.
“While the guidance is a welcome step, the final document should provide greater clarity on clinical-evaluation protocols and algorithm change-management, particularly for AI/ML-based medical devices.”
By
HB Team
