Enabling a responsive and agile performance evaluation of AI-based digital diagnostics

Wednesday, 1 March | 15:00–16:15 CET | Online
Organized by AI for Good

Register here

Artificial intelligence (AI)-based technologies have been used successfully to support the goals of SDG3 – Ensure healthy lives and promote well-being for all at all ages – such as the use of AI-based image analysis for augmenting interpretation of X-rays for tuberculosis screening and diagnosis, increasing diagnostic accuracy in evaluation of cough and auscultation of lung sounds, etc. However, the lack of performance data on AI-based diagnostic technologies in low- and middle-income country (LMIC) settings and across a representative dataset contribute to a disorganized digital health technology landscape that may be filled with ineffective or biased AI-based diagnostic solutions. This gap not only allows for the adoption of digital health solutions that may not be suited for scale-up, but also places a time-intensive task on the end-user or individual implementing organizations of reviewing and identifying effective digital health technologies. Generating and disseminating evidence on the performance of digital diagnostic technologies based on standardized and objective assessment criteria using representative datasets can address this gap, catalyze the adoption of evidence-based digital health technologies, and subsequently lead to effective and safe use of digital health technologies among LMIC end users. This session would look at questions on performance evaluation of AI-based diagnostics while ensuring that the policy and regulatory approval timelines and requirements are in step with the pace of evolution of these technologies.

Rigveda Kadam, Deputy Director, Digital Health at FIND will give a presentation on our work at this event.