Abstract
Continuous glucose monitoring (CGM) has transformed diabetes mellitus management, evolving from a supportive monitoring tool to a central pillar of care. For people with type 1 diabetes and many insulin-treated individuals with type 2 diabetes, CGM now directly informs treatment decisions, especially when integrated with automated insulin delivery (AID) systems. In these hybrid closed-loop systems, sensor data drives real-time insulin adjustments, meaning that accuracy is not just a matter of measurement quality; it is a matter of patient safety. However, the primary accuracy measure currently used, the mean absolute relative difference (MARD), is increasingly inadequate for guiding clinical decisions. MARD offers a single averaged number under controlled conditions, but it does not capture the timing, direction, or clinical consequences of sensor errors. This is particularly problematic in AID systems, where even minor inaccuracies may lead to inappropriate insulin dosing, increasing the risk of hypoglycemia or hyperglycemia. Given the centrality of CGM in modern diabetes care, a more comprehensive evaluation approach is urgently needed, one that reflects real-world performance, prioritizes patient safety, and addresses the diverse contexts in which CGM devices are used. This editorial presents an opinion-based perspective, proposing a four-dimensional framework for CGM evaluation beyond the traditional reliance on MARD.