Rethinking Meal Management in Automated Insulin Delivery: From Carbohydrate Counting to Simplified Approaches

To review the spectrum of meal-management strategies currently used with automated insulin delivery (AID) systems, combining published evidence with contemporary clinical experience.

This narrative review examined published approaches to prandial management with AID and incorporated the experience of ten French diabetes specialists from high-volume adult and pediatric AID centers. Strategies were categorized according to the degree of carbohydrate estimation, user involvement, and cognitive burden.

Although accurate carbohydrate counting remains the reference approach, experts described a continuum of increasingly simplified strategies, including semi-quantitative meal announcements, fixed repetitive meal announcements, insulin-driven announcements based on the intended insulin dose, and proxy announcements using fictitious carbohydrate entries primarily to signal meal initiation. Semi-quantitative approaches currently have the strongest evidence base and may provide an effective alternative to precise carbohydrate counting, whereas several other strategies used in routine practice remain largely unevaluated prospectively. Despite their differences, these approaches share the objective of preserving timely prandial insulin delivery and, whenever possible, bolus-calculator use while reducing educational requirements and cognitive burden.

Meal management with AID should be individualized according to patient abilities, preferences, eating patterns, treatment burden, and system characteristics rather than viewed as a binary choice between precise carbohydrate counting and no meal announcement. As optional-announcement hybrid closed-loop and fully closed-loop technologies emerge, the required level of user involvement may progressively decrease, further shifting meal management toward simpler and more autonomous approaches.