P-173: Validation of a visual carbohydrate estimation method for mixed Indian meals in type 1 diabetes
M Dhingra, V R, S RastogiAbstract
Introduction
Precise carbohydrate counting is a cornerstone of Type 1 Diabetes (T1D) management but is difficult to implement in real-world settings, especially with heterogeneous mixed Indian meals. This often leads to estimation errors, reduced adherence, and glycemic variability. Simplified visual estimation methods may offer a practical alternative but require validation.
Objectives
To evaluate the accuracy and clinical impact of a structured visual carbohydrate estimation (guestimation) method in individuals with type 1 diabetes.
Methods
100 individuals with type 1 diabetes were trained in a hand-based, portion-referenced estimation method adapted for Indian meals. Over 4 weeks, estimated carbohydrate values were compared with weighed food records. Continuous glucose monitoring data were analyzed for Time in Range (70–180 mg/dL), postprandial excursions, and glycemic variability (CV%). User-reported ease and adherence were also assessed.
Results
Carbohydrate estimation error reduced from 30.8% at baseline to 17.2% post-training (p<0.001). Time in Range improved from 52.6% to 68.4% (p<0.001), and postprandial excursions decreased by 23%. Glycemic variability showed a modest reduction (CV 34.6% to 31.5%, p<0.05). Participants reported greater confidence, reduced cognitive effort, and improved consistency in insulin dosing, particularly during mixed or unplanned meals
Conclusions
A structured visual approach to carbohydrate estimation can improve both accuracy and glycemic outcomes in type 1 diabetes. This method may offer a practical alternative to precise carbohydrate counting, particularly in settings where meals are complex and real-time calculation is challenging.