Assessing the Accuracy of Continuous Glucose Monitoring Metrics: The Role of Missing Data and Imputation Strategies

Abstract

Aim: This study aims to evaluate the accuracy of continuous glucose monitoring (CGM)-derived metrics, particularly those related to glycemic variability, in the presence of missing data. It systematically examines the effects of different missing data patterns and imputation strategies on both standard glycemic metrics and complex variability metrics. Methods: The analysis modeled and compared the effects of three types of missing data patterns-missing completely at random, segmental, and block-wise gaps-with proportions ranging from 5% to 50% on CGM metrics derived from 14-day profiles of individuals with type 1 and type 2 diabetes. Six imputation strategies were assessed: data removal, linear interpolation, mean imputation, piecewise cubic Hermite interpolation, temporal alignment imputation, and random forest-based imputation. Results: A total of 933 14-day CGM profiles from 468 individuals with diabetes were analyzed. Across all metrics, the coefficient of determination (R²) improved as the proportion of missing data decreased, regardless of the missing data pattern. The impact of missing data on the agreement between imputed and reference metrics varied depending on the missing data pattern. To achieve high accuracy (R² > 0.95) in representing true metrics, at least 70% of the CGM data were required. While no imputation strategy fully compensated for high levels of missing data, simple removal outperformed others in most scenarios. Conclusion: This study examines the impact of missing data and imputation strategies on CGM-derived metrics. The findings suggest that while missing data may have varying effects depending on the metric and imputation method, removing periods without data is a general acceptable approach.

Keywords: CGM; accuracy; continuous glucose monitoring; diabetes; imputation; metrics; missingness; temporal alignment imputation.