A Deep Neural Network Application for Improved Prediction of $\text{HbA}_{\text{1c}}$ in Type 1 Diabetes

HbA
_1c
is a primary marker of long-term average blood glucose, which is an essential measure of successful control in type 1 diabetes. Previous studies have shown that HbA
_1c
estimates can be obtained from 5-12 weeks of daily blood glucose measurements. However, these methods suffer from accuracy limitations when applied to incomplete data with missing periods of measurements. The aim of this article is to overcome these limitations improving the accuracy and robustness of HbA
_1c
prediction from time series of blood glucose. A novel datadriven HbA
_1c
prediction model based on deep learning and convolutional neural networks is presented. The model focuses on the extraction of behavioral patterns from sequences of self-monitored blood glucose readings on various temporal scales. Assuming that subjects who share behavioral patterns have also similar capabilities for diabetes control and resulting HbA
_1c
, it becomes possible to infer the HbA
_1c
of subjects with incomplete data from multiple observations of similar behaviors. Trained and validated on a dataset, containing 1543 real world observation epochs from 759 subjects, the model has achieved the mean absolute error of 4.80 ± 0.62 mmol/mol, median absolute error of 3.81 ± 0.58 mmol/mol and R
²
of 0.71 ± 0.09 on average during the 10 fold cross validation. Automatic behavioral characterization via extraction of sequential features by the proposed convolutional neural network structure has significantly improved the accuracy of HbA
_1c
prediction compared to the existing methods.