Background and Aims

Cybersecurity in eHealthcare is a growing concern, and in particular, in diabetes care, where vast amounts of data are being generated by the recent surge in wearable devices, such as continuous glucose monitoring (CGM). The use of open-source platforms, for example, NightScout and OpenAPS, has increased the risk of data breaches and might represent a privacy concern for some users. In this work, we aim to demonstrate that it is possible to identify CGM data at an individual level by standard machine learning techniques.

Methods

The publically available REPLACE-BG dataset containing 226 adult participants with type 1 diabetes wearing CGM over 6 months was used. A support vector machine (SVM) binary classifier aiming to determine if a CGM data stream belongs to an individual was trained and tested for each subject in the dataset. Eleven standard glycaemic metrics were employed to generate the feature vector for the SVM. Data points in the training and testing datasets were generated by evaluating the selected glycaemic metrics over multiple incidences of one-month time windows. In order to increase the number of data points, a sliding time window was employed.

Results

The mean and standard deviation of sensitivity, specificity, and accuracy results on the testing data set were 0.80±0.24, 0.97±0.03, and 0.89±0.12, respectively.

Conclusions

This work demonstrates that it is possible to determine with relatively high accuracy if a CGM data stream belongs to an individual. The proposed approach can be used as a digital CGM ‘fingerprint’ or for detecting glycaemic changes within an individual (e.g. illness).

Abstract Body

The Bio-inspired Artificial Pancreas (BiAP) is an advanced hybrid closed-loop insulin delivery system based on mathematical modelling of the pancreatic beta-cell physiology. BiAP incorporates an innovative adaptive meal-insulin bolus calculator which uses artificial intelligence to provide adaptive and individualised mealtime insulin dosing by learning from past post-prandial glycaemic outcomes, user behaviour, and controller’s functioning.

The BiAP control algorithm is designed for embedded low-power solutions. It has been implemented in a dedicated microchip-based handheld device and, more recently, in an iPhone connected to a Dexcom G6 continuous glucose sensor, a Tandem t:slim AP insulin pump, and a dedicated remote web-based platform. BiAP has been successfully assessed in-clinic and an ambulatory crossover randomised controlled trial is planned to evaluate its longer-term clinical effectiveness.

In this talk, we review the latest algorithmic and software developments within BiAP, as well as the results of a realistic in-silico head-to-head comparison of the BiAP controller and an open-source do-it-yourself (DIY) artificial pancreas controller. Finally, the latest developments on enhancing BiAP with a machine learning-based meal detection algorithm will be introduced.