Digital Health and Machine Learning Technologies for Blood Glucose Monitoring and Management of Gestational Diabetes

Abstract

Innovations in digital health and machine learning are changing the path of clinical health and care. People from different geographical locations and cultural backgrounds can benefit from the mobility of wearable devices and smartphones to monitor their health ubiquitously. This paper focuses on reviewing the digital health and machine learning technologies used in gestational diabetes - a subtype of diabetes that occurs during pregnancy. This paper reviews sensor technologies used in blood glucose monitoring devices, digital health innovations and machine learning models for gestational diabetes monitoring and management, in clinical and commercial settings, and discusses future directions. Despite one in six mothers having gestational diabetes, digital health applications were underdeveloped, especially the techniques that can be deployed in clinical practice. There is an urgent need to (1) develop clinically interpretable machine learning methods for patients with gestational diabetes, assisting health professionals with treatment, monitoring, and risk stratification before, during and after their pregnancies; (2) adapt and develop clinically-proven devices for patient self-management of health and well-being at home settings ("virtual ward" and virtual consultation), thereby improving clinical outcomes by facilitating timely intervention; and (3) ensure innovations are affordable and sustainable for all women with different socioeconomic backgrounds and clinical resources.

Fig. 1. Digital health for antenatal and postnatal health and care in hospital, community and home care environments

Fig. 2. Self-monitoring glucose meters

(a) fingertip blood glucose testing with mobile connections [20, 21], (b) four generations of blood glucose meters (c. 1987-2005): Top left: Reflolux S (Accu-Chek III in the U.S.), by Boehringer Mannheim, 2-minute read time, based on reflectance; top right: ExacTech Card, by MediSense, 30-second read time, electrochemical test stripe; bottom left: FreeStyle, by TheraSense, 15-second read time, electrochemical test stripe; bottom right: Freestyle Mini, by Abbott, 7-second test time, electrochemical test stripe. [22].

Fig. 3. Noninvasive enzyme-based glucose monitoring sensing systems through different contact agents and body sensors

(a) contact lens glucose sensor on tear, (b) saliva glucose monitoring strip on saliva, (3) needle-type glucose sensor on insulin sensitivity factor (ISF), and (d) wearable glucose monitoring patch on sweat. [33]

Fig. 4. Noninvasive continuous glucose sensing techniques

(a) skin-like glucose biosensor [41], (b) wearable-band type near infrared (NIR) optical biosensor [42], (c) sensing through fluorescent labelling [38], and (d) microwave sensors [40].

Fig. 5. Taxonomy of models for blood glucose prediction.

Fig. 6. An example of a hybrid approach using the physiological model and data-driven model [76]

Fig. 7. The architecture of GluNet [79].

Fig. 8. LSTM time-series prediction model with deep residual network [85]

Fig. 9. The block diagram of the proposed DRF model with the actor-critic architecture, reproduced without changes from [87]