Wrist-worn wearables are used for fitness and health monitoring. This global expansion of wearable technology opens up opportunities for the diagnosis and management of chronic conditions. Diabetic patients have a two to three-fold higher risk of developing cardiovascular disease and that cardiovascular diseases accounted for 29.2.% of all deaths in Singapore. The wearable device is a promising avenue that allows for continuous monitoring of the large population of patients. Its ubiquitous and easy to use nature is an added advantage for its implementation. In this study, the investigators aim to leverage existing photoplethysmography (PPG) technology, together with artificial intelligence, to accurately monitor blood glucose levels in a continuous and non-invasive manner. A simple non-invasive tool to monitor blood glucose will be developed, and alerts will be issued when the blood glucose levels fall in the unhealthy range. A standard glucometer will be used to calibrate and validate PPG measurements of blood glucose. This study aims to recruit 500 participants from KK Women's and Children's Hospital.
Wrist-worn wearables are currently being used for fitness and health monitoring. The global expansion of wearable technology combined with smartphones access creates new questions and opportunities in the diagnosis and management of chronic conditions. High-end consumer wearables have integrated green light reflection photoplethysmography (PPG) sensors into their products. A PPG is an optically obtained plethysmogram that can be used to detect blood volume changes within mascrovasculature. Smartphones, smartwatches and heart rate tracking devices are the most commonly used devices to feature PPG. In a published literature assessing smartphone apps using PPG for heart rate monitoring, it has found that these devices are reasonably accurate, with correlation coefficients \> 0.93 and mean absolute percentage errors ranging from 3.3% to 6.2%. Although PPG sensors were initially only designed to track heart rate, there has been a push to use these with algorithms in the detection of arrhythmias such as AF, and other fields. Diabetic patients have a two to three-fold higher risk of developing cardiovascular disease, and that cardiovascular diseases accounted for 29.2% of all deaths in Singapore. Wearable devices are ubiquitous, easy to use, and may allow for screening and further monitoring of a large population of patients. This research proposes to leverage PPG technology, together with artificial intelligence, and incorporate this into affordable wearable lifestyle devices, wrist-worn and in-ear, to accurately monitor continuously and non-invasively glucose levels in humans. A simple non-invasive tool to monitor blood glucose will be developed, and alerts will be issued when the blood glucose level falls in the unhealthy range. The standard glucometer will be used to calibrate and validate the PPG measurements of blood glucose. This study targets to recruit and measure the blood glucose of 500 participants from KK Women's and Children's Hospital. The primary aims of this study are: (1) Calibrate and validate PPG measurements of blood glucose, obtained both the wrist-worn and in-ear PPG devices, against the standard glucometer; (2) To develop a risk prediction model to identify subjects with blood glucose in the unhealthy range, using both subject characteristics and important features extracted from the PPG measurements using machine learning techniques. The secondary aim of this study is to validate that in-ear and wrist-worn wearables both provide relative accurate heart rate and heart rate interval measurements.
Study Type
INTERVENTIONAL
Allocation
NA
Purpose
OTHER
Masking
NONE
Enrollment
500
Wrist-worn wearable
In-Ear wearable
KK Women's and Children's Hospital
Singapore, Singapore
Capillary blood glucose level using the glucometer and PPG signal readings
Raw PPG signal readings obtained from the devices will be collected via an application on a phone. Capillary glucose reading will be obtained using a glucometer at the same sitting.
Time frame: Both the PPG signals and capillary blood glucose will be collected at the same time during the visit lasting up to 2 hour per subject. A total of 500 subjects will be recruited up to 6 months
Risk prediction model of diabetes using PPG features on the wearables
This study aims to develop means for non-invasive measurement of blood glucose and to assist with the early detection of diabetes among subjects who do not exhibit any noticeable signs of the disease. PPG data will be collected from the wearables, and the standard glucometer will be used to calibrate and validate these PPG measurements of blood glucose. The objective is to calibrate and evaluate PPG as a means for assessing and monitoring blood glucose levels. The effects of each of the PPG variables collected will be assessed by its ability to detect an increased blood glucose level, in terms of sensitivity, specificity, and positive and negative predictive values. A risk prediction model will be developed to identify subjects with blood glucose in the unhealthy range, using both subject characteristics and important features extracted from the PPG measurements using artificial intelligence and machine learning techniques.
Time frame: The study will be completed within 6 months involving 500 subjects. Data analysis and risk prediction model will be completed up to 12 months
Validate that in-ear and wrist-worn wearables both provide relative accurate heart rate and heart rate interval measurements.
Smartwatches in the market have been shown to provide relatively accurate data as to the measurement of heart rate and this feature has been incorporated into newly developed smartwatches. In this study, PPG data collected will be used to assess by its ability to detect heart rate and heart rate intervals, in terms of sensitivity, specificity, and positive and negative predictive values.
Time frame: Subjects will be assessed at one visit study. Total duration of the study is 15mins for subjects whose capillary blood glucose is greater than or equal to 11.1mmol/L, and about 90minutes for those less than 11.1mmol/L.
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.