392. Analysis of GDm-health data: New insights into gestational diabetes management

• Published in: Pregnancy hypertension 2018-10, Vol.13, p.S47-S48
• Main Authors: Mackillop, Lucy dr, Velardo, Carmelo dr, Hirst, Jane dr, Gallen, Ian dr, Farah, George dr, Tarassenko, Lionel prof
• Format: Article
• Language: English
• ISSN: 2210-7789; 2210-7797
• DOI: 10.1016/j.preghy.2018.08.141

Little evidence guides the optimal monitoring and management of blood glucose (BG) in women with gestational diabetes mellitus (GDM). Guidelines are based on population-based thresholds and care is not tailored to the needs of the mother and her baby. We aim to explore the utility of data analytics for identifying novel outcome predictors in women with GDM based on trends in capillary BG, demographic, clinical and pharmacological data, as captured from a digital BG management system (GDm-health). We developed a smartphone-assisted BG monitoring system which aims to improve the care of women with GDM. We describe the initial analysis of data from women receiving their care using GDm-health at two NHS Trusts between 2014 and 2018. Patient demographic and contextual data including age, booking weight, height, screening criteria for GDM and results, medication use and timing of initiation, qualitative data on diet and lifestyle, HbA1c and maternal and neonatal outcomes will be analysed and corelated with BG data. Of 1661 women with diabetes in pregnancy, 1446 had GDM. Analysis was performed on data from 876 women with GDM who had completed a pregnancy. We collected 112,997 readings; (mean 203, SD 181). 36,164 readings were tagged to breakfast, 31,115 to lunch, and 40,316 to evening meal. 5402 were untagged. 267 of women required pharmacological treatment during their pregnancy with detailed information of dose and timing treatment. Using large data sets to link BG data with demographic data, therapeutic data, and data on maternal and neonatal outcome will enable new insights to refine targets for BG management during pregnancy; automate decision making around medication commencement and dose adjustment; identify those at most at risk of post-partum hyperglycaemia and identify potential digital biomarkers as predictors of maternal and neonatal outcomes.