Diabetes in pregnancy is characterised by high levels of glucose in the blood (hyperglycemia) which may develop during pregnancy (a condition known as gestational diabetes mellitus, or GDM) or in some cases, may already be present in the form of pre-existing type 1 or type 2 diabetes. GDM is increasing in prevalence worldwide, driven by demographic and lifestyle changes. In the UK, prevalence of GDM is currently estimated to be 20 .
Approximately half of women with a history of GDM will go on to develop type 2 diabetes within five to ten years after delivery. Children born to women with GDM are also at an increased risk of developing Type 2 diabetes later in life .
If not managed effectively, hyperglycemia in pregnancy can cause serious maternal and neonatal complications. The risk is greatly reduced through tight blood glucose control during pregnancy.
GDm-Health™ allows women to log blood glucose measurements on a smartphone app which can be connected to a blood glucose meter, and remotely share these results with their healthcare provider.
Additional information can be attached to blood glucose readings including comments, such as meal information, medication taken, and call-back requests. Clinical care staff can review and monitor blood glucose results and metadata in near real-time and intervene if needed. Nearly a third of acute Trusts across England now use GDm-Health, where it has replaced the traditional method of women recording readings in a paper diary to be reviewed in clinic every one to two weeks.
Since it’s commercial release by Sensyne Health, over 12,000 women have actively used the app to record their blood glucose readings up to six times a day, as well as other relevant data and demographic and clinical information (e.g. age, ethnicity and maternal and neonatal outcomes).
Introducing the data
The GDm-Health dataset now represents the largest comprehensive GDM dataset in the world. It is a live system that is increasing in size as more NHS Trusts adopt the system, and more women start using the system. Compliance rates are higher than standard care  with one Trust reporting a 98% compliance level amongst women who are prescribed GDm-Health .
- Data from 34 trusts are gathered in separate, isolated cloud instances
- State-of-the-art privacy preserving tools are used to anonymize data (ICO guidelines)
- De-identified and anonymised data are stored on a secure cloud storage accessible from specific locations and to named individuals performing the analysis (virtual cold-room)
- Analysis is performed on the de-identified and anonymised dataset
Broad, connected data
In addition to the breadth of the dataset, the database contains many connected datapoints. BG readings are linked to medications, prandial category, meal details as well as maternal & neonatal outcomes, complications and other data. All the aforementioned are used in clinical practice for the management of the woman.
A dataset to enable deeper research
The occurrence and type of complication present in the database, offer the opportunity to further investigate causal links between, for example, specific risk factors and maternal complications, leading to new interventions and treatments aimed at reducing these complications.
Management of diabetes in pregnancy requires a treatment plan that typically consists of diet and exercise advice, followed by pharmacological treatment if required (metformin or insulin). The various stages of this management plan are recorded in the database, enabling detailed analysis of the factors contributing to each intervention.
Reducing time to intervention may be a critical factor in reducing the risks to mother and child. By analysing detailed patterns in the data with machine learning methods, we can develop algorithms to reliably predict the likelihood of a patient needing pharmacological treatment, thereby helping clinicians to make timely decisions that may improve outcomes for pregnant women and their babies.
Through real-time monitoring of patients, the GDm-Health database has developed into a dense, longitudinal dataset that presents many opportunities for clinical research and discovery. For example, data collected by the system has enabled the design of algorithms derived from real-world data obtained from GDm-Health, which could help healthcare providers optimise their clinical decision making and allow intervention, including medication, to be delivered earlier.
- IDF Diabetes Atlas : https://diabetesatlas.org/data/en/country/209/gb.html
- Fraser, A., Lawlor, D.A., 2014. Long-Term Health Outcomes in Offspring Born to Women with Diabetes in Pregnancy. Curr. Diab. Rep. 14, 489. https://doi.org/10.1007/s11 892-014-0489-x
- Mackillop L, Hirst JE, Bartlett KJ, Birks JS, Clifton L, Farmer AJ, Gibson O, Kenworthy Y, Levy JC, Loerup L, Rivero-Arias O, Ming WK, Velardo C, Tarassenko L ‘Comparing the Efficacy of a Mobile Phone-Based Blood Glucose Management System With Standard Clinic Care in Women With Gestational Diabetes: Randomized Controlled Trial’ JMIR Mhealth Uhealth 2018;6(3):e71
- Hirst JE, Mackillop L, Loerup L, Kevat DA, Bartlett K, Gibson O, Kenworthy Y, Levy JC, Tarassenko L, Farmer A. Acceptability and user satisfaction of a smartphone-based, interactive blood glucose management system in women with gestational diabetes mellitus. J Diabetes Sci Technol. 2015 Jan;9(1):111-5