GlucoMine: A Case for Improving the Use of Wearable Device Data in Diabetes Management
summary
The paper builds an algorithm called GlucoMine that discretizes CGM traces, applies a grid-based week-time stacking and extracts recurrent adverse events from 3–6 months of data for 54 people with type 1 diabetes. It shows standard practice of reviewing 14 days misses patterns: 96% of subjects have ≥15 minutes of recurrent events, 85% ≥1.5 hours, 65% ≥3 hours, and clinicians report an 89% perceived benefit for an automated long-range pattern extractor; a clinician study compares strengths and weaknesses of conventional plots versus the new view.
AI-written summary · about
related talks
22mCrafting a View of Self-Tracking Data in the Clinical Visit
01h 14mDigitization and Consumerization of Healthcare: The Evolving Role of Devices, Data, and Analytics
27mRevolutionizing Medical Monitoring with High-Performance ECG Apps
14mTelling Traces
01h 04mTake Control of the Data of You
11mBig Data Moonshots and Ground Control
17mEnhancing Benefit Adjudication Through Graph Node Embedding, Clustering, and Outlier Detection
17mQuantifying Sources and Types of Smartwatch Usage Sessions