Timely Interventions for Diabetes Excellence (TIDE) is an open-source tool for the semi-automated analysis of population-level continuous glucose monitoring (CGM) data and for the algorithmic prioritization of patients with type 1 diabetes (T1D) developed by SURF Stanford Medicine.
The patient prioritization algorithm uses a flag system based on interpretable clinical criteria, such as the percentage of glucose readings between 70 and 180 mg/dL (time-in-range or TIR), below 70 mg/dL (hypoglycemia) and below 54 mg/dL (clinically significant hypoglycemia). The clinical indicators are calculated on the patients' CGM time-series over the previous 14 days.
The tool displays both aggregate population-level data and individual patient-level data, in which case the specific patient to view is selected interactively by the user by clicking on one of the patient's clinical indicators in the patients panel. The tool is intended to be used by care providers on a weekly basis in order to identify the patients in more urgent need of review.
This repository contains the Python implementation of TIDE, which is also available in R and Tableau. The code uses synthetic CGM time-series stored in a static CSV file, as this allows developers to clone the repository and run the code directly without the need of obtaining the API credentials from a specific CGM device provider. The code in this repository has been deployed on AWS Elastic Beanstalk without changes, and the web interface can be accessed here.
-
Clone the repository.
git clone https://github.com/flaviagiammarino/stanford-tide-dashboardcd stanford-tide-dashboard -
Install the requirements.
pip install -r requirements.txt -
Run the app.
python3 application.py
[1] Ferstad, J.O., Vallon, J.J., Jun, D., Gu, A., Vitko, A., Morales, D.P., Leverenz, J., Lee, M.Y., Leverenz, B., Vasilakis, C. and Osmanlliu, E., 2021. Population‐level management of type 1 diabetes via continuous glucose monitoring and algorithm‐enabled patient prioritization: Precision health meets population health. Pediatric Diabetes, 22(7), pp.982-991. doi:10.2196/27284
[2] Scheinker, D., Gu, A., Grossman, J., Ward, A., Ayerdi, O., Miller, D., Leverenz, J., Hood, K., Lee, M.Y., Maahs, D.M. and Prahalad, P., 2022. Algorithm-Enabled, Personalized Glucose Management for Type 1 Diabetes at the Population Scale: Prospective Evaluation in Clinical Practice. JMIR diabetes, 7(2), p.e27284. doi:10.1111/pedi.13256
[3] Scheinker, D., Prahalad, P., Johari, R., Maahs, D.M. and Majzun, R., 2022. A New Technology-Enabled Care Model for Pediatric Type 1 Diabetes. NEJM Catalyst Innovations in Care Delivery, 3(5), pp.CAT-21. doi:10.1056/CAT.21.04386