Exploring Location Data as a Predictor for Blood Glucose in Type 1 Diabetes: A Systematic Review (Preprint)
Preprint
- 6 November 2025
- preprint
- Published by JMIR Publications Inc. in JMIR Preprints
Abstract
BACKGROUND Managing Type 1 Diabetes (T1D) requires regular glucose monitoring and appropriate insulin dose adjustments. Although the use of continuous glucose monitoring (CGM) sensors has been beneficial, there are still inherent delays in CGM measurements and insulin onset of action, making accurate glucose prediction essential. Smartphones can collect Global Positioning System (GPS) data that can be converted into location categories (e.g., “gym,” “cafe,” “restaurant”), which provide information about a person’s location and offer insight into their behavior, and both location and behavior may influence blood glucose levels. OBJECTIVE This systematic review aims to evaluate existing research on the use of location category data as a predictor of blood glucose fluctuations in individuals with diabetes. It explores whether such data have been used to identify location categories where people with diabetes are more likely to be out of range, potentially supporting timely corrective actions. METHODS The systematic review was conducted following PRISMA guidelines, identifying studies examining the use of semantic or geographic location category data for blood glucose prediction in individuals with diabetes. Eligible studies were analyzed for the location category data used, predictive modeling approaches, and outcome measures. RESULTS 665 screened studies, only three met the inclusion criteria. All were from a single research project involving 40 individuals with Type 2 Diabetes (T2D), monitored over a period of 3 days. These studies utilized geographic and temporal data but did not classify places by location category. No studies investigated the use of the location category in the context of T1D. CONCLUSIONS No studies have used location categories to predict blood glucose levels in individuals with T1D. Limited research in T2D has incorporated GPS data, but without identifying specific place types such as restaurants, gyms, or workplaces. In contrast, mental health research has effectively applied location-based methods to predict stress, anxiety, and depression, showing that the places people visit and the time they spend there reflect important behavioral patterns. Because diabetes management also relies on daily behaviors such as eating, physical activity, and routine, applying these methods from mental health research may provide new insights into how specific locations influence blood glucose variability and support more timely, personalized diabetes management strategies.Keywords
This publication has 27 references indexed in Scilit:
- Association of secondhand smoke exposure with all-cause mortality and cardiovascular death in patients with hypertension: Insights from NHANESNutrition, Metabolism and Cardiovascular Diseases, 2024
- Daily mobility patterns in patients with bipolar disorder and healthy individualsJournal of Affective Disorders, 2021
- Design and evaluation of a context-aware model based on psychophysiologyComputer Methods and Programs in Biomedicine, 2020
- An integrated framework for using mobile sensing to understand response to mobile interventions among breast cancer patientsSmart Health, 2020
- GPS mobility as a digital biomarker of negative symptoms in schizophrenia: a case control studynpj Digital Medicine, 2019
- Understanding behavioral dynamics of social anxiety among college students through smartphone sensorsInformation Fusion, 2019
- Improving momentary stress measurement and prediction with bluetooth encounter networksSmart Health, 2018
- Hypoglycemia in DiabetesNursing Clinics of North America, 2017
- Understanding Post-Prandial Hyperglycemia in Patients with Type 1 and Type 2 Diabetes: A Web-based Survey in Germany, the UK, and USADiabetes Therapy, 2016
- Diabetes and behavior: A paradigm for health psychology.American Psychologist, 1983