Blood Glucose Prediction: The Lived Experience of Individuals with T1D

In this project, we focused on understanding the lived experience of individuals with type 1 diabetes (T1D) using personalized blood glucose predictions in their everyday lives.

Motivation: Blood glucose predictions could help Patients manage their condition proactively, in the example below, injecting additional insulin now could prevent high blood glucose values in an hour.

Problem: Integrating blood glucose predictions into the everyday decision-making of individuals with T1D may prove difficult. This difficulty stems from the highly individualized self-management practices needed to meet clinical guidelines, as well as prediction uncertainty and interpretation.

Goal:  With this project we wanted to understand whether people with T1D would want or need blood glucose predictions as well as their unique preferences, the importance of social context, and interaction preferences in regard to blood glucose prediction technologies.

Approach: We designed, developed, and deployed a mobile application (MOON-T1D), that simulated blood glucose level predictions, as depicted below. We conducted a 5-day experience sampling method study framed by two qualitative interviews with 15 individuals with T1D.

Results: Our reflexive thematic analysis revealed how predictions could facilitate or of participants, as well as the complex role of emotions and trust surrounding predictions. Our findings showed that participants wanted predictions for self-management, that they would take action in response to them and the substantial effect they would have on their management practices.

Accepted paper at CHI24