Multimodal In-Vehicle Hypoglycemia Warning for Drivers With Type 1 Diabetes: Design and Evaluation in Simulated and...

Area of Science:

• Human-Computer Interaction
• Diabetes Management
• Automotive Safety

Background:

• Hypoglycemia poses risks to cognitive function and driving safety.
• In-vehicle voice assistants have been explored for hypoglycemia warnings but may startle drivers.
• Combining voice warnings with ambient LEDs offers a potential solution to mitigate startling effects.

Purpose of the Study:

• To assess the impact of in-vehicle multimodal warnings on drivers with type 1 diabetes.
• To evaluate emotional reactions and technology acceptance related to hypoglycemia alerts.
• To compare the effectiveness and preference of different warning modalities.

Main Methods:

• Quasi-experimental design with simulated and real-world driving studies.
• Independent variables: blood glucose phase and warning modality (voice, LED, voice+LED).
• Dependent variables: physiological (skin conductance) and subjective emotional reactions (Affective Slider), technology acceptance.

Main Results:

• Self-reported emotional reactions were significantly affected by warning modality and blood glucose phase in real-world driving.
• The voice+LED warning modality was consistently rated as most effective and urgent across both driving conditions.
• Voice+LED and voice warnings were preferred over standard warnings, with voice+LED being highly preferred in simulated driving.

Conclusions:

• Multimodal alerts, particularly voice combined with LEDs, show potential for enhancing hypoglycemia management in drivers.
• Further research is warranted to optimize in-vehicle health warning systems for improved driving safety.
• This study advocates for the integration of advanced warning systems to support individuals with diabetes during driving.