Development of a Voice-Enabled Automated Sleep Survey System

Area of Science:

• Healthcare technology
• Sleep medicine
• Artificial Intelligence in healthcare

Background:

• Hospitalized patients often experience poor sleep quality.
• Existing sleep assessment methods can be time-consuming and lack comprehensive environmental data.
• There is a need for efficient, patient-centric tools to manage sleep disturbances in clinical settings.

Purpose of the Study:

• To develop and evaluate a conversational agent for enhanced sleep quality assessment in hospitalized patients.
• To identify environmental and care-related factors impacting patient sleep using AI.
• To explore the potential of non-pharmacological interventions based on AI-driven sleep analysis.

Main Methods:

• Development of a conversational agent integrating GPT-4 and large language models.
• Utilized the Richards-Campbell Sleep Questionnaire and additional environmental questions.
• Processed patients' spoken responses to identify sleep-impacting factors.
• Assessed system's recognition accuracy for sleep-related factors.

Main Results:

• The conversational agent successfully streamlines sleep data collection and analysis.
• The system accurately identifies environmental (e.g., temperature, lighting) and care-related factors affecting sleep.
• Demonstrated high recognition accuracy, indicating potential for effective sleep quality management.

Conclusions:

• Conversational AI offers a promising approach to patient-centric sleep assessment in hospitals.
• The developed agent can identify key factors influencing sleep disturbances.
• This technology supports the advancement of non-pharmacological interventions for improving hospital sleep quality.