講演情報
[SY-66-02]AIOT (AI+IOT) based prediction system for suicide/aggressive behavior in psychiatric wards
*Hwa-Young Lee Lee1 (1. Soonchunhyang University Cheonan Hospital (Korea))
キーワード：
AI、IOT、suicide、Aggression
Objective: Aggression is a psychiatric emergency and predicting it in psychiatric inpatients enhances the efficacy and safety of patient management. We established a real-time vital sign monitoring system based on AIoT technology in the psychiatric ward and compiled a collection of clinical scales to predict and evaluate aggression in hospitalized patients.
Methods: Existing scales were comprehensively reviewed to select items for a set of clinical scales that could predict a crisis that immediately preceded aggressive behavior. The work is based on an understanding of how aggressive behavior develops, and the contributory factors. To establish a ward environment based on AIoT technology, a monitoring dashboard, vision sensors, and object interaction sensors were implemented.
Results: Seven clinical scales, including the Nurses’ Global Assessment of Suicide Risk, the Positive and Negative Syndrome Scale (PANSS), the Modified Overt Aggression Scale (MOAS), the Inpatient Aggression Prediction Scale, the Broset Violence Checklist (BVC), the Staff Observation Aggression Scale-Revised (SOAS-R), and the State-Trait Anger Expression Inventory (STAXI), were used in the aggression monitoring protocol. Among all modalities, vision-based monitoring showed the highest predictive response (68.8%), followed by clinical assessments (53%)
Conclusion: The establishment of a ward environment utilizing AIoT technology to enable the early prediction of aggression/suicide in psychiatric inpatients is anticipated to aid in creating a safer ward environment through further refinement processes.
Methods: Existing scales were comprehensively reviewed to select items for a set of clinical scales that could predict a crisis that immediately preceded aggressive behavior. The work is based on an understanding of how aggressive behavior develops, and the contributory factors. To establish a ward environment based on AIoT technology, a monitoring dashboard, vision sensors, and object interaction sensors were implemented.
Results: Seven clinical scales, including the Nurses’ Global Assessment of Suicide Risk, the Positive and Negative Syndrome Scale (PANSS), the Modified Overt Aggression Scale (MOAS), the Inpatient Aggression Prediction Scale, the Broset Violence Checklist (BVC), the Staff Observation Aggression Scale-Revised (SOAS-R), and the State-Trait Anger Expression Inventory (STAXI), were used in the aggression monitoring protocol. Among all modalities, vision-based monitoring showed the highest predictive response (68.8%), followed by clinical assessments (53%)
Conclusion: The establishment of a ward environment utilizing AIoT technology to enable the early prediction of aggression/suicide in psychiatric inpatients is anticipated to aid in creating a safer ward environment through further refinement processes.