Examining Passively Collected Smartphone-Based Data in the Days Prior to Psychiatric Hospitalization for a Suicidal...

Area of Science:

• Digital phenotyping and mental health research.
• Suicide risk detection and prediction.
• Mobile health (mHealth) and digital biomarkers.

Background:

• Digital phenotyping is increasingly used in clinical research, but passive assessment for suicide risk remains underexplored.
• Screenomics, a novel digital phenotyping approach, captures smartphone activity through screenshots, offering a unique data stream.
• This study explores screenomics for detecting suicide risk, a critical area with limited passive assessment research.

Observation:

• The study reviewed two participants with recent suicidal ideation, analyzing their passive (screenomics) and active (ecological momentary assessment - EMA) risk profiles.
• Data were collected over 35 days of 5-second screenshot intervals and 28 days of 6-times-daily EMA prompts.
• Analysis focused on suicide-related content, risk factors, and social content from both screenomics and EMA data.

Findings:

• Participants showed decreased EMA compliance and increased overall phone usage, particularly social use, during suicidal crises preceding hospitalization.
• Screenomics identified key precipitating factors like physical pain and loneliness, which were not adequately captured by self-report (EMA).
• Passive screenomics data provided granular insights into digital interactions leading up to a suicidal crisis.

Implications:

• Passively collected screenomics data holds significant potential for understanding and predicting suicidal crises.
• Integrating screenomics with EMA offers a more comprehensive assessment of psychological processes during high-risk periods.
• This approach can reveal novel risk factors not detectable through self-report alone, improving suicide prevention strategies.