Scalable Precision Psychiatry With an Objective Measure of Psychological Stress: Prospective Real-World Study

Area of Science:

• Digital psychiatry
• Artificial intelligence in mental health
• Biometric data analysis

Background:

• Objective assessment of mental well-being is crucial for advancing precision psychiatry.
• Current subjective measures may lack the fidelity needed for unbiased therapeutic recommendations.

Purpose of the Study:

• To compare the efficacy of a precision psychiatry therapy recommendation algorithm trained with objective psychological stress data versus subjective assessments.
• To evaluate if objective data improves the accuracy of predicting effective therapy sessions.

Main Methods:

• Trained a therapy recommendation algorithm using data from 67 "super users" over 28+ psychotherapy sessions.
• Utilized objective stress level (OSL) from facial biomarkers and subjective ecological momentary assessments (EMAs) for stress (SRS) and mood (SRM).
• Compared algorithm predictions trained on delta OSL, delta SRS, and delta SRM against self-selected therapy efficacy.

Main Results:

• Objective stress measure (∆OSL) showed differentiated therapeutic efficacy compared to subjective measures (∆SRS, ∆SRM).
• Algorithm training with objective data (∆OSL) significantly improved prediction efficacy over self-selected therapy (P<.01).
• Subjective EMA training showed non-significant trends, and self-selected sessions were overrepresented in recommendations.

Conclusions:

• A rapid, scalable, objective measure of psychological stress, combined with a recommendation algorithm, can identify effective therapy autonomously.
• Objective mental well-being data holds significant potential for improving precision psychiatry and meeting global mental health demand.