Realising stratified psychiatry using multidimensional signatures and trajectories

Area of Science:

• Psychiatry
• Computational psychiatry
• Translational medicine

Background:

• Traditional psychiatric diagnoses are categorical labels for heterogeneous disorders.
• Current quantitative methods in psychiatry assume categorical diagnoses and yield average treatment responses.
• This approach sacrifices valuable predictive information for personalized treatment.

Purpose of the Study:

• To operationalize a framework for 'stratified psychiatry' enabling personalized treatment.
• To develop multidimensional representations of disorder definition, clinical state, and treatment outcomes.
• To demonstrate the application of this framework using real-world clinical trial data.

Main Methods:

• Proposed a formal multidimensional representation of disorder definition and clinical state.
• Redefined outcomes as multidimensional constructs to reveal patient-specific differences.
• Applied the framework to the Clinical Antipsychotic Trials of Intervention Effectiveness dataset.

Main Results:

• Demonstrated heterogeneity in patient clinical states and treatment trajectories lost in traditional approaches.
• Showcased how multidimensional outcomes expose within- and between-patient differences.
• Identified existing clinical trials that could be reanalyzed with the proposed framework.

Conclusions:

• Quantitative methods can realize stratified psychiatry through multidimensional modeling.
• This approach has implications for clinical trial design and personalized treatment in psychiatry.
• Probabilistic reasoning tools are necessary for implementing stratification effectively.