Health risk prediction models incorporating personality data: Motivation, challenges, and illustration

Area of Science:

• Health Informatics
• Psychological Medicine
• Biostatistics

Background:

• The era of big data in health necessitates advanced prediction models.
• Genomic data is commonly used, but psychosocial data offers a novel augmentation.
• Personality measures are proposed to enhance traditional health predictors.

Purpose of the Study:

• To provide an overview of medical risk prediction models.
• To discuss the rationale for integrating personality data into health risk prediction.
• To propose principles for the effective use of personality data in healthcare.

Main Methods:

• Review of medical risk prediction models.
• Discussion of integrating personality measures.
• Development of a personality-augmented prediction model for mild cognitive impairment.
• Evaluation of model performance and clinical utility.

Main Results:

• Three principles proposed: specific outcomes, incremental validity, and responsible model-building.
• A personality-augmented model for mild cognitive impairment was developed and evaluated.
• The study highlights the potential and challenges of personality data in clinical settings.

Conclusions:

• Personality data holds significant potential for enhancing health risk prediction models.
• Achieving clinical utility requires addressing challenges in real-world implementation.
• An iterative, programmatic approach is necessary for successful integration into healthcare.