Application of artificial intelligence to measure and predict patient values and preferences: a scoping review

Abstract

Patients' voices are often difficult to capture directly in healthcare decisions. This scoping review examines how artificial intelligence (AI) has been applied to measure and predict patient values and preferences, aiming to evaluate its potential to generate reliable, patient-centered evidence, identify opportunities and challenges, and explore AI tools in literature reviews. Analyzing 67 studies, we summarize how AI processes diverse data sources such as social media, clinical records, and patient surveys to extract population- and individual-based patient values and preferences. Researchers have applied AI for efficient data preprocessing, extraction, analysis, integration, and modeling. Despite promising validation results (e.g., >80% accuracy in data preprocessing), key challenges remain, such as data quality issues, lack of real-world validation, and ethical concerns. This review underscores the potential of AI in patient values and preferences research and calls for greater transparency and real-world implementation to better align healthcare delivery with patient needs.

Ähnliche Arbeiten