Are We Missing the Environmental Factors in AI-Based Fall Risk Models?: A Systematic Review.

Area of Science:

• Gerontology
• Computer Science
• Public Health

Background:

• Falls are a significant risk for older adults, often linked to home environmental hazards.
• Environmental factors are modifiable and crucial for fall prevention strategies.
• Current AI fall prediction models primarily focus on individual factors, neglecting environmental influences.

Purpose of the Study:

• To systematically review the integration of environmental factors into AI-based fall risk prediction models.
• To summarize AI approaches and performance in predicting falls among community-dwelling older adults.
• To assess the role of environmental data in enhancing AI fall prediction models.

Main Methods:

• Systematic review adhering to PRISMA guidelines.
• Searched six major electronic databases from inception to December 2025.
• Included studies using AI models to predict falls in older adults, incorporating environmental factors.

Main Results:

• Nine studies met inclusion criteria, utilizing supervised machine learning, computer vision, or robotics.
• Environmental factors were diverse, from checklists to sensor/vision data.
• Inclusion of environmental features improved model discrimination (AUC-ROC 0.67-0.76) and identified hazards.

Conclusions:

• Environmental factors are underrepresented in current AI fall prediction models.
• Standardized, context-aware environmental data integration can enhance AI model relevance and preventive utility.
• Future research should focus on incorporating comprehensive environmental data for more effective fall prevention.