Evaluating the Prognostic and Clinical Validity of the Fall Risk Score Derived From an AI-Based mHealth App for Fall...

Area of Science:

• Gerontology
• Public Health
• Biomedical Engineering

Background:

• Falls are a major public health issue, exacerbated by an aging population, leading to severe health consequences.
• Identifying high-risk individuals, particularly older adults in residential care, is critical for effective fall prevention.
• Multifactorial, personalized fall risk assessment is recommended by current guidelines.

Purpose of the Study:

• To evaluate the prognostic validity of the Fall Risk Score (FRS) using real-world longitudinal data.
• To establish clinical relevance by determining FRS threshold values and minimum clinically important differences.
• To assess the FRS's predictive accuracy for future falls in older adults.

Main Methods:

• A retrospective cohort study of 617 older adults in German residential care facilities.
• Utilized the LINDERA mobile health app for fall risk assessment and collected longitudinal data.
• Employed quadratic regression and Spearman correlation to analyze the association between FRS and subsequent falls.

Main Results:

• A higher FRS at baseline significantly correlated with an increased number of falls at follow-up (ρ=0.960, P<.001).
• Strong correlations were observed in subgroups, including those with slower gait speeds (ρ=0.954) and walking aid users (ρ=0.955).
• Established FRS thresholds: 45% (6 months), 32% (12 months), and 24% (24 months) predict a fall.

Conclusions:

• The FRS demonstrates strong prognostic validity for predicting falls in older adults, particularly in identified subgroups.
• Findings support a stratified approach to fall risk assessment and highlight the need for early, personalized interventions.
• The study provides valuable data on FRS clinical relevance, aiding in fall prevention strategies for vulnerable populations.