FALLPREDICT: Predicting Fall-Related Health Care Encounters in Persons with Vascular Lower Limb Amputation Using...

Area of Science:

• Rehabilitation Medicine
• Biomedical Engineering
• Gerontology

Background:

• Lower limb amputation (LLA) increases fall risk, particularly in patients with diabetes and peripheral artery disease (PAD).
• Effective fall prediction models are crucial for implementing targeted prevention strategies in this vulnerable population.

Purpose of the Study:

• To develop and validate the FALLPREDICT model for predicting 12-month fall risk in patients post-LLA.
• To identify key predictors of falls in individuals undergoing transtibial (TT) or transfemoral (TF) amputation.

Main Methods:

• Retrospective cohort study of 1,690 Veterans with dysvascular LLA (TT or TF).
• Data sourced from VA Corporate Data Warehouse and National Prosthetics Patient Database.
• Incident falls identified via ICD-10 codes within 12 months of definitive lower limb prosthesis (LLP) prescription.

Main Results:

• The FALLPREDICT model, with nine predictors (including amputation level, age, BMI, and prior falls), demonstrated good discrimination (c-statistic 0.71).
• The model showed satisfactory calibration, providing reliable absolute risk estimates for falls.
• Key predictors included amputation level, age, BMI, prior falls, depression, peripheral neuropathy, kidney dialysis, and COPD.

Conclusions:

• The FALLPREDICT model is a valuable tool for identifying patients with LLA at high risk of future falls.
• This risk stratification can inform the development of personalized and targeted fall intervention strategies.