Predictive model for falling in Parkinson disease patients.

Area of Science:

• Neurology
• Gerontology
• Movement Disorders

Background:

• Falls are a significant complication in advancing Parkinson's disease (PD).
• Existing risk factors for falls in PD are insufficient for reliable prediction.
• Predicting future falls is crucial for managing Parkinson's disease progression.

Purpose of the Study:

• To develop a multivariate model for predicting falls in Parkinson's disease patients.
• To identify key predictors of falling in individuals with PD.
• To enhance fall risk assessment in clinical settings for PD.

Main Methods:

• A prospective cohort study involving 49 Parkinson's disease patients.
• Development and evaluation of a multivariate predictive model for falls.
• Area Under the Receiver Operating Characteristic Curve (AUC) analysis was used to assess model performance.

Main Results:

• The final multivariate model included PD duration, Freezing of Gait (FOG), Audit of Cognitive Emoticons (ACE), and physical activity.
• The model demonstrated high predictive performance with an AUC of 0.9282.
• The model achieved 89.83% correct classification, 92.68% sensitivity, and 83.33% specificity.

Conclusions:

• The developed multivariate model exhibits high performance in predicting falls among Parkinson's disease patients.
• This predictive tool can aid clinicians in identifying individuals with PD at higher risk of falling.
• Further research can refine this model for broader clinical application in Parkinson's disease management.