Development and validation of a clinical prediction model for first fall in early Parkinson's disease: a study of two...

Falls are frequent and debilitating complications in Parkinson's disease (PD), with a substantial risk present even in early stages. Predicting the first fall is critical for preventive interventions, yet existing models are often unsuitable for fall-naive, early PD patients due to their reliance on fall history.

This study aimed to develop and externally validate a clinical prediction model for first falls in a population of fall-naive, early PD patients.

This prognostic model study used data from two cohorts: the Parkinson's Progression Markers Initiative (PPMI) for model development (n = 283) and internal validation (n = 120), and an independent Chinese cohort for external validation (n = 150). Participants were fall-naive with early PD (Hoehn and Yahr stage 1-2) and were followed for 36 months. The primary outcome was time to the first fall. A Cox proportional hazards model was developed using readily accessible clinical variables. Model performance was assessed using discrimination (C-index, AUC), calibration, and decision curve analysis.

During follow-up, 16.9% of PPMI participants and 20.7% of the Chinese cohort experienced a first fall. The final model incorporated five independent predictors: lower body mass index, asymptomatic orthostatic hypotension, lower Montreal Cognitive Assessment score, a Geriatric Depression Scale-15 score > 5, and a higher postural instability and gait disorder score. The model demonstrated good discrimination with an optimism-corrected C-index of 0.844 in the training set and maintained its performance in both internal (C-index: 0.768) and external validation (C-index: 0.825). Decision curve analysis indicated that the model demonstrated superior clinical net benefit for predicting falls over 36 months compared to 18 months.

A history of falls is not necessary to predict the first fall in early PD. Our externally validated model, based on five easily ascertainable clinical factors, provides a practical tool for early risk stratification and can help guide individualized preventive strategies to delay or prevent initial falls in this vulnerable population.