Evaluating multivariable clinical prediction models for Parkinson's disease prognosis: A scoping review

Multivariable clinical prediction models combine two or more patient predictor variables to calculate the probability of a clinical outcome. In persons with Parkinson's disease (PD), clinical prognostic models can be used to predict disease progression or the occurrence of a clinical event. These models can aid in PD prognostication and complement clinical judgment by informing medical decision-making and resource allocation. In this scoping review of the literature from inception through May 2025, we present data from 74 multivariable prognostic models in PD, organized by clinical outcome. We also describe model development, validation, and calibration, if performed, and identify knowledge gaps. By summarizing existing multivariable clinical prognostic models in PD, this review is designed to inform prognostication and foster the development of more accurate, validated prognostic models for use in clinical practice.