Construction and verification of a nomogram model for predicting the risk of post-stroke spasticity: a retrospective study

TL;DR

This study developed and validated a nomogram model to predict the risk of post-stroke spasticity using clinical data from over 300 patients.

Contribution

A novel nomogram model for predicting post-stroke spasticity was constructed and validated using routine clinical data and machine learning techniques.

Findings

The model achieved an AUC of 0.844 in the training set and 0.842 in the validation set, showing strong predictive accuracy.

Calibration curves demonstrated good agreement between predicted and observed probabilities in both training and validation sets.

Decision curve analysis and clinical impact curve confirmed the model's practical applicability in clinical settings.

Abstract

Objective: The aim of this study is to develop and validate a nomogram model for predicting the risk of post-stroke spasticity (PSS).

Methods: A retrospective study collected data from 366 stroke patients admitted to Guangzhou University of Chinese Medicine’s Dongguan Hospital between January 2022 and April 2025. PSS was defined as a Modified Ashworth Scale (MAS) score ≥1 within 3 months after stroke. The patients were divided into training and validation sets using a 7:3 ratio. Using multivariate logistic regression analysis and least absolute shrinkage and selection operator (Lasso) regression, a predictive model was built, and a nomogram was created for clinical application. The receiver operating characteristic (ROC) curve and calibration curve were plotted to assess the discrimination and calibration of the model. Decision curve analysis (DCA) and the clinical impact curve (CIC) were used to evaluate the model’s clinical applicability and usefulness.

LASSO-logistic regression analysis identified seven predictors associated with PSS: C-reactive protein, albumin, creatine kinase, fasting blood glucose, hyperlipidemia, sleep disorders, and manual muscle testing (MMT) score at admission. The model had an area under the curve (AUC) of 0.844 (95% CI: 0.793–0.896) in the training set and 0.842 (95% CI: 0.765–0.920) in the validation set, which means it was good at making predictions. The calibration curves showed excellent agreement between predicted and observed probabilities in the training set. Good calibration was maintained in the validation set, indicating only minimal overestimation of risk. DCA and CIC both agreed that the nomogram model could be used in a wide range of therapeutic situations.

The nomogram based on routine clinical data in this study, after internal validation, can effectively predict the risk of PSS and provides a practical decision-making tool for clinicians. However, future multi-centre external validation is still required to confirm its broad applicability.

Figures