Models Containing the Braden Scale Predict Activities of Daily Living Improvement in Post‐Stroke Hemiplegia: A Single Center Retrospective Cohort Study

BACKGROUND AND PURPOSE: This study sought to develop and validate a predictive model integrating the Braden scale to forecast the recovery of activities of daily living (ADL) in patients with post-stroke hemiplegia. Additionally, it aimed to evaluate the effectiveness of a comprehensive rehabilitation program. METHODS: A retrospective analysis was conducted on 324 patients with hemiplegic stroke admitted to our hospital between August 2023 and August 2025. Candidate variables were selected from demographics, comorbidities, neuroimaging features, laboratory findings, early treatment, complications, Braden scale scores and ADL assessments. Patients were categorized into low and high ADL recovery groups based on the median post-treatment ADL score. A logistic regression model was developed and presented as a nomogram. Model performance was assessed using receiver operating characteristic (ROC) analysis, calibration curves, and decision curve (DCA) analysis, followed by internal validation. Additionally, in a subgroup of 82 patients who underwent comprehensive rehabilitation, detailed functional assessments including motor function, balance, walking ability, and activity of daily living (ADL) were conducted before and after intervention. RESULTS: A total of 324 patients with post-stroke hemiplegia were included in the analysis. Comparative analysis between the high and low ADL recovery groups revealed that the high recovery group was significantly younger (p < 0.001) and had a higher proportion of females. This group also exhibited lower CRP levels (p = 0.003), higher albumin (ALB, p < 0.001) and cholesterol (CHOL) levels (p = 0.039), a lower rate of multi-lesion involvement (p < 0.001), a higher rate of thrombolysis (p = 0.019), a lower incidence of infection (p < 0.001), a lower rate of specific medicine intake (p = 0.021), higher baseline ADL scores at admission (p < 0.001), and higher Braden scale scores both before and after treatment (p < 0.001). Multivariable logistic regression identified sex, thrombolysis, infection incidence, and post-treatment Braden scale as independent predictors of ADL recovery. The nomogram showed excellent discrimination, with an AUC of 0.888 (95% CI: 85.4%-92.2%). At the optimal cut off value of 0.55, the model achieved a sensitivity of 84.3% and a specificity of 76.6%. In the rehabilitation subgroup (N = 82), significant post-treatment improvements were observed in motor function (upper limb p < 0.001, hand p = 0.002, lower limb p < 0.001), balance (Berg Balance Scale and Fugl-Meyer assessment; both p < 0.001), walking capability (p < 0.001), and overall performance (p < 0.001). DISCUSSION: The findings indicate that sex, incidence of infection, thrombolysis, and the post-treatment Braden scale were independent predictors of ADL recovery. Moreover, the comprehensive rehabilitation program was associated with significant functional improvements in motor skills, balance, ADL, and walking proficiency.