A nomogram incorporating sleep duration, ADL, waist circumference, cognitive function, handgrip strength, age, and depression demonstrated good discrimination for predicting frailty risk in middle-aged and older adults with cardiovascular disease, with AUC values of 0.851 and 0.861 in training and validation cohorts respectively.
Key Findings
Results
Frailty was identified in 12.5% of middle-aged and older adults with cardiovascular disease in the study sample.
148 of 1184 participants were identified as frail
Participants were aged ≥45 years with cardiovascular disease
Data were drawn from the 2015 China Health and Retirement Longitudinal Study (CHARLS)
The sample was randomly divided into training and validation cohorts at a 7:3 ratio
Results
Seven independent predictors of frailty were identified through LASSO regression followed by multivariable logistic regression.
Predictors identified were: sleep duration, activities of daily living (ADL), waist circumference, cognitive function, handgrip strength, age, and depression
LASSO regression was used for initial variable selection to reduce overfitting
Multivariable logistic regression was subsequently used to construct the nomogram model
The study used a cross-sectional design
Results
The nomogram demonstrated good discrimination for predicting frailty in both training and validation cohorts.
AUC of 0.851 (95% CI: 0.814–0.888) in the training cohort
AUC of 0.861 (95% CI: 0.804–0.917) in the validation cohort
Performance was consistent between cohorts, suggesting the model does not substantially overfit
Results
The nomogram showed good calibration between predicted and observed frailty outcomes.
Calibration was assessed using calibration plots and the Hosmer-Lemeshow test
Hosmer-Lemeshow test yielded P > 0.05 in both cohorts, indicating no significant lack of fit
Good agreement was found between predicted probabilities and observed outcomes
Results
Decision curve analysis indicated favorable clinical utility of the nomogram for frailty prediction.
Decision curve analysis (DCA) was used to evaluate clinical net benefit across a range of threshold probabilities
DCA results indicated the nomogram provided favorable clinical utility
The authors conclude the tool may facilitate early screening and risk stratification in clinical practice
Yang X, Zhou H, Huang C, Yuan M, Du X, Zhang C. (2026). Prediction Model for Frailty in Middle-Aged and Older Adults with Cardiovascular Disease.. Vascular health and risk management. https://doi.org/10.2147/VHRM.S581066