Development and Validation of A Screening Tool for Sarcopenia in Community-dwelling Older Adults: A Diagnostic Cross-Sectional Study

Background:To develop a high-performance, easy-to-operate prediction tool for sarcopenia screening in Chinese community-dwelling older adults. Methods: This cross-sectional study conducted in community settings, using univariate and multiple logistic regression to select predictors and develop prediction models. A user-friendly nomogram was developed based on the final model. The model was evaluated by the area under the receiver operating curve (AUROC) for discrimination, calibration curves for calibration, and decision curve analysis (DCA) curve for clinical utility. Results: A total of 2453 individuals aged ≥60 years were included. Participants were randomly divided into the training and validation sets in a 7:3 ratio. In the training dataset, optimal model(Model 1) includes six variables: age, sex, BMI, calf circumference, diastolic blood pressure, and sitting duration (hours) with an AUROC of 0.8720(95%CI, 0.8510-0.8930), and the simplified model(Model 2) includes only the first four variables from the optimal model with an AUROC of 0.8688(95%CI,0.8476-0.8900). In the validation dataset, the AUROC for the simplified model demonstrated reasonable concordance (0.8465,95%CI, 0.8102-0.8828)(Model 3). At a threshold probability of 0.111, the sensitivity and specificity were as follows: 0.8683(95%CI, 0.8258-0.9108) and 0.7186(95%CI, 0.6957-0.7416) for simplified model, 0.7941(95%CI,0.7156-0.8726) and 0.7062(95%CI,0.6707-0.7416) for validation model. Calibration curves showed good agreement .The DCA showed the prediction model had clinical utility for screening sarcopenia among Chinese older adults. Conclusions: The study developed a high-performance prediction tool that serves as a low-cost, user-friendly, and freely available tool for community and self-screening of sarcopenia, facilitating personalized early prevention strategies and optimizing healthcare resource utilization. Clinical trial number: Not applicable.