A nomogram for predicting osteoporotic fracture: Establishment and validation of based on a retrospective multicenter study

Purpose We here aimed to develop a nomogram to identify patients with a high risk of osteoporotic fracture. Methods We conducted a multicentre hospital study. A development cohort consisting of patients from three hospitals was used to identify the predictors of osteoporotic fracture through univariate and multivariable logistic regression analyses and to construct a nomogram. The C-statistic, calibration plot, and decision curve analysis were calculated to evaluate discrimination, calibration, and clinical usefulness of the nomogram, respectively. The nomogram was further validated in the validation cohort (1 hospital) and internally validated by bootstrap. Results A total of 27,658 patients were enrolled from January 2018 to December 2022. Osteoporotic fracture was confirmed in 15,324 (71.2%) of 21,525 and 4,030 (65.7%) of 6,133 in development and validation cohorts respectively. Gender, increased age, urbanization, osteoporosis, hypoproteinemia, Parkinson’s disease, hypertension, heart failure, and chronic kidney disease were independent risk factors for osteoporotic fracture. The C-statistic was 0.82 (95% CI , 0.81–0.82) based on the development cohort. Similar C-statistic values were achieved during internal (0.82 [95% CI , 0.81–0.82]) and external validation (0.71 [95% CI , 0.70–0.73]). Calibration plots were well fitted and DCA curves indicated that the clinical validity of the model was best when the threshold probability was 0.4–1.0. Conclusion The nomogram established in this study could better predict the risk of osteoporotic fracture. After considering and discussing the prediction with patients, physicians can establish a rational therapeutic plan.