Development and validation of a novel screening tool for osteoporosis in older US adults: The NHANES cross-sectional study

Abstract

Purpose: Osteoporosis (OP) is a common disease among adults aged >50 years. At present, the main approach to screen or to diagnosis OP is mainly via bone mineral density (BMD) testing, which might not be optimal for OP screening. This study aimed to develop and validate a convenient and effective prediction model for screening OP based on the demographic information, medical history, and lifestyle habits in the elderly in the United States.

Methods: All data were collected from the National Health and Nutrition Survey database. Participants aged ≥50 years with complete BMD data were included in this study. Twelve candidate predictors were initially selected to develop the prediction model. Final predictors screening and model development were based on multivariate logistic regression. Model discrimination (C statistic) and calibration (Brier scores) were calculated to evaluate the performance of the model. Internal validation was performed using the bootstrap resampling technique, and external validation was based on the validation cohort.

Results: The screening tool was developed with individual patient data from 1941 patients and validated with data from 1947 patients after the development of the model. Seven predictors (patient age, sex, race, body mass index, physical activity, sleep duration, and history of fracture) were included in the final prediction model, and the final model had a C statistic of 0.849 [95% confidence interval (CI): 0.820-0.878] and Brier scores of 0.062 [95% CI: 0.054-0.070] on the development cohort. For the validation of the developed model, the results showed a C statistic >0.800 and Brier scores <0.070, irrespective of internal validation or external validation.

Conclusions: A novel screening tool for OP in the elderly, which has excellent discrimination and useful calibration, has been developed and externally validated. Considering its simplicity, generalizability, and accuracy, this tool has the potential to become a practical mean for the elderly to screen OP.

Keywords: External validation; National Health and Nutrition Survey; Osteoporosis; Prediction model; The elderly.