Development and validation of the Chinese Osteoporosis Screening Algorithm (COSA) in Identification of People with High Risk of Osteoporosis

Abstract Osteoporosis is a prevalent but under-diagnosed disease partly due to the limited availability of dual-energy X-ray absorptiometry (DXA) machines in many countries. To enhance the public awareness and facilitate diagnosis of osteoporosis, we aimed to develop a new Chinese Osteoporosis Screening Algorithm (COSA) and compare its performance with the Osteoporosis Self-Assessment Tool for Asians (OSTA) to identify people at high risk of osteoporosis. A total of 4,747 postmenopausal women and men aged≥50 years from the Hong Kong Osteoporosis Study were randomly split into a development (N=2,373) and an internal validation cohort (N=2,374). An external validation cohort comprising 1,876 community-dwelling subjects was used to evaluate the positive predictive value (PPV). Among 11 predictors included, age, sex, weight, and history of fracture were significantly associated with osteoporosis after correction for multiple testing in the development cohort. Age- and sex-stratified models were developed separately for women aged<65 years, women aged≥65 years, and men due to the presence of significant sex and age interactions. The AUC of the COSA in the internal validation cohort was 0.761 (95% CI: 0.711-0.811), 0.822 (95% CI: 0.792-0.851), and 0.946 (95% CI: 0.908-0.984) for women aged<65 years, women aged≥65 years, and men, respectively. The COSA outperformed OSTA in identifying people with osteoporosis in all the three groups. In the external validation cohort, the PPV of COSA was 40.6%, 59.4%, and 19.4% for women aged<65 years, women aged≥65 years, and men, respectively. In conclusion, we have developed and validated a new osteoporosis screening algorithm, COSA, specific for Hong Kong Chinese.