Development and validation of the Chinese osteoporosis screening algorithm (COSA) in identification of people with high risk of osteoporosis

Abstract

Objectives

To enhance the public awareness and facilitate diagnosis of osteoporosis, we aim to develop a new Chinese Osteoporosis Screening Algorithm (COSA) to identify people at high risk of osteoporosis.

Methods

A total of 4747 postmenopausal women and men aged ≥ 50 from the Hong Kong Osteoporosis Study were randomly split into a development (N = 2373) and an internal validation cohort (N = 2374). An external validation cohort comprising 1876 community-dwelling subjects was used to evaluate the positive predictive value (PPV).

Results

Among 11 predictors included, age, sex, weight, and history of fracture were significantly associated with osteoporosis after correction for multiple testing. Age- and sex-stratified models were developed due to the presence of significant sex and age interactions. The area under the curve 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, women aged ≥ 65, and men, respectively. The COSA demonstrated improved reclassification performance when compared to Osteoporosis Self-Assessment Tool for Asians. In the external validation cohort, the PPV of COSA was 40.6%, 59.4%, and 19.4% for women aged < 65, women aged ≥ 65, and men, respectively. In addition, COSA > 0 was associated with an increased 10-year risk of hip fracture in women ≥ 65 (OR, 4.65; 95% CI, 2.24–9.65) and men (OR, 11.51; 95% CI, 4.16–31.81).

Conclusions

We have developed and validated a new osteoporosis screening algorithm, COSA, specific for Hong Kong Chinese.