Development of Osteoporosis Screening Algorithm for Population Aged 50 Years and above in Klang Valley, Malaysia

Abstract

Background: The current osteoporosis screening instruments are not optimized to be used among the Malaysian population. This study aimed to develop an osteoporosis screening algorithm based on risk factors for Malaysians. Methods: Malaysians aged ≥50 years (n = 607) from Klang Valley, Malaysia were interviewed and their bone health status was assessed using a dual-energy X-ray absorptiometry device. The algorithm was constructed based on osteoporosis risk factors using multivariate logistic regression and its performance was assessed using receiver operating characteristics analysis. Results: Increased age, reduced body weight and being less physically active significantly predicted osteoporosis in men, while in women, increased age, lower body weight and low-income status significantly predicted osteoporosis. These factors were included in the final algorithm and the optimal cut-offs to identify subjects with osteoporosis was 0.00120 for men [sensitivity 73.3% (95% confidence interval (CI) = 54.1%–87.7%), specificity 67.8% (95% CI = 62.7%–85.5%), area under curve (AUC) 0.705 (95% CI = 0.608–0.803), p < 0.001] and 0.161 for women [sensitivity 75.4% (95% CI = 61.9%–73.3%), specificity 74.5% (95% CI = 68.5%–79.8%), AUC 0.749 (95% CI = 0.679–0.820), p < 0.001]. Conclusion: The new algorithm performed satisfactorily in identifying the risk of osteoporosis among the Malaysian population ≥50 years. Further validation studies are required before applying this algorithm for screening of osteoporosis in public.