Performance of four clinical screening tools to select peri- and early postmenopausal women for dual X-ray...

Several methods to select postmenopausal women for dual X-ray absorptiometry (DXA) have been proposed. We decided to compare the performance of three clinical decision rules (SCORE, ORAI, OST) with the usual case-finding strategy based on the presence of a major risk factor for future fracture (CFMRF). The study subjects were 2009 healthy, white, peri- or early postmenopausal women participating in the Danish Osteoporosis Prevention Study (DOPS). DXA results expressed as T-scores and scores on SCORE, ORAI, OST and CFMRF were extracted from the DOPS database. First, we evaluated the screening tools as originally described by the developers. The resulting sensitivities and specificities ranged from 18% to 92% and from 66% to 85%, respectively. Only OST achieved a high sensitivity (92%) with respect to femoral neck T-score < or = -2.5; however, the sensitivity with respect to lumbar spine T-score < or = -2.5 was only 51%. Next, the performance of the screening tools was evaluated against T-score < or = -2.0 (and T-score < or = -2.5) in at least one of the regions: femoral neck, total hip or lumbar spine. Using ROC curve analysis, we determined cut-offs yielding sensitivities as close as possible to 90%. The CFMRF and the ORAI tool were too coarse to yield 90% sensitivity. The performances of OST and SCORE were equal from a clinical perspective in that the sensitivities and the specificities varied from 89% to 94% and from 23% to 28%, respectively. The performance of CFMRF was no better than could be expected by chance, yielding a sensitivity of 19% and a specificity of 85%. Applying SCORE or OST 75% of the women would have to be referred for densitometry to identify 90% of the women with T-score < or = -2.0 (or T-score < or = -2.5) in at least one region. In conclusion, our results question the utility of all the evaluated tools for screening peri- and early postmenopausal women for low BMD. However, if a decision on referral has to be made, it may be based on the simple OST rule, which performed as well as or better than any of the other tools.