Info

a Number of women with no missing responses on all variables used in the predictor. Reproduced with permission of the publisher from ref. 1. Permission conveyed by the Copyright Clearance Center.

a Number of women with no missing responses on all variables used in the predictor. Reproduced with permission of the publisher from ref. 1. Permission conveyed by the Copyright Clearance Center.

The SCORE questionnaire performed well in both the development and validation cohorts. Sensitivities of 0.89 and 0.91 indicate that 89 and 91% of the women in the two cohorts with femoral neck T-scores of -2 or less were at or above the SCORE threshold of 6 and would have been appropriately referred for bone density testing. In order to achieve this high degree of detection, the lower specificities of 0.50 and 0.40 in the 2 cohorts must be accepted. Of the women in the 2 cohorts 50 to 60% with femoral neck T-scores greater than -2 would also be referred for testing because their test scores were at or above the SCORE threshold of 6.

The initial publication of the design of the SCORE questionnaire and its utility was greeted with skepticism. Prior to the development of SCORE, most attempts to use clinical risk factors to select women for bone density testing had not resulted in adequate sensitivity and specificity. The questionnaire itself appeared deceptively simple in its final form, belying the very complex statistical analysis of more than 350 variables. It was also noted that 24% of the women in the validation cohort had rheumatoid arthritis. This cast doubt on the applicability and validity of the SCORE questionnaire to the general population in which such a high percentage of rheumatoid arthritis would not be expected. Finally, the femoral neck T-scores used in the development and validation phase of SCORE were based on the manufacturer's reference databases, not the NHANES III database (3). A Lunar DXA device was used in slightly more than half the cases in both cohorts with a Hologic or Norland DXA device used in the remaining cases. It was thus initially unclear how the adoption of the NHANES III database T-scores might affect the choice of the SCORE threshold for identifying individuals with a high probability of low bone mass at the femoral neck.

Other researchers attempted to validate the SCORE questionnaire in different study populations. Cadarette et al. (4) utilized the SCORE questionnaire in a population of 398 postmenopausal women with an average age of 64.5 years. Of the 398 women, 86.7% were Caucasian. Bone density was measured at the femoral neck and PA lumbar spine with a Hologic DXA device and T-scores were calculated using the manufacturer's reference data. The original SCORE threshold of 6 was used to identify those women who had a high probability of having a T-score of -2 or less at the spine, femoral neck, or both. The sensitivity and specificity for identifying women with a low BMD at either the spine or femoral neck in this study were 0.90 and 0.32, respectively. This means that 90% of the women with a low BMD at either the PA spine or femoral neck would be referred for bone density testing. Conversely, 68% of women with a BMD not considered low at either site would also be referred for testing, but 32% would not. When the analysis

Table 8-4

Sensitivity and Specificity for Various SCORE Thresholds in Two Age Groups

Table 8-4

Sensitivity and Specificity for Various SCORE Thresholds in Two Age Groups

Group

n

SCORE cutpoint

Sensitivity

Specificity

50-59 years

Natural Arthritis Relief

Natural Arthritis Relief

Natural Arthritis Relief details a unique method of reversing Rheumatoid Arthritis Symptoms by removing numerous arthritis triggers as well as toxins using a simple 5 step natural process.

Get My Free Ebook


Post a comment