Developing Model-based Bone Density Reference Values for African-American Women

点击翻译按钮获取中文摘要

Project Summary / Abstract Clinical reference values are typically derived from Caucasians, with limited sample sizes and simplistic statistics. As a result, applying these reference values to racial/ethnic minorities often causes misdiagnosis. Our long-term goal is to develop innovative methodologies to produce and validate precision reference values. The reference values of bone mineral density (BMD) are proving increasingly inaccurate since a substantial number of patients who sustain fragility fractures have what the standard T-score method defines as a normal BMD value. The T-score method was originally proposed only for postmenopausal Caucasian women, and as a result there is no empirical method designed specifically to define BMD reference values for racial/ethnic minorities. Therefore, with life expectancy increasing, osteoporotic fractures in minorities including African- American women are becoming a major public health issue. The objective of this application is to develop a novel method to determine precision BMD reference values for African-American women, a group whose osteoporotic fractures have previously been under-recognized and undertreated. On the basis of preliminary data produced by the applicant, the central hypothesis is that for African-American women a new method based on “the best statistical model” of normal BMD will be a significantly better predictor of osteoporotic fracture than the T-Score method. This hypothesis will be tested by pursuing three specific aims: 1) identifying the best statistical model of normal BMD; 2) determining the model-based precision threshold of BMD; and 3) validating the precision reference values for African-American women. For Aim 1, this project will leverage multiple existing NHANES databases to develop a highly accurate statistical model of normal BMD for African- American women. For Aim 2, a precision threshold that matches an individual's characteristics will be determined by utilizing the best BMD model and cohort data containing fracture outcome. Under Aim 3, Study of Osteoporosis Fracture data will be utilized to validate the model-based method by comparing its predictive accuracy for osteoporotic fracture with traditional methods. This innovative approach will replace the traditional “one-size-fits-all” clinical reference value, fundamentally shifting current research and clinical practice paradigms from one static cutoff point for everyone to precision thresholds fitting individual characteristics. The new approach will overcome the limitations of the T-score method, thus increasing the prediction accuracy of the WHO's Fracture Risk Assessment Tool, and the accuracy of osteoporosis diagnoses in general. The approach developed herein can be extended to the development of many other types of precision reference values. In addition, this project will engage students in the research of precision medicine through the use of “big data,” thus providing them with hands-on research experience and stimulating their research interest in health disparities.