Our overall goal is to predict an individual woman's risk of developing clinically significant post-menopausal osteoporosis, using data from longitudinal studies. Three steps will be taken to develop the prediction formulas: 1) identifying early predictors of rate of bone loss, 2) using an individual's past bone mass to predict the future, and 3) using the predicted bone mass in estimating the risk of fractures. We will also compare the rates of bone loss at the appendicular and axial skeleton. In addition, we will develop and implement a quality control system for repeated bone measurements. To achieve the goals, statistical procedures will be developed to analyze repeated measurements made at irregular time points. The new methods will be applied to the data collected from ongoing and proposed longitudinal studies. Some of the new procedures will involve empirical Bayes regression. Other techniques we will use include survival analysis, and piecewise regression. The long-term objective is to identify individuals who are at high risk of developing symotomatic osteoporosis so that they can be treated prophylactically.
Showing the most recent 10 out of 22 publications
