The Global Longitudinal Study of Osteoporosis in Women (GLOW) was an international study of osteoporosis in women over 55 years. The major goals of the study were to provide insights into the management of fracture risk. The data GLOW500 contains a small sub-set of the much larger data from six American centres. Subjects were followed for two years, and the outcome of interest was any fracture during follow-up (fracture=1).
- Describe graphically the distribution of age, body mass index (BMI) and smoking status in those who had a prior fracture and those who did not have a prior fracture. You should report appropriate summary statistics and justify your choice.
- (ii) Next, consider the association between age and the risk of a fracture during follow-up. Test the null hypothesis that average age is the same for those who experience a fracture during follow-up and those who do not and report relevant summary statistics. Report the probability of a fracture during follow-up by age-group. Report should include estimated probabilities and 95% confidence intervals along with a test of statistical significance.
- Based on the results in (i) and (ii) and without conducting further analysis, discuss how the association between prior and future fracture may be confounded by age. In the context of this question, what would it mean to say that age is an effect modifier? You should present your solution in report format. Graphs may be cut and pasted from Stata into your report, but tables should not. Titles should be placed above tables and below graphs.
