Statistical Challenges in Modeling Health Symptoms During Menopausal Transition

• Activity Number: 316
• Type: Invited
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Annette Dobson*+ and Gita Mishra
• Companies: University of Queensland and University of Queensland
• Keywords: menopausal symptoms ; longitudinal analyses ; Clusters

Abstract

Despite the growing interest in the reporting of women's health symptoms during the middle years of life, there have been few longitudinal investigations. In most studies, data on women's health symptoms are usually obtained from a checklist of 20 or so symptoms. However, how these symptoms vary over time or how they vary with biological age has not been fully explored. In this paper, we discuss the main design problems associated with the measurement of age at menopause, for example for women on hormone therapy, and the statistical challenges involved in analysing data from such longitudinal studies. We propose two methods for assessing the clustering and continuity of health symptoms during menopause transition that can provide insights into the role of menopausal age on the aetiology of these symptoms. We believe that these approaches will have a wider application in longitudinal ana