Detangling the Effect Between Rate of Change and Within-Subject Variability in Longitudinal Risk Factors and Associations with a Binary Health Outcome

• Activity Number: 335
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Mary Sammel*+
• Companies: University of Pennsylvania
• Keywords: longitudinal data ; variability ; cohort study

Abstract

To evaluate the effect of longitudinally measured risk factors on the subsequent development of disease, often summary measures of these factors are calculated to capture features of the risk profile. These methods consider correlations among repeated measurements on subjects as a nuisance. Using an example of hormone profile changes in the menopausal transition, we demonstrate that residual variability in subject measures, in addition to risk factor profiles, may also be important in predicting future health outcomes of interest.

We explore joint models allowing us to structure within- and between-subject variability from longitudinal studies, and combine variance structures with mean structures such as mean longitudinal profiles to better understand the relationship between longitudinally measured risk factors and health outcomes.

In the context of a 14 year longitudinal cohort study of ovarian aging among women approaching menopause, we hypothesized that fluctuations in hormone levels, rather than absolute levels, predicted menopausal symptoms. Increased efficiency in estimating associations of interest are demonstrated over a simplified 2-stage estimation approach.