A joint modeling approach to study the association between subject-level longitudinal marker variabilities and repeated outcomes

• DOI: 10.48550/arxiv.2309.08000
• Publication Date: 2023-01-01
• Authors: Irena Chen; Zhenke Wu; Sioḃán D. Harlow; Michelle M. Hood; Carrie Karvonen‐G...; Michael R. Elliott
• Keywords: Longitudinal Study; Affect; Bone Health

Women are at increased risk of bone loss during the menopausal transition; in fact, nearly 50\% women's lifetime occurs this time. The longitudinal relationships between estradiol (E2) and follicle-stimulating hormone (FSH), two hormones that change have characteristic changes transition, health outcomes complex. However, addition to level rate E2 FSH, variability these across transition may be an important predictor health, but question has yet well explored. We introduce a joint model characterizes individual mean trajectories residual variances links trajectories. In our application, we found higher FSH was associated with declines mineral density (BMD) before menopause, association moderated over time after transition. Additionally, E2, not variability, slower decreases also include simulation study shows naive two-stage methods often fail propagate uncertainty individual-level variance estimates, resulting estimation bias invalid interval coverage