Permanent URI for this collection
Browse
Recent Submissions
Publication Statistical Methods for Analyzing Recurrence of Cardiovascular Events(2024-09) Miu, BingCardiovascular disease (CVD) is the leading cause of morbidity and mortality worldwide. Conventional research often focuses on the time to the first CVD event, ignoring subsequent events crucial for fully understanding the disease burden. Accounting for recurrent events is important for studying chronic diseases like CVD. Despite recent advancements, analyzing recurrent CVD events poses several challenges due to their multiple types, varying severities, and diverse associated risk factors. Our work addresses these challenges by proposing new statistical methods and employing advanced approaches to study CVD events using data from the UK Biobank. In Chapter 1, we evaluate the impact of physical activity (PA) on recurrent CVD events using a semiparametric proportional mean model. To comprehensively analyze recurrent CVD events of multiple types and severities, we suggest adopting a weighted composite endpoint that includes all recurrent CVD and terminal events. Additionally, we apply an isotemporal substitution model to further investigate the effects of different daily activity intensities on these outcomes. In Chapter 2, we extend the semiparametric proportional mean model introduced in Chapter 1 and propose a dynamic regression model to analyze the time-varying effects of PA on the weighted composite endpoint of recurrent CVD hospitalizations and deaths. In Chapter 3, we conduct a mediation analysis to study the influence of early life experiences, such as childhood maltreatment, on recurrent CVD events. Our aim is to gain deeper insights into the dynamics of recurrent CVD events and facilitate the development of more effective prevention strategies.