Prerequisites: SPHBS 818 AND (CASMA 581 or 582). - The course will cover advanced topics in survival analysis, giving an overview of important analysis techniques that are students are likely to encounter in public health research with time-to-event data. The aim of the course will be to build a strong understanding of the theoretical foundations of the methods, while also connecting the theoretical methods to the real-world public health problems that require them. This combination will allow the students to move forward with both theoretical dissertation work in survival analysis and collaborative work requiring complex survival analysis applications in public health problems with longitudinal data. After taking the course, students will have: - An understanding of the theoretical concepts needed to understand and contribute to the survival analysis methodological literature, and the open areas of research in the field. - An ability to appropriately identify the issues that arise with time-to-event or survival data in public health research, and a toolkit for applying the appropriate methods to address them. The topics will include: - Counting processes and martingales: counting processes and martingale theory are a way to frame events occurring over the course of time and are commonly used tools to develop statistical methods for time-to-event data, and to prove their statistical properties. This is an important topic for students interested in including any survival analysis in their dissertation work to be able to understand and contribute to the literature. We will revisit some topics from prior survival courses using counting processes (ex. Logrank test, Cox proportional hazards model). - Joint modeling: joint modeling is used to assess how longitudinal covariates (common in public health research) affect a time to event outcome. - Restricted mean survival: An approach to summarizing survival data that is used as an alternative to the more commonly used proportional hazards model, and avoids some of its pitfalls. - Pseudo observations: this method transforms survival data with censoring to a dataset of pseudo observations that are not censored allowing for the use of traditional statistics methods outside of survival analysis. - Causal inference: causal inference with survival outcomes has unique challenges, including the interpretation of common summary measures like hazard ratios as causal quantities. - Recurrent Events: most of the methods that students have been exposed to will be for events that occur once (e.g. Death, hospital discharge). Recurrent event methods move a step further into methods that occur repeatedly (e.g. repeat infections, tumor regressions, repeated interactions with the healthcare system).

Note that this information may change at any time. Please visit the MyBU Student Portal for the most up-to-date course information.