My name is Ruoyu Miao. I am a research associate in the Department of Obstetrics and Gynaecology at the University of Cambridge.
I did joint models for mean and variance structures as a postdoc in the Department of Mathematics at the University of Manchester (2022-2023). We proposed to model the within-subject variance of the longitudinal response and assume it’s associated with both the longitudinal and survival processes through random effects, which is motivated by the high within-subject variability in systolic blood pressure to correctly model that variability and improve the cardiovascular risk prediction.
One of my research areas is 'Joint modelling approaches for multi-outcome data'. My research focuses on two concentrations: (1) A joint latent class model of longitudinal and survival data with the time-varying membership probability; (2) A general time-varying joint latent class model with covariance modelling.
As for the joint latent class model, we aim to propose a more general JLCM to address the issue of the existence on the time-varying jumping behaviours heterogeneous subgroups and model heterogenous random covariance matrix. The primary contribution of my work is that I extended JLCMs by assuming a time-varying membership probability, allowing individuals to jump between classes across time. This can be motivated, e.g., by patients changing treatments regimes to address changes in their health status. The proposed time-varying JLCM can relax the time-invariant restrictions of JLCM and capture jumping behaviours for longitudinal markers. In my second piece of work, I proposed the general joint latent class model with both time-varying class membership probabilities and covariance modelling. The key contribution of this general JLCM are classifications are allowed to change with time while also allowing us to capture the potential effects of covariates on the association between the longitudinal and survival processes.
Now I am doing a Postdoc at the Department of Obstetrics & Gynaecology, School of Clinical Medicine. My current research in the University of Cambridge is the proteomic analysis of blood samples from the Pregnancy Outcome Prediction Study (POPS) cohorts. The aim of the study is to determine the association between maternal serum levels of metabolites and the risk of adverse pregnancy outcome.
