Biostatistics Seminar Series: Pang Du, PhD

Title: "Variance Change Point Detection under A Smoothly-changing Mean Trend with Application to Liver Procurement"   Abstract: The extensive literature on change point analysis can be roughly divided into the parametric and nonparametric categories. Both categories of models require a sudden change in the data distribution, either in a few parameters or in the distribution as a whole. In this paper we are concerned with the scenario that during the observation period the variance of data may make a significant jump while the mean of data changes in a smooth fashion. It is motivated from a liver procurement experiment where the surface temperature of the organ in perfusion shows a significant drop in variance while maintaining a smoothly-changing mean profile. This drop is likely due to the sudden deterioration in the viability of the organ and thus carries critical clinical information in the quality assessment of the organ. Blindly applying the existing change point analysis methods to the example can yield erratic change point estimates since the smoothly-changing mean violates their common assumption of a sudden change. In this paper, we propose a penalized weighted least squares approach that naturally integrates the variance change point detection and the smooth mean function estimation. An iterative procedure is introduced to obtain all the parameter estimates. Given the variance components, the mean function is estimated by smoothing splines as the minimizer of the penalized weighted least squares. Given the mean function, we propose a likelihood ratio test statistic for identifying the variance change point. The null distribution of the test statistic is derived. Simulations show excellent performance of the proposed method. Application to the liver procurement experiment data offers numerical support to the non-invasive way of assessing organ viability through surface temperature monitoring.