Health Research and Policy

Abstract

DATE: September 27, 2012
TIME: 1:15 - 3:00 pm
LOCATION: Li Ka Shing Center for Learning (LKSC)
291 Campus Dr, Room LK 209
1:15 pm - 3:00 pm
TITLE: Predictive Precedence of Organ Recovery in Injury
SPEAKER: Junhee Seok, PhD, Postdoctoral Research Scholar
Department of Statistics, Stanford

Injury is a complex disease that often induces a systemic response of the whole body system such as multiple organ dysfunctions. In intensive care, the recovery of major organ functions is essential to the overall recovery averting death. Despite its importance, investigating relations among recovery of different organs and clinical outcomes has been significantly limited by the lack of data collection and appropriate analysis methods. Throughout the effort of a large-scale multidisciplinary research consortium, we have collected the functional trajectories of major organs of 1,875 severe trauma patients from six clinical centers in the United States. With a newly developed multivariate method, here we analyzed the times to recovery events of six major organs: neurological, respiratory, cardiovascular, renal, hepatic and hematologic systems.

The analysis of organ recovery events actually falls into the category of failure time analysis or survival analysis. Analyzing the failure times of multiple non-terminal events is a keen interest in many fields; however, estimating the joint distribution of the failure times in a nonparametric way is not straightforward because some failure times are often right-censored and only known to be greater than observed follow-up times. Although the nonparametric estimation for multivariate survival times has been studied in the literature, there is no good practical solution yet. Here, we proposed a nonparametric Bayesian approach for directly estimating the density function of multivariate survival times, where the prior is constructed based on the optional Polya tree. We investigated several theoretical aspects of the procedure and derived an efficient iterative algorithm for implementing the Bayesian procedure. The empirical performance of the method was examined via extensive simulation studies.

By applying the proposed method to the analysis of organ recovery times, we found a preferred organ recovery sequence, that is the hematologic and hepatic systems recovered first, the respiratory system in the middle, then the neurological and renal systems, and finally the cardiovascular system recovered. The late recovery of complicated patients was mainly due to the late recovery of the respiratory system, and the recovery of the neurological, renal and cardiovascular systems strongly depended on the respiratory recovery.

The functional states of these two organs were adapted to early-stage predictions together with immunological states represented by gene expression of white blood cells, which resulted in a successful prediction for the final patient outcome without complications. The overall results shed lights on prospective medical practices of prioritized treatments and monitoring on organ functions.

Suggested reading:
Oakes, “Biometrika Century: Survival Analysis”, Biometrika, 2001, 88: 99-142.

Wong and Ma, “Optional Polya Tree and Bayesian Inference”, Annals of Statistics, 2010, 38: 1433-1459.

Cuschieri et al, “Benchmarking Outcomes in the Critically Injured Trauma Patient and the Effect of Implementing Standard Operating Procedures”, Annals of Surgery. 2012, 255(5):993-9.

Back to Workshop in Biostatistics -->

Stanford Medicine Resources:

Footer Links: