Developing Robust Risk Models for the Persistent Inflammation, Immunosuppression, and Catabolism Syndrome

Martingale Research Corporation, Plano TX

Investigators

Abstract

The Persistent Inflammation, Immunosuppression, and Catabolism Syndrome (PICS) is a newly identified medical condition involving extended stays in intensive care units that reduces quality of life for over 250,000 surgical patients in the U.S. each year, and adds nearly $100 billion annually in health-related costs. The availability of new clinical data characterizing PICS provides the opportunity to develop patient-level PICS risk models for trauma and emergency surgery research, epidemiologic, clinical diagnostic, and therapeutic studies.Such models are invaluable as decision tools that better inform practitioners, administrators, and policy makers in order to effectuate improved patient quality of life and reduced health care costs. Despite widespread use of patient-level prediction models for clinical events, patient-level multivariate PICS risk models are not currently available. Moreover, recent advances in statistics that can be applied to robust risk modeling of underlying pathologies are required, but not readily accessible to researchers. Thus, utilizing improved statistical methods for developing a PICS risk model that can reveal the etiology for the disease, not merely predict onset, would significantly help scientists understand, assess, prevent, and treat PICS. This Phase I study investigates the feasibility of applying a Best Approximating Model (BAM) method to develop PICS risk models on a NIGMS-sponsored research center dataset. The BAM method is a systematic model development approach that combines robust estimation, specification analyses, stochastic/exhaustive model search, and model validation within the single model selection/validation framework of a generalized additive model. A BAM is designed to handle common problems encountered in developing predictive and explanatory risk models including possible model misspecification, missing values, and overfitting; as well as multicollinearity, small sample size bias, and Type I error inflation due to multiple model comparisons. In this Phase I study, the BAM software functionality will be extended to support development of robust PICS risk models, followed by a series of simulation studies designed to evaluate its performance. The simulation studies will also characterize the advantages of the BAM strategy for developing a robust PICS risk model over conventional statistical methods such as stepwise regression. Feasibility study results will: 1) develop and evaluate the performance of a PICS risk model on a NIGMS-sponsored research center dataset, 2) disseminate findings to scientists studying PICS, 3) establish feasibility for Phase II PICS risk model development, validation, evaluation, and dissemination, and 4) provide the foundation for Phase III commercialization of an advanced risk modeling technology.