PFI:AIR - TT: A Clinical Predictive Model Based Smart Decision Support System for Congestive Obstructive Pulmonary Disease (COPD) related Re-hospitalization

Florida Atlantic University, Boca Raton FL

Investigators

Abstract

This PFI: AIR Technology Translation project focuses on translating smart predictive modeling technologies to develop a Clinical Decision Support System (CDSS) to fill the need for reducing re-hospitalization for patients with Congestive Obstructive Pulmonary Disease (COPD). COPD affects almost 24 million people in the U.S. and is the 3rd leading cause of death. The CDSS is important because it provides better quality of care for patients while providing significant costs savings related to hospital readmissions. The project will result in a prototype of a smart CDSS, which will enable practitioners to better understand if a COPD patient is expected to be a candidate for hospital readmission within 30 days of discharge. This cloud-based CDSS will have the following unique features: ability to utilize patient information available either in electronic or plain text format, novel algorithms derived from big data analytics, natural language processing and machine learning and ability to provide alerts to healthcare professionals about high readmission risk patients. These features provide the following advantages: improving the efficiency of the care process through system accessibility, effectively utilizing and integrating relevant patient information from various sources, reducing total cost of care through early and preventative intervention, and improving the quality of care and quality of life of the patients. Further, this CDSS will help reduce the variability in care so that hospitals with historically lower performance can benefit from best practices. This project addresses the following technology gaps as it translates from research discovery toward commercial application in the area of clinical decision support systems. The predominant technical gaps are a lack of fusion of structured and unstructured data for the development of predictive models, a limited use of natural language processing, and a significant lack of integration of local and global clinical information. This project provides a comprehensive platform to address these gaps by integration the structured local and global clinical data available in either CCD, CCR or CCDA format with the unstructured clinical data such as physician notes and laboratory reports. Further the system utilizes UIMA based cTAKES natural language processing sub-system that is developed for COPD to analyze this unstructured data. The predictive capabilities of CDSS are based on multiple leading-edge technologies including data mining, machine learning, natural language processing and data visualization techniques. In addition, graduate students involved in this project will have opportunities to learn the interdisciplinary domain of medical information systems and their real-world applications. In addition, these students will be mentored in technology transfer and commercialization activities. Larkin Community Hospital (largest D.O. program in the U.S.) and Humana (the largest health plan in Florida) will provide clinical data and expertise to the project. This project could be effectively deployed at hospitals, healthplans, accountable care organizations and managed care organizations for avoiding hospital readmissions related to COPD thereby, providing significant savings related to hospital readmissions while improving quality of care.