ACTSI-supported Analytics Information Warehouse Recognized Internationally

ACTSI’s Biomedical Informatics Program (BIP) and interim director Andrew Post, MD, PhD, work closely with Emory Healthcare to provide clinical informatics expertise in information systems and quality improvement projects, particularly in the area of analytics and information system support for hospital discharge planning. The development of the Analytics Information Warehouse (AIW), a clinical data warehousing project, also featured in the Journal of Biomedical Informatics, was recently selected as one of the best “Big Data – Smart Health Strategies” papers in the 2014 yearbook of the International Medical Informatics Association (IMIA).

AIW aims to develop predictive tools for readmissions and other adverse outcomes and complications in targeted patient populations including heart failure and transplant. The tool was implemented and deployed in the local community, utilizing it to identify drivers of hospital readmission within 30-days. The project examined the association of over 100 derived variables representing disease and co-morbidities with readmissions in five years of data, concluding that a platform for detecting and managing such “clinical phenotypes” could accelerate quality improvement and comparative effectiveness studies. “The AIW is a key part of the BIP’s aim to create centralized and easy access to the over 7.6 million patient records in Emory’s clinical data warehouse, and advanced analytics for using that data to enhance care and accelerate research,” said Dr. Post.

IMIA Yearbook of Medical Informatics is a non-profit project aiming to advance the field of medical informatics, as well as to stimulate dissemination and exchange of information between researchers and the health and biomedical informatics community at large. Schattaer Publishing Company, located in Stuggart, Germany, formally recognized the AIW, which leverages innovative artificial intelligence-based technologies to restructure up to tens of millions of electronic patient records into a form far better suited to analytics, research, and comparison across institutions and populations.