» Back to Table of Contents
The Informatics, Decision Enhancement, and Analytic Sciences (IDEAS) 2.0 Center started as a Targeted Research Enhancement Program (TREP) in 2003, progressed to a Research Enhancement Award Program (REAP) in 2008, and became a Center of Innovation (COIN) in 2013. The mission of our COIN is to implement novel interventions, promote cross-center collaboration, and engage operational partners to improve the health of Veterans. Through innovation, we strive to act as an engine for change.
From the beginning, informatics has served as the unifying theme of our Center. In Salt Lake City, interest in informatics is “infectious,” meaning that, eventually, all of our investigators catch some form of the “informatics bug.” Yet, our investigators also possess diverse types of methodological expertise, encompassing causal inference, computer simulation, natural language processing (NLP), cognitive task analysis, and ethnographic observation.
The subject of our Collaborative Research to Enhance and Advance Transformation and Excellence (CREATE), “Cognitive Support for Therapeutic-Decision Making,” effectively weaves together these interests. Additional areas of focus within our Center include antibiotic resistance and health care-associated infections, post-deployment health, and, as an emerging focus, rural health.
This article briefly highlights recent areas of investigation to underscore the approach we have taken to leverage informatics to advance scientific knowledge and VA health care. In 2009, the leaders of the VHA MRSA initiative, Drs. Rajiv Jain and Gary Roselle, sought our assistance to address the lack of availability of electronic microbiology data. Working in concert with another operational partner, Dr. Chris Nielson, we developed and validated an informatics pipeline to convert text-based microbiology reports into structured, coded data. Tables containing millions of rows of data were added to VINCI as a new data resource.1
Establishing new centralized data resources in microbiology broadly benefited research and operations. Our detailed analysis of MRSA screening tests yielded insights about the role of readmission to amplify the impact of even modest reductions in MRSA transmission.2 Electronic microbiology data has made it feasible to develop new models for surveillance using predictive analytics and to implement algorithms to automate estimation of rates of health care-associated infection.
A natural extension of this work in microbiology was to examine antibiotic prescribing practices. Using a variety of data resources, including bar coded medication administration data, we characterized variation in antibiotic use across VA inpatient and outpatient settings. Moreover, we developed novel tools to analyze and visualize population-health data. We are working closely with the Antimicrobial Stewardship Task Force and with collaborators at the VA Greater Los Angeles Healthcare System to test different implementation strategies to reduce inappropriate antibiotic use.
Our work on microbiology and medications is part of the broader effort within our center to process and analyze VA’s big data. Our research in NLP, originally supported by the Consortium for Healthcare Informatics Research (CHIR), as well as by VINCI, has engaged a large number of collaborators at other VA centers. Across a variety of clinical domains, we showed that information extracted from text data improved upon classifications based on ICD-9 diagnosis codes and other forms of structured data alone.
The experience of CHIR and VINCI demonstrated the wide applicability of NLP in health services research, including in quality measurement, clinical phenotyping, and decision support. Many of our current projects involve the processing of hundreds of thousands or even millions of documents. Several of these studies fit within our post-deployment health focus area, with engagement of operational partners, such as the National Center on Homelessness among Veterans, and the War Related Illness and Injury Study Center.
Our CREATE will lead to the development of novel systems, such as “Veterans Like Mine” that tap into the vast experience of care within VA to retrieve and display information about other patients similar to the Veteran at hand. The purpose of these systems is to facilitate the management of uncertainty, the assessment of treatment options, and the prediction of clinical outcomes. VA’s big data has the potential to advance the use of evidence to inform experience and, in turn, convert experience into new evidence.
Electronic health records and decision-support systems constitute an original focus of our Center. Spanning a decade of research, we have characterized various types of limitations of VA’s health information technologies. Just as in the private sector, VA’s systems need to be redesigned to enhance cognitive support for care that is team-based, patient-centered, and safe. Our distinctive contribution to these efforts is to develop and test innovations that are strongly guided by theory. As our partnerships in VA informatics have evolved, several of our investigators have stepped into roles as VA operational leaders, while continuing to direct or collaborate in research. The effect of these dual roles is to enhance the impact of our scientific endeavors, through input on design and evaluation of implemented programs.
- Jones, M. et al. “Identification of Methicillin-resistant Staphylococcus Aureus within the Nation’s Veterans Affairs Medical Centers Using Natural Language Processing,” BMC Medical Informatics and Decision Making 2012; 12:34.
- Jones, M. et al. “Relationships Between the Importation, Transmission, and Nosocomial Infections of Methicillin- Resistant Staphylococcus aureus: An Observational Study of 112 Veterans Affairs Medical Centers,” Clinical Infectious Diseases 2014; 58(1):32-9.