• April 7 - 11, San Francisco

    2014 Joint Summits on Translational Science

2014 CRI Panels

There are four CRI Panels scheduled for CRI 2014. Full descriptions, date and time are available in the Itinerary Planner.

CRI02: Late Breaking Panel - Direct Use of EHR for Clinical Trial Data Collection Instead of a Dedicated Research EDC System

Presenters
Vojtech Huser, NIH Clinical Center; Keith Marsolo, Cincinnati Children's Hospital; Jon McKeeby, NIH Clinical Center; Brian Wells, Cleveland Clinic Research Institute

Abstract
Seamless collection of clinical trial data in the context of routine care is a significant clinical research informatics (CRI) challenge. In recent years, despite existence of dedicated research electronic data capture (EDC) systems, such as REDCap or MediData Rave, collecting clinical trial data or registry data directly within an Electronic Health Record System (EHR) is emerging as a viable platform for certain type of studies. This panel will explore challenges and applicable standards in this domain and feature 3 case studies where research data capture is done in an EHR system.

CRI05: Panel - Big Data Analytics in Clinical Research

Presenters
Lewis Frey, University of Utah School of Medicine; Leslie Lenert, Medical University of South Carolina; Scott DuVall, University of Utah School of Medicine/VA Salt Lake City Health Care System; Lisa Dahm, University of California Irvine Medical Center

Abstract
The panel will discuss improvements and issues with the use of big data methodologies for predictive analytics in clinical research. The panel is made up of four clinical informatics researchers that are involved with the development of big data systems in healthcare. Dr. Lisa Dahm is the Director of Clinical Informatics at Irvine and oversees the development and operations of Saritor Hadoop Distributed File System, a big data solution for Irvine's electronic medical records. Dr. Scott Duvall is the Associate Director for the VA's corporate data warehouse VINCI, which includes all VA patient records. Drs. Leslie Lenert and Lewis Frey are Co-PIs on a NIH funded Clinical Personalized Pragmatic Prediction of Outcomes (Clinical3PO) big data system deployed at the VA. The Clinical3PO initiative is focused on predictive analytics within the VA using similarity matching technology. Dr. Lenert will present an overview of the Clinical3PO system and its implications for clinical care. Dr. Frey will discuss the technology development and preliminary results from the near-term prediction algorithm within the Clinical3PO system. Hadoop and other big data systems provide an ecosystem that is affordable, scalable and highly available, while allowing clinical research and clinical practice to coexist in the same system.

CRI09: Panel - TRANSFoRm Digital Infrastructure: The Architecture for the Learning Healthcare System in Europe

Presenters
Vasa Curcin, Imperial College London; Theodoros Arvanitis, University of Warwick; Piotr Brodka, Wroclaw University of Technology; Derek Corrigan, Royal College of Surgeons of Ireland; Brendan Delaney, King's College London

Abstract
The Learning Healthcare System (LHCS) refers to the close coupling of clinical research and the translation of research into practice in a cycle of continuous improvement. This vision permeates multiple domains, clinical as well as technical, and its realization is dependent on establishing standardized, secure, and traceable flows of data between these domains to maximize the research and clinical benefits. This panel presents the model-driven software architecture designed in the TRANSFoRm project (www.transformproject.eu), a large EU FP7 Integrated Project to develop a digital infrastructure for the LHCS in European Primary Care. The discussion will cover various components of the system, comparing them with similar tools in USA and Europe, and analyze how our modular approach supports collaboration with related efforts.

CRI16: Panel - Architectures for Data Standardization and Interoperability in Patient Centered Outcomes Research

Presenters
James Campbell, University of Nebraska; Russ Waitman, University of Kansas; Abel Kho, Northwestern University Feinberg School of Medicine; Thomas Campion, Weill Cornell Medical College; Samuel Rosenbloom, Vanderbilt University

Abstract
The Patient Centered Outcomes Research Institute's (PCORI) Clinical Data Research Network (CDRN) initiative promises to test the reusability of electronic health records (EHR) and other data sources to support comparative effectiveness research. This effort is concurrent with a national investment in EHRs compliant with Nationwide Health Information Network (NwHIN) standards designed to develop interoperable shared data. The interoperation and utility of this data is untested by clinical research at a national scale. This panel will bring together four recently funded CDRNs who will describe the approaches to interoperability, data models, and standardization they are incorporating in their network.

CRI22: Panel - Implementation of Cloud Service vs. Locally Produced Clinical Prediction Rules to Assess Traumatic Brain Injury Risk in Children: A Multi-Center Study

Presenters
Marilyn Paterno, Partners Healthcare/Brigham and Women's Hospital; Peter Dayan, Columbia University; Eric Tham, University of Colorado/Children's Hospital Colorado; Howard Goldberg, Partners Healthcare/Brigham and Women's Hospital; Robert Grundmeier, The Children's Hospital of Philadelphia; Nathan Kuppermann, University of California Davis

Abstract
The overall goal of this multi-center study is to decrease inappropriate use of cranial CT for children with minor blunt head trauma (BHT) by creating a generalizable model to translate evidence into clinical practice. Participating sites used either a web-based, platform-independent Clinical Decision Support (CDS) Service provided by the Enterprise Clinical Rules Service (ECRS) team at Partners HealthCare System (PHS) or locally produced CDS (i.e. using the EHR's CDS rules engine) developed at a central site and exported to sites selecting the local CDS option. This panel will describe the process of creating specific, computable knowledge from evidence for use across multiple institutions. A key contribution to the field of generalizable computer decision support that we provide is our experience using the same decision support content in disparate CDS systems. Our learning goals for this panel are three-fold: [a] to understand the processes needed to provide CDS for multiple sites both within a local EHR internal rules engine and from an external, remote, cloud-based CDS service; [b] to consider the pros and cons of each approach; and [c] to understand how best to provide shareable, reusable, scalable, and maintainable CDS. We will provide initial findings from implementation from two sites in our assessment. All panelists are key participants in this research, and each brings specific expertise in his/her presentation area.