OpenCDS: Overview, Demonstration, and Potential Implications for HeD Use Case 2 January 10, 2012 Health edecisions All-Hands Call Kensaku Kawamoto, MD, PhD Director, Knowledge Management and Mobilization, Univ. of Utah Assistant Professor, Dept. of Biomedical Informatics, Univ. of Utah Founder, OpenCDS (www.opencds.org) kensaku.kawamoto@utah.edu David Shields Senior Data Warehouse Architect, Univ. of Utah Lead Engineer, OpenCDS david.shields@utah.edu
HeD Use Case 2 CDS Guidance Requestor CDS Guidance Supplier CDS Request (patient data + context) OpenCDS CDS Guidance (guidance + service structure)
OpenCDS (www.opencds.org) Goal Facilitate widespread availability of advanced CDS capabilities through open-source, collaborative development of standards-based CDS infrastructure, tooling, and high-value content Methods Encapsulate CDS capabilities as Decision Support Services (DSSs) that can be integrated with various clinical information systems Support relevant CDS standards Leverage open-source components (e.g., JBoss Drools, Apelon Distributed Terminology System) 1.1 release freely available under Apache 2 opensource license
Featured Collaborators
OpenCDS Architectural Overview Patient data, knowledge modules to use Conclusions about patient Standard Interface: HL7 Decision Support Service (DSS) Standard (http://hssp-dss.wikispaces.com) Trigger Queries for required pt data Patient Data Sources Standard Data Models: Client Decision HL7 Virtual Support Medical Apps Record (vmr) Standard [ http://wiki.hl7.org/index.php?title= Virtual_Medical_Record_(vMR) ] Institution A OpenCDS Decision Support Service Knowledge Modules Client Decision Support Apps Queries for required pt data Patient Data Sources Institution B
CDS Services Example VMR Eval. Result Decision Support Service CCD EHR System Pt data Patient Data Sources
Support for Multiple Knowl. Rep. Approaches DSS Client Patient data Patient-specific care advice HL7 Decision Support Service Interface OpenCDS Adapters External CDS Engines/Services (e.g., SEBASTIAN, ECRS, WarfarinDosing.org) OpenCDS Knowledge Management and Execution Platforms
WarfarinDosing.org Integration DSS Client Patient data Patient-specific care advice HL7 Decision Support Service Interface WarfarinDosing.org OpenCDS Adapter WarfarinDosing.org
WarfarinDosing.org (Stand-Alone)
WarfarinDosing.org (Stand-Alone)
WarfarinDosing.org Integration via OpenCDS
Immunization Calculation Engine (ICE) Platform DSS Client Patient data Patient-specific care advice HL7 Decision Support Service Interface OpenCDS ICE Adapter ICE Knowledge Management and Execution Platform
ICE Immunization Schedule Manager
ICE Immunization Rule Manager
ICE Test Manager
OpenCDS HeD Use Case 1 Platform DSS Client Patient data Patient-specific care advice HL7 Decision Support Service Interface OpenCDS Use Case 1 Adapter HeD Use Case 1 CDS Knowledge Artifacts
OpenCDS Guvnor Platform DSS Client Patient data Patient-specific care advice HL7 Decision Support Service Interface OpenCDS Guvnor Adapter OpenCDS Guvnor Knowledge Management and Execution Platform
Demo
Integration Architectures EHR system calls the DSS Current status: some early adoption and announced plans for adoption by at least one major vendor Vendors increasingly providing custom development platforms that allow many things, including using DSS Vendors also exploring client-level configuration of DSS use HeD UC 2 will make this approach much more scalable External CDS system interfaced with EHR calls DSS HeD UC 2 will also make this approach more scalable More support for standard EHR APIs (e.g., for data read/write and order placement) desired for moving this approach forward, but out of scope for HeD UC 2 (Ref: Kawamoto et al., AMIA Proc, 2012)
Lessons Learned and Implications Lesson Potential Implication for HeD UC 2 Simplicity is key for implementability Key to interoperability is common, detailed data specifications Retrieving data is by far the most time-consuming component Data serialization/deserialization is computationally expensive Authoring tooling needs to be user friendly and intuitive Terminology management is key End-to-end validation is critical Keep approach as simple as possible Focus upcoming work on data specifications Support only sending required data Support sparse XML serialization format N/A UC 2 allows use of any authoring environment Tightly define terminology bindings Support running service as-if fixed point in time; define services in testable manner
Financial support Acknowledgements NHGRI K01 HG004645 (PI: K. Kawamoto) University of Utah Dept. of Biomedical Informatics University of Utah Information Technology Services Utah Beacon Community Subcontract (PI: Bruce Bray) VHA Office of Knowledge Based Systems (Director: Steve Brown)
Acknowledgements Numerous OpenCDS collaborators www.opencds.org Featured Collaborators From University of Utah: David Shields Reid Holbrook, MD Reed Barney Cheri Hunter Bruce Bray, MD Jennifer Garvin, PhD Gang Luo, PhD From HLN Consulting, LLC: Daryl Chertcoff Sam Nicolary Paul Sanders Mingyuan Zhang Jason Jacobs
Thank You! Kensaku Kawamoto, MD, PhD Director, Knowledge Management and Mobilization, Univ. of Utah Assistant Professor, Dept. of Biomedical Informatics, Univ. of Utah Founder, OpenCDS kensaku.kawamoto@utah.edu David Shields Senior Data Warehouse Architect, Univ. of Utah Lead Engineer, OpenCDS david.shields@utah.edu
www.opencds.org
Backup Slides (partial content of live demo)
OpenCDS Guvnor Platform Rules
OpenCDS Implement. Underlying Details
OpenCDS Guvnor Platform Decision Table
OpenCDS Guvnor Platform Flow Diagrams
OpenCDS Guvnor Platform Testing
OpenCDS Guvnor Platform Batch Testing