Developing a Learning Health System HIMSS Session ID: 226 Larry Sitka, MS, BS James Forrester, MS
Agenda Introductions The Learning Health System Vision Larry Sitka, Founder of the Lexmark Acuo VNA Putting the Learning Health System into Action Jim Forrester, University of Rochester Medical Center Closing Comments Q/A
The Learning Health System Vision
5 A Healthcare System Today the U.S. system has a healthcare focus Reactive Healthcare System The entire system is proceduralbased Procedural-based reimbursement model Vendors build products based on procedures and not patient outcomes Healthcare means the patient is already sick. Are we too late in the process? vs A Health System The world is teaching vendors we must focus on a health system Proactive/Suggestive Health System Reimbursements based on patient outcomes Departments go from revenue generating to capital costs Roadblocks to entry are being eroded across localities We still have vendor lock and vendor block Proactive/Suggestive Health System is a Learning Health System A Proactive/Suggestive Healthcare Content Management Platform for a Learning Healthcare System must exist
6 Patient s Outcome is the Center of the Universe, Not the Product in a Learning Health System vendor neutral archive enterprise content management EMR Cloud SAN NAS CAS
What it is An ONC proposed ecosystem comprised of providers and controlled by the patient. Creates new knowledge. Is supported by a wide variety of systems. Lexmark surrounds a chaotic HIT environment. Characterized by Continuous learning cycles A broad array of stakeholders Extends beyond the enterprise to external care contributors Supportive LIVE Infrastructure
The Learning Healthcare System maps well to Mature VNAs Learning Healthcare System Build upon the existing HIT system One size does not fit all Empower individuals Leverage the market Simplify Maintain modularity Consider the current environment and support multiple levels of advancement Focus on value Scalability and universal access Mature VNAs Can install under/over traditional PACS Interoperable now and into the future; Canonical data models Reduces complexity, eliminating vendor lock and vendor block Consolidates storage and improves its management Modular scalability; Adopts new technology and virtualization; Additional benefits emerge Savings helps offset investments Interoperability through certified standards
A Learning Health System and the Mature VNA The foundation of a Learning Healthcare System is a Mature VNA 3 Year Agenda 2015-2017 Send, receive, find and use a common clinical data set to improve health and healthcare quality 6 Year Agenda 2018-2020 Expand interoperable health IT and users to improve health and lower cost 10 Year Agenda 2021-2024 Achieve a nationwide learning health system 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 Ability to acquire, send, find, receive. Canonicalization of content metadata Provides a means of perceived centralization of content (both storage and application) Data ownership by the healthcare organization. Ownership of the content on behalf of the patient Canonical data mapping not just inbound but outbound. Interoperability through standards Interoperability using secure and authorized Open Image Exchange for sharing Leverage the existing HIE and private networks All built around this new patientcentered health system PIX Manger Services Today s Mature VNA Figure 10. Connecting Health and Care for the Nation: A Shared Nationwide Interoperability Roadmap Version 1.0
10 Industry Driver #1: Unprecedented Demand for Information Genomics Zettabytes BIG DATA Analytics Define outcomes Patient Required by a Learning Health System The next 1000+ concurrent users ALWAYS ONLINE Oncologist Surgeon Referring Physician Radiologist HIM Physician Digital Pathology Supporting Content Radiology/Cardiology Imaging Scanned Documents Exabytes Petabytes Terabytes
Industry Driver #2: PACS Redefined The Separation of the PACS Components (IHE Actors) into Enterprise Imaging Viewing Component Diagnostic Clinical Specialized Multiple sub specialties Workflow Component Modality worklist Physicians worklist Exam state Dictation Credentialing Physician load balancing Archiving (Content Manager) Secure storage, distribution, routing, canonical data models Migrations, PID resolution, SOA access Data security, data preservation, data separation All content location services (DICOM and other content)
Today s Environment Silos of vendor locked and blocked information EMR Custom Interfaces Cardiology/Radiology PACS Other ologies ECM PDF, JPG, TIFF, TXT,.RAW Proprietary File Access Vendor Lock? Vendor Block? Individual Departmental Storage Silos Multiple if any DR plan? PHI Exposure? $$$?
Healthcare Content Management System Radiology PACS DICOM World Cardiology PACS Image Enabled UniViewer EMR RIS/HIS/EMPI Other Content World ECM CONTENT (JPG, TIFF, PDF,.RAW, ETC.) ECM CAPTURE DICOM DICOM DICOM / WADO / QUIDO / STOW RESTful Webservices MINT Dynamic Encrypted URL via HL7 HL7 Perceptive Search XDS Web Services EMR Open Image Exchange PIX Manager (IHE) Mature VNA Enterprise Workflow Image Services Management Web Services Federation and Morphing XDS.b Reg/Rep (XDS) User Based ACL MSAD Cardiology PACS VNAMed Custom DICOM Interfaces Radiology PACS VNA Semantix DICOM Assisted HL7 Supplement-95 Vendor Neutral Archive Migration Utility SOA Service Bus Architecture DICOM Virtualization Clinical Information Lifecycle Management Store Storage Virtualization Data Lifecycle Management Mature VNA Storage Virtualization Managed Shares Secure Access Protecting PHI Archive ECM Integration (api) VNA IHE Audit VNA HA Business Continuity XDS/XDS-I ECM DICOM Verification & Extraction non-dicom Utility Support IHE ITI Database Intelligence Layer Proprietary File Access CIFS, NFS, API HTTP/REST iscsi Enterprise Content Management SAN MAIN FACILITY SCSI CAS/COS CLOUD NAS SATA TAPE DISASTER RECOVERY SITE OTHER
What s Next for Interoperability in Healthcare 21 st Century Cures Act Data Locking and Data blocking What are the consequences? Lost MU certifications for the Vendor Provider will have reduced reimbursements and ultimately complete loss. SMART on FHIR as the platform to link apps to the EMR SMART HealthIT is the interface between healthcare data and innovation. The goal of SMART is audacious and can be expressed concisely: an innovative app developer can write an app once, and expect that it will run anywhere in the health care system. Further, that one app should be readily substitutable for another. When apps are substitutable, they compete with each other which drives up quality and down price. SMART and the App Store for Health model. 1 4
Open Content Exchange Leverage Healthcare Content Management Apps to enable health information exchange between facilities Secure Open Image/Content Exchange Healthcare DURSA Governance Agreement (Data Use Reciprocal Support Agreement) Media Writer PACS Scan Content Upload Service Content Download Service Content Upload Service Content Download Service Content Upload Service Content Download Service Sending PACS Receiving PACS HCO A (Un)Affiliated HCO B HCO A (Un)Affiliated HCO B HCO A (Un)Affiliated HCO B Enterprise Image Connectivity VNA + ECM Enterprise Viewing + Image Sharing Multifunction products (MFP)
The Learning Health System Implementation at the University of Rochester Medical Center
Monthly Image Volume Provides a metric that is understandable by C-Suite execs Useful for monitoring and trending Used to trend average images per study as compared to patient outcomes. As average image count per study increases, does diagnostic accuracy increase? Does turn around time of diagnostic read improve?
1,600,000,000 Monthly Image Volume Total Images in Archive (VNA) 1,400,000,000 1,200,000,000 1,000,000,000 800,000,000 600,000,000 400,000,000 200,000,000 0 Jan-14 Apr-14 Jul-14 Oct-14 Jan-15 Apr-15 Jul-15 Oct-15 Jan-16
Image Access from EMR Provides a metric that is understandable by C-Suite execs Useful for understanding clinical workflow (referring provider) Useful for capacity planning
Foreign Image Ingestion Useful for understanding confirmation of gatekeeper functionality Useful for understanding trends and resulting implications Such as consult load on radiologist for foreign studies
Foreign Image Ingestion Foreign Film Studies 10000 9000 8000 7000 6000 5000 4000 3000 2000 Received Push to PACS Linear (Received ) Linear (Push to PACS) 1000 0 Jan-14 Feb-14 Mar-14 Apr-14 May-14 Jun-14 Jul-14 Aug-14 Sep-14 Oct-14 Nov-14 Dec-14 Jan-15 Feb-15 Mar-15 Apr-15 May-15 Jun-15 Jul-15 Aug-15 Sep-15
Foreign Image Ingestion 700,000 Enterprise Imaging Studies 600,000 500,000 400,000 300,000 200,000 100,000 0 FY11 FY12 FY13 FY14 FY15 Foreigns 25307 32577 39250 49323 57631 Rad Dx 518728 526028 525051 548342 581220
VNA Image Study Prefetch Useful for understanding customer\end user experience Useful for finding anomalous usage patterns
VNA Image Study Prefetch 600,000 Images Retrieved from Archive (VNA) to Viewer Cache 500,000 400,000 300,000 200,000 Ad Hoc Retrieval PreFetch Ortho PreFetch Rad Linear (Ad Hoc Retrieval) Linear (PreFetch Ortho) Linear (PreFetch Rad) 100,000 0
VNA Image Study Prefetch 200,000 Images Retrieved from Archive (VNA) to Viewer Cache 180,000 160,000 140,000 120,000 100,000 80,000 Ad Hoc Retrieval Linear (Ad Hoc Retrieval) 60,000 40,000 20,000 0 5/2 5/2 5/2 5/2 5/3 5/3 6/1/ 6/2/ 6/3/ 6/4/ 6/5/ 6/6/ 6/7/ 6/8/ 6/9/ 6/1 6/1 6/1 6/1 6/1 6/1 6/1 6/1 6/1 6/1 6/2 6/2 6/2 6/2 6/2
Create Historic Imaging Order Set Create historic order load from VNA image study archive meta data Imaging studies migrated in to UR Medicine VNA from foreign or internal archive Studies are modified for demographic information upon migration to match UR Medicine Epic demographics Upon migration a data extract is queried from UR Medicine VNA SQL database which includes study information such as modality type, exam code, exam date. This extract is converted to HL7 and fed in to imaging downstream systems so that they can now readily access the imaging studies from VNA. Can also be fed in to EMR as reportable order. This is a powerful tool for imaging study migration and consolidation
Image Study Viewed by Hour of Day Image study views in viewer Grouped by hour of day and day of week Average over 52 weeks Diagnostic and enterprise combined data set Useful for understanding image study viewing patters Used to plan scheduled downtimes Used to understand impact of unscheduled downtimes
Image Study Viewed by Hour of Day Image Study Views By Hour 2500 2000 1500 1000 500 0 Sun 0:00 Sun 4:00 Sun 8:00 Sun 12:00 Sun 16:00 Sun 20:00 Mon 0:00 Mon 4:00 Mon 8:00 Mon 12:00 Mon 16:00 Mon 20:00 Tues 0:00 Tues 4:00 Tues 8:00 Tues 12:00 Tues 16:00 Tues 20:00 Weds 0:00 Weds 4:00 Weds 8:00 Weds 12:00 Weds 16:00 Weds 20:00 Thurs 0:00 Thurs 4:00 Thurs 8:00 Thurs 12:00 Thurs 16:00 Thurs 20:00 Fri 0:00 Fri 4:00 Fri 8:00 Fri 12:00 Fri 16:00 Fri 20:00 Sat 0:00 Sat 4:00 Sat 8:00 Sat 12:00 Sat 16:00 Sat 20:00
Study Completed Imaging studies tech complete timestamp Grouped by hour of day and day of week Average over 52 weeks Useful for understanding imaging study acquisition patterns
Study Completed 250.0 Completed Studies By Hour 200.0 150.0 100.0 50.0 0.0 Sun 0:00 Sun 4:00 Sun 8:00 Sun 12:00 Sun 16:00 Sun 20:00 Mon 0:00 Mon 4:00 Mon 8:00 Mon 12:00 Mon 16:00 Mon 20:00 Tues 0:00 Tues 4:00 Tues 8:00 Tues 12:00 Tues 16:00 Tues 20:00 Weds 0:00 Weds 4:00 Weds 8:00 Weds 12:00 Weds 16:00 Weds 20:00 Thurs 0:00 Thurs 4:00 Thurs 8:00 Thurs 12:00 Thurs 16:00 Thurs 20:00 Fri 0:00 Fri 4:00 Fri 8:00 Fri 12:00 Fri 16:00 Fri 20:00 Sat 0:00 Sat 4:00 Sat 8:00 Sat 12:00 Sat 16:00 Sat 20:00
ED CT Data Similar to prior dataset but specific to trauma center ED CT acquisition Useful for understanding implications of planned and unplanned downtimes for ED trauma center Demonstrates\confirms known ED encounter and usage patterns Useful to plan staffing for 3D\advanced visualization processing lab
ED CT Data 400 SMH ED CT Studies 2014 by Begin Time 350 300 250 200 150 100 50 0 Sun 0:00 Sun 4:00 Sun 8:00 Sun 12:00 Sun 16:00 Sun 20:00 Mon 0:00 Mon 4:00 Mon 8:00 Mon 12:00 Mon 16:00 Mon 20:00 Tues 0:00 Tues 4:00 Tues 8:00 Tues 12:00 Tues 16:00 Tues 20:00 Weds 0:00 Weds 4:00 Weds 8:00 Weds 12:00 Weds 16:00 Weds 20:00 Thurs 0:00 Thurs 4:00 Thurs 8:00 Thurs 12:00 Thurs 16:00 Thurs 20:00 Frid 0:00 Frid 4:00 Frid 8:00 Frid 12:00 Frid 16:00 Frid 20:00 Sat 0:00 Sat 4:00 Sat 8:00 Sat 12:00 Sat 16:00 Sat 20:00
ED CT Data 2500 Total EDCT Volume 2014 Grouped by Hour of Day 2000 1500 1000 500 0 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 0 1 2 3 4 5 6
Questions Larry Sitka, MS, BS, Principle Solutions Architect, Founder Acuo VNA Lexmark Healthcare James Forrester, MS, Director of Product Management and Imaging Informatics University of Rochester Medical Center