Clinical Decision Support
Consortium
Agenda
1. Welcome, introductions and study overview (Blackford Middleton, subcontract PI, 10 min)
2. Recommendations and site visits (Dean Sittig, Co-investigator and KMLA and Recommendations Team Lead, 10 min)
3. CDSC customers’ experience (20 min)
• PHS (Adam Wright, PI and Demo Team Lead)
• Regenstrief Institute (Brian Dixon, Research Scientist) • NextGen (Sarah Corley, MD, CMO)
4. Knowledge Layers, Knowledge Authoring Tool and Health eDecisions (Aziz Boxwala, Co-investigator and KTS Team Lead, 10 min)
5. Wrapping AHRQ contract and CDSC V2 (Lana Tsurikova, Co-investigator, RPM and Research Management Team Lead, 15 min)
CDS Consortium
Study Overview
Blackford Middleton, MD, MPH, MSc
Subcontract Principal Investigator
Chief Informatics Officer and Professor of Biomedical Informatics, and of Medicine (with tenure) at Vanderbilt University
CDSC Overview
Clinical Decision Support Consortium (CDSC)
• Base Year One and Two: March 2008 – June 2010 • Optional Year One: July 2010 – July 2011
• Optional Year Two: July 2011 – July 2012 • Optional Year Three: July 2012 – July 2013
Participating Organizations:
• Started from 11 entities
• Currently includes 31 organizations – 8 healthcare institutions
– 9 academic institutions – 14 vendors
CDSC Goal and Significance
• Goal: To assess, define, demonstrate, and evaluate best practices for knowledge management and clinical decision support in healthcare
information technology at scale – across multiple ambulatory care settings and EHR technology platforms.
• Significance: The CDS Consortium will carry out a variety of activities to improve knowledge about decision support, with the ultimate goal of
supporting and enabling widespread sharing and adoption of clinical decision support.
1. Knowledge Management Life Cycle
2. Knowledge Specification
3. Knowledge Portal and Repository
4. CDS Public Services and Content
5. Evaluation Process for each CDS Assessment and Research Area 6. Dissemination Process for each Assessment and Research Area
Timelines
CDSC Headlines Preview
• CDS Web Services work at Scale – across multiple sites and multiple EMRs
• An Enhanced CCD can serve as the Patient Data Payload (with some limitations)
• Knowledge Artifacts Can Be Collaboratively Authored and Shared Across Diverse Care Delivery Organizations in an Open Knowledge Repository
CDSC Next Steps: v 2.0
• Future research and development roadmap to focus on extensions to patient data payload, additional content areas, and generalizing the approach to create “CDS Ecosystem”
– Additional content areas (MU, Pharmacogenomics, care coordination, chronic care management)
– Additional functional areas (Order sets, infobuttons, documentation templates)
• Individual sites developing rules services (for $)
• The CDSC collectively exploring v 2.0 options (more to follow)
Recommendations Summary
Dean Sittig, PhD
Recommendations for HIT Vendors
• Expand use of CCD standard
• Continue support for Gov’t-approved controlled clinical vocabularies
• Continue to work with and support the on-going HL7 knowledge representation initiatives
• Support the HL7 Infobutton standard • Implement standard data triggers
• Provide appropriate insertion points in the clinical workflow for CDS interventions to be delivered
Comparison of Clinical KM Capabilities
• Commercially-available and leading
internally-developed electronic health records.
• Qualitative research program:
– barriers and facilitators to successful adoption and use – Reviewed teaching facilities with long-standing EHR R&D
programs
• Are commercially-available EHRs capable
– clinical knowledge management features, functions, tools, and techniques
– required to deliver and maintain the CDS interventions – required to support the recently defined “meaningful use”
Comparison of Clinical KM Capabilities (cont.)
• 17-question survey about the vendor’s EHR,
CDS-related system tools and capabilities that each vendor provides, and clinical content.
• Majority of the systems were capable of performing
almost all of the key knowledge management functions we identified
• The transformation of the healthcare enterprise is
achievable using commercially-available, state-of-the-art EHRs.
BMC Med Inform Decis Mak. 2011 Feb 17;11:13. doi: 10.1186/1472-6947-11-13.
J Am Med Inform Assoc. 2012 Nov-Dec;19(6):980-7. doi: 10.1136/amiajnl-2011-000705. J Am Med Inform Assoc. 2011 May 1;18(3):232-42. doi: 10.1136/amiajnl-2011-000113. BMC Med Inform Decis Mak. 2012 Feb 14;12:6. doi: 10.1186/1472-6947-12-6.
Site Visit to Vendors and Vendors’
Customers Implementing
CDSC Services
Dean Sittig, PhD
We First Visited CDS Content Vendors
• Zynx Health, First Databank, UpToDate
• Focus was on CDS in general
• Big themes were
– We are in this together [with clinical
organizations and the EHR industry] so need
to work together
– We are like Switzerland [we do not practice
medicine]
Ash JS, et al. Studying the vendor perspective on clinical decision support.
Virtual Site Visits with EHR Vendors
• NextGen, GE and UMDNJ sites
• Focus was on service oriented architecture for CDS
• Big themes were
– SOA is the future
– The challenges are great, including lack of
interest in CDS among customers and standards
issues
Study Updates
and Demo Results
Adam Wright, PhD
ECRS Rule Authoring Input (CCD) SMArt VMR Open EHR / Recc Translations Normalization
Partners Trial Results
• Trial running since May 2010
• Currently active in 2 out of the original 4 clinics
• Calls have been consistently high
• Service was not running between Dec 2010 and
May 2011
• Ongoing advanced analysis of the data
– Clinical & Execution performance
BRIAN E. DIXON, MPA, PHD, FHIMSS
ASSISTANT PROFESSOR OF HEALTH INFORMATICS, IUPUI SCHOOL OF INFORMATICS AND COMPUTING
RESEARCH SCIENTIST, REGENSTRIEF INSTITUTE
RESEARCH SCIENTIST, CENTER FOR IMPLEMENTING EVIDENCE-BASED PRACTICE, DEPARTMENT OF VETERANS AFFAIRS, HEALTH SERVICES
RESEARCH & DEVELOPMENT SERVICE
Regenstrief Institute
CDSC Experience
Two Independent Phases
y
Phase 1 – Limited Pilot
{ July – December 2011{ 3 primary care physicians
{ Display of reminders in general “inbox”
y
Phase 2 – Expanded Pilot
{ June – December 2012{ 19 primary care physicians (all docs at 2 clinics) { Display of reminders in CPOE module of EHR
Lessons Learned
y
CDSC Service is analogous to homegrown reminders
{ Recent analysis found strong correlation; article for reviewy
Technical integration into EHR straightforward
{ Not “easy” but manageable{ Works better with SOA/modular CPOE
y
Challenges remaining
{ Terminology mappingImplementing CDSC Web
Service –NextGen and
WVPHA
Sarah Corley, MD, FACP, FHIMSS
Background
• NextGen EHR/CDSC web service
integration completed in 2012
• Client site testing with test patients
completed and ready to move to
production
• Project was taken on as a proof of concept
project from NextGen’s perspective
Initial Challenges
• Legal agreements
• NextGen had to perform mapping from
ICD-9 to SNOMED codes and NDC to
RxNorm initially
• Some diagnoses were interpreted narrowly
for CDSC
• Eclampsia does not represent a
pregnant patient
• Diabetes in pueperium, baby delivered
does not represent diabetes
Initial Challenges
• Allergies had to be mapped from UNII to
RxNorm
• Structured PE findings had to be codified
– Diabetic foot exam
• The pregnancy information was in a CCD
dedicated section rather than a subsection
of the problem list
• The patient data needed to be
de-identified
Workflow Challenges
• CDS in NextGen EHR is comprehensive &
actionable
• How to best display CDSC recommendations
within workflow?
• Who should see recommendations
• How to pass requests efficiently
Future Development
• Imported recommendations need to be
actionable
• Duplicate recommendations need to be
stripped if they are already in EHR
• More high value CDS needs to be
provided
– Radiology appropriateness indicators
– Cardiology appropriateness indicators
NextGen EHR/CDSC Web
Service Integration Details
Schematic of NextGen/CDSC
Integration
NextGen template is populated by user Interface calls CDSC server with NextGen data CDSC data returned Stored procedure is calledSchematic of NextGen/CDSC
Integration
NextGen template is loaded Check if CDSC data is present Display CDSC button on template CDSC button hidden No YesHarmonization of Standards
• MU2 set a broader scope of vocabulary
standards
• The cCDA has expanded standardized
content and should be used
• Transport and display of content should be
standardized
– e.g. Direct for transport
– E.g. wrapped CDA for content coming back
• These will reduce barriers to wider vendor
participation
Actionable CDS
• Consider using MU requirements for
reconciliation as a tool to import
recommendations for medications now
with goal for importing lab tests and
diagnostic studies in the future as
structured data so it can be imported into
EHR and ordered without transcription.
NextGen Next Steps
• Move to production server
– Barriers have been interoperability staff
constraints due to MU 2
Questions
Knowledge Layers,
Knowledge Authoring Tool
and Health eDecisions
Aziz A. Boxwala, MD, PhD
on behalf of the
Overview
• CDSC Knowledge Layers
• CDSC Knowledge Authoring Tool
• Health eDecisions update
Knowledge Representation Approach
• Goals
– Rapid translation of evidence into CDS knowledge
– Implementable in different settings and using different CDS tools and technologies
• Multilayered knowledge representation
framework
– Increasing structure and refinement in
successive layers
Multilayered Framework
Published Guideline Semi-structured Recommendation StructuredRecommendation Executable Rules
Order Sets in CPOE system
Narrative Guideline
Screening for High Blood Pressure
Reaffirmation Recommendation Statement U.S. Preventive Services Task Force (USPSTF)
The U.S. Preventive Services Task Force (USPSTF) recommends screening for high blood pressure in adults aged 18 and older. (This is a grade "A" recommendation)
Narrative Guideline
Screening for High Blood Pressure
Reaffirmation Recommendation Statement U.S. Preventive Services Task Force (USPSTF)
The U.S. Preventive Services Task Force (USPSTF) recommends screening for high blood pressure in adults aged 18 and older. (This is a grade "A" recommendation)
Semi-Structured Recommendation
Meta data
Title: Screening for High Blood Pressure Reaffirmation Recommendation Statement
Developer: U.S. Preventive Services Task Force (USPSTF)
Strength of recommendation: Grade A
Clinical Scenario:
Patient age ≥18 years
Blood pressure not obtained in the last year
Clinical Action:
Obtain and record blood pressure
Semi-Structured Recommendation
Meta data
Title: Screening for High Blood Pressure Reaffirmation Recommendation Statement
Developer: U.S. Preventive Services Task Force (USPSTF)
Strength of recommendation: Grade A
Clinical Scenario:
Patient age ≥18 years
Blood pressure not obtained in the last year
Clinical Action:
Obtain and record blood pressure
Structured Recommendation
Meta data
Title: Screening for High Blood Pressure
Developer: CDS Consortium
Derived from: USPSTF BP Screening Semistructured Rec.
Applicable Scenario
Data Mapping: BPRecordedInLastYear: Observation = VitalSign-> select(code.equals(BPLoincCode) and vsDataTime.within(12, months))
Logical Condition: BPRecordedInLastYear->notEmpty()
Recommended Action: VitalSign(code: BPLoincCode)
Structured Recommendation
Meta data
Title: Screening for High Blood Pressure
Developer: CDS Consortium
Derived from: USPSTF BP Screening Semistructured Rec.
Applicable Scenario
Data Mapping: BPRecordedInLastYear: Observation = VitalSign-> select(code.equals(BPLoincCode) and vsDataTime.within(12, months))
Logical Condition: BPRecordedInLastYear->notEmpty()
Recommended Action: VitalSign(code: BPLoincCode)
Arden Syntax Rule
knowledge evoke: …
data:
BPRecordedInLastYear := read last{table=‘RES’, code=‘12345-0’} PCPemail := read {…};
Adult := …;
logic:
if (adult is false) then conclude false;
if (BPRecordInLastYear is null) then conclude true;
action:
Write ‘Patient has not had a blood pressure screening in the last year’ at PCPemail;
Arden Syntax Rule
knowledge evoke: …
data:
BPRecordedInLastYear := read last{table=‘RES’, code=‘12345-0’} PCPemail := read {…};
Adult := …;
logic:
if (adult is false) then conclude false;
if (BPRecordInLastYear is null) then conclude true;
action:
Write ‘Patient has not had a blood pressure screening in the last year’ at PCPemail;
Preliminary Assessment
• Survey of 19 CDS experts from Partners,
Kaiser, VA, Regenstrief
• Assessed impact of layered model on five
dimensions of GLIA: decidability,
executability, presentation, flexibility, and
computability
• The results suggest that structured actions
are more implementable than semi-structured
ones. This effect was not seen for clinical
Knowledge Document Structure (L3)
CDS Knowledge Document
CDS Knowledge Document
Knowledge Module: Reminder Rule
Knowledge Module: Reminder Rule
Knowledge Module: Order Set
Knowledge Module: Order Set
Knowledge Module: Documentation Template
Knowledge Module: Documentation Template
Knowledge Module: Info Button
Knowledge Module: Info Button
Knowledge Module Knowledge Module Action Action Behavior Behavior Presentation Presentation Metadata Metadata Data Data
KAT Status
• Currently deployed on the web in test
mode
– If you would like to check it out, please
contact Lana Tsurikova or me
• In-progress
– Terminology search using BioPortal
• Planned
Health eDecisions
• S&I Framework project
• To identify, define and harmonize standards that facilitate the emergence of systems and services whereby shareable CDS interventions can be
implemented via:
– Standards to structure medical knowledge in a
shareable and executable format for use in CDS, and
(Use Case 1 – CDS Artifact Sharing)
– Standards that define how a system can interact with and utilize an electronic interface that provides
helpful, actionable clinical guidance (Use Case
Status of Health eDecisions
• Created a knowledge artifact schema (in part based on L3) that has been applied to
– Event-condition-action rules – Order sets
– Documentation templates
• Balloted as HL7 DSTU in Jan 2013 – Passed ballot
– Significant number of comments related to alignment with HQMF
– Updating the ballot materials for publication as DSTU – Pilot projects started
Thanks
• CDSC team members
• Agency for Healthcare Research and
Quality
• Health eDecisions team members and
community
Wrapping up AHRQ Contract
CDSC Chapter 2
Lana Tsurikova, MSc, MA
AHRQ Asked Us
Evaluation
Evaluation DisseminationDissemination
Demonstration
Demonstration
Implementation
8 clinical sites 5 EHR systems
Accomplishments
CDS Services
WVP Health Authority (NextGen), Salem, OR UMDNJ (GE) UMDNJ (GE) Newark, NJ Newark, NJ Cincinnati Children Cincinnati Children’’ss Nationwide Children Nationwide Children’’ss Ohio Ohio NYP NY PHS ChildrenChildren’’s Hospitals Hospital Colorado
Colorado Kaiser Roseville
UC Davis
Kaiser Sacramento Kaiser San Rafael Kaiser San Francisco California Wishard Hospital Wishard Hospital Indianapolis, IN Indianapolis, IN 1.7M CDS transactions 240 users
Accomplishments
Knowledge Management
11 clinical rules 11 clinical rules 50+ classification rules 50+ classification rules375 immunization schedule rules
Accomplishments
Accomplishments
Dissemination
Dissemination
24 published papers
24 published papers
16 papers in progress
16 papers in progress
11 sets of recommendations
11 sets of recommendations
CDS Grand Challenges
Manage large clinical knowledge databases
Create an internet-accessible, clinical decision support repository
Create an architecture for sharing executable CDS modules Disseminate best practices
Prioritize CDS content development and implementation
Additional Challenges
• Lack or ambiguity of standards
• Terminology alignment
Additional Products
Clinical Clinical Governance Governance Committee Committee Legal Framework Legal Framework 2 Years Hongsermeier et al.,AMIA Annu Symp Proc. 2011
Turechek et al.,
What It Took
$6.5M AHRQ Contract # HHSA290200810010 In-kind Contribution Tools and Time 90+ Researchers and CollaboratorsClinical Outcomes
• CDS services perform well
• Sites aim to increase participation by
adding clinics or clinicians
• SOA-based approach to CDS is feasible
The CDSC Influence -
Standards
• The Health eDecisions (HeD) Knowledge Artifact
Schema was largely based in large part on the
CDSC L3 artifact; i.e., the approach of using one
schema to express different types of CDS artifacts
such as rules, order sets, and documentation
templates
• HeD also uses concepts from L3 such as behavior,
action groups and actions, and various elements
from the metadata
Wrapping up CDSC V1
Wrapping up CDSC V1
•
7/8/2013
•
Final report
•
•
Complete
Complete
existing and new
existing and new
demonstrations
demonstrations
•
•
Continue work
Continue work
on publications
Reflections
Reflections
•
Leadership
•
Pre-competitive R&D
CDSC Chapter 2 – What’s next
•
Clinical content
•
Standards
•
Integrations
•
Meaningful Use Stage 2
Future Members
1. Healthcare service providers
2. EHR and content vendors
3. Insurance companies
4. HIT community, guidelines developers,
specialty societies
CDSC Chapter 2 – How
• PHS CDS Lab
• Academic-Industrial Collaborative
• CDS Institute or CDS National Center for
Excellence
Acknowledgements
Principal Investigator (PI): Adam Wright, PHD (2/2013-7/2013)
Subcontract PI: Blackford Middleton, MD, MPH, MSc (3/2008 – 2/2013)
CDSC Team Leads:
Research Management Team: Lana Tsurikova, MSc, MA
KMLA/Recommendations: Dean F. Sittig, PhD
Knowledge Translation and Specification: Aziz Boxwala, PhD
KM Portal: Tonya Hongsermeier, MD, MBA
CDS Services: Howard Goldberg, MD
CDS Demonstrations: Adam Wright, PhD
CDS Dashboards: Jonathan Einbinder, MD
Evaluation: David Bates, MD, MSc
Thank you!
Lana Tsurikova, MSc, MA rtsurikova@partners.org
www.partners.org/cird/cdsc