Single Source: Use of Patient Care Data in Support of Clinical Research Liora Alschuler, Landen Bain, Rebecca Kush, PhD October 2003
Two Different Worlds Patient Care World Clinical Research World
Different World Views Patient Care World -Multiple data sources and data types. -HL7 v2.x a pervasive standard. -Electronic health records assembled from multiple sources. Clinical Research World -Carefully controlled data. -Each trial s data independent. -CDISC the emerging standard. -Data flows from sites to CROs to sponsors to FDA. Electronic Health Record Patient Care World -Clinicians want to see everything they can get -Date are organized around the patient. Clinical Research World -Bio-statisticians tightly control what is gathered. - Data are organized around a trial.
This hasn t worked I d like some data, please. Electronic Health Record Sure, why not?
Shift focus to the front end Electronic Health Record Connectors -Standards: HL7 s CDA; CDISC s ODM. -Technology: XML -Single Source document approach
And unify the data capture step Electronic Health Record CDA A Single Source Document for Data Capture ODM
Single Source: how is it different? esource & electronic data capture redundant with creation of clinic note full burden lies on investigator require that information reside in EMR proprietary data formats CDA & CDISC in single-source capture trial data, merge it into clinic note (reuse) works with current technology, workflow (EMR optional destination, not required) open, non-proprietary data formats
Business Drivers Need to shorten cycle time of drug development: Drug company R & D is too high; Government pressure. Drive for personalized, prospective medicine. Patient and investigator convenience. HHS desire to link research data to physician behavior. Rationalize provider document workflow.
Obstacles Two different world views between patient care and clinical research: Bio-statisticians want perfect data; Clinicians want everything they can grab. Lack of a structured vocabulary. Privacy concerns.
Project opportunities Starbright a small clinical trial at Duke s Clinical Research Institute (DCRI) DCRI s Data Mgt and IT groups are cooperating with the trial s investigator. Goal is to capture both patient care and clinical trial data at the front end, and populate both data streams, using CDA and CDISC standards
Participants: Starbrite Trial CDISC Duke Clinic Duke Clinical Research Institute (DCRI) investigatiors Liora Alschuler, Landen Bain, Rebecca Kush, PhD, Meredith Nahm, Steve Woody, Sally Cassells, Richard Low, John Madden, MD technology partners, Microsoft (primary), Arbortext, Topsail, Sentillion.
CDA & CDISC in Starbrite Trial Manual creation and re-entry of CRF HIS validation lab, ADT, meds, source documents LIS display manual entry to CRF re-key CRF db CLINIC CRO lab, ADT, meds, source documents HIS LIS display dictate chart note Current processes (dual source) Redundant creation of chart note
CDA & CDISC in Starbrite Trial Merged workflow: clinical trial data captured in initial CDA, re-used in chart note HIS validation lab, ADT, meds, source documents LIS display CDA(1) CDA(2) complete chart note ODM ecrf db Proposed processes (single source) CLINIC CRO
CDA & CDISC in Starbrite Trial features: Single Source contributes to patient chart, not the reverse, optimizes clinical workflow no requirement to create/extract from EMR fewer privacy and regulatory issues can be driven from electronic protocol uses HL7 CDA and CDISC ODM
Duke Clinical Document Repository Clinic data input Sentillion Microsoft InfoPath XML/Office 2003 CDA Starbrite CSW/ Document management TopSail Integration Inbox Microsoft Word XML/Office 2003 Inbox transcription workstation Arbortext/ Documentum Documentum Transformation Publisher Starbrite Clinical Document Management Initial Architecture to DCRI as ODM.XML Clinic workstation: Browser and Starbrite 1) Create new record for each visit: contains patient data, visit data, templated information 2) Trial data input: nurse, incremental, realtime data integrity checks 3) Trial data completion: physician, real-time data integrity checks. First CDA complete, trial data (ecrf) extracted, transformed to ODM. 4) Chart data creation: physician: prepopulate chart CDA using trial data, cut & paste from other records 5) complete note using voice interface to transcription system 6) review, sign and export to repository
Single-source: General solution Clinical Document Repository LIS, RIS, HIS, EHR Clinical to sponsor as Data ODM.XML Repository CDA Diverse input devices/applications Clinical input XML/CDA compliant transcription transcription workstation Clinical Document management Clinic workstation: Browser and Composition Engine 1) Create new record for each visit: contains patient data, visit data, templated information 2) Trial data and clinical data input: nurse, incremental, real-time data integrity checks; input through diverse charting and forms applications (all input that is known to be required) 3) Trial data and clinical data completion: physician, real-time data integrity checks; input through real-time query, retrieval from document and data archives, (LIS, RIS, HIS, EHR); transcription 4) Chart note and trial data completion: review & sign; send note to repository, send trial data to sponsor.
CDA & CDISC in Starbrite Trial System features/requirements: multi-stage, incremental document creation optimal re-use (minimal redundant entry) minimal change to current workflow create both ODM-compliant XML for trials and CDA-compliant XML for clinical records, mapping between ODM.xml and CDA.xml low cost rapid development optimal use of technology partners, off-the-shelf technology
CDA & CDISC in Starbrite Trial Status, October, 2003: preliminary technical design and information analysis complete information design, in progress: Liora Alschuler, Sally Cassells (Lincoln Technologies, CDISC) completing data CDA/ODM data mapping technical implementation: kick-off meeting October 14, Durham, NC (attending: Microsoft, Arbortext, Topsail, DCRI) project pilot in place early Q1, 2004
Single-source: beyond Starbrite Next steps: Design and prototype a general solution Expand scope: integrate with standards-based electronic protocol Implement at one or more additional sites Establish base of industry partnerships to sustain second stage design, testing, implementation (Ringside Seat).