Transcript Presentation - Association for Pathology Informatics

Caisis 4.0: Re-Designing the
Data Supply Chain
Paul Fearn, MBA
Memorial Sloan-Kettering Cancer Center
APIII – Sep 10, 2007
Caisis Project Goals






Integrate research and clinical data management
activities and systems to improve quality/efficiency
Optimize data format and organization for processing
by both humans and computers
Usability - “To be widely accepted by practicing
clinicians, computerized support systems for decision
making must be integrated into the clinical workflow.
They must present the right information, in the right
format, at the right time, without requiring special
effort. In other words, they cannot reduce clinical
productivity” – Brent C. James, NEJM 2001
Facilitate collaboration through widespread adoption
of an open source system (adopted by 15 sites in
four countries, data for over 165,000 patients)
Develop economies of experience, scale and scope
Do better science! (reproducible results)
Supported by National Cancer Institute grant R01-CA119947
2
Caisis 4.0 Technology/Architecture






Web-based (and cross-browser compatible)
Microsoft SQL Server, ASP.NET, C# platform
No special toolkits, frameworks or proprietary
modules needed beyond .NET platform
Open source license (GPL) to facilitate
innovation and collaboration with other sites
XML/metadata-driven user interface
Designed to include new modules and plug-ins
3
Caisis 4.0 User Interface
4
Data Supply Chain Concepts

Data/information - HPI, billing and diagnosis codes,
annotation for specimens, medical record, research
datasets, tumor registry reports, adverse event reports
Consumers – patients, clinicians, investigators,
statisticians, medical records, billing
 Suppliers/sources – patients, physicians, institutions,
departments, systems, “silos”, other s (eg SSDI)
 Processing/activities – physician, data manager,
investigator, clinical and research operations
 Distribution – manual data entry, ETL, real-time
 Storage – “inventory”, “warehouses”, databases and
information systems
 Management/coordination – design and sustain
Hugos, M. Essentials of Suppy Chain Management, 2nd Edition, 2006
HBR on Supply Chain Management, 2006
5
Figuring Out the Data Supply Chain
New Visit
Note
Lab
Report
Billing
System
Medical
Record
Radiology
Report
Path
Report
Tx Summary
F/U Visit
Note
Clinical Data
Warehouse
Tumor
Registry
Research
Database
Data | Consumer | Supplier | Processing | Distribution | Storage | Mgmt
6
Workflow Design: Follow-up Visit

Beginning of visit




Consumer(s): MD
Data: relevant PMH, HPI, recent results,
symptoms, medications, QOL
Upstream supplier(s): Patient, Lab, Radiology,
Pathology, EMR
End of visit



Downstream consumer(s): patient, billing,
medical records, scheduling, researchers
Data: prescriptions, plan, education, encounter
bill, documentation, status
Supplier(s): MD
7
eForms
8
High
Data Feed Prioritization
Protocol
SSDI
DemoProcedures
Accruals
Appts
graphics
Low
Collection Cost
Lab Values
>6 Week Lag
Velocity
Real-Time
Where is the “biggest bang for the buck”? Where is the “low-hanging fruit”?
9
“Swim-Lanes” and Silos
Understanding Data Storage and Processing
10
Quality Effects of Integration
Clinic Workflows
 Populate clinic forms
from research database
 Multiple people view,
enter and update data
 Collect research data
during clinical workflows
Research Workflows
 Fill gaps / correct errors
 Identify analysis outliers
 Longitudinal follow-up
11
Data “Supply Chain” Analogy







Data / information: in its most raw, granular form
Consumers: Who needs what data or information? When, where
and how? What format?
Suppliers / sources: Who generates/collects what data elements?
When, where and how? What format?
Processing / activities: Who can most efficiently or effectively
process what data? When, where and how?
Distribution: Who transports what data?
When, where and how? What format?
Storage: Who stores what data in a warehouse or database?
Where and how? What format?
Management / coordination:



Capture data as far upstream as possible
Minimize steps, especially manual ones (OHIO)
Organize chain of collection, movement, storage and processing to
efficiently deliver data or information to consumer JIT for use
12
Free Software and Collaboration
To demo, download or get more information visit
http://Caisis.org
13
MSKCC Caisis Team - 2007
Avinash Chan
Kevin Regan
Frank Sculi
Vicki Cameron
Paul Alli
Beth Roby
Jason Fajardo
Kerry McCarthy
Brandon Smith
Not pictured: Tumen Tumur, Kinjal Vora
14
Appendix: Caisis Project Timeline

Microsoft Access databases
 1999 ProstateDB 1.0
 2000 PRDB / Prostabase
 ColdFusion & SQL Server web-based database
 2002 Valhalla 1.0 – 1.1


2003 Valhalla 1.2 (7,994 patients)


Prostate
Billing/EMR compliant populated clinic forms
Microsoft.NET & SQL Server web-based database
 2004 Caisis 2.0 – 2.1 (26,470 patients)


2005 Caisis 3.0 – 3.1 (44,000 patients)


Prostatectomy eForm, protocol manager, tumor maps
2006 Caisis 3.5 – (55,000 patients)


Integrated bladder, kidney, testis
GU and Urology Prostate Follow-up eForms
2007 Caisis 4.0 – (65,000 MSKCC patients)

Metadata-driven, dynamic forms, new diseases and eForms
15
Appendix: Caisis Next Steps, 1 of 2

BISTI/National Cancer Institute grant R01-CA119947







Restructure data model to accommodate other diseases through
metadata-driven fields and dynamically generated web forms
Migrate dataset production algorithms, nomograms, longitudinal
patient follow-up tools, project tracking and other prototyped
features into the Caisis framework
Make Caisis compatible with interoperability standards from the
Biomedical Informatics Grid (caBIGTM)
Support adoption and collaborative development of Caisis by
maintaining the Caisis.org website, web conferences and face-toface meetings, issue tracking, and training and documentation
Simplify installation, configuration, security, auditing,
customization and ongoing maintenance
Program the web-based user interface for compatibility with all
major web browsers
Improve the system’s scalability and portability
16
Appendix: Caisis Next Steps, 2 of 2

eForms
Form tracking and email system for
scheduled surgeries and clinic visits
 Shift navigation from passive to directing
and “pulling” users through tasks
 Reduce physician time and clicks to
complete forms

Specimen tracking module
 Plugins

17
Appendix: Multi-Institutional Adoption / Collaboration
Over 15 sites, 400 users, and 165,000 patients
1.
2.
3.
4.
5.
Baylor College of Medicine
Cancer Research UK - London
Case Western Reserve University
Cleveland Clinic
Eastern Virginia Medical Center
6.
Helios/Wuppertal
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
George Washington University
McGill University
MD Anderson Cancer Center
Memorial Sloan-Kettering Cancer Center
North Shore Long Island Jewish Health System
Ottawa Hospital – Civic Campus
Seattle Consortium (Fred Hutchinson / Univ of Washington)
Stiftung biobank-suisse
University of Alabama – Birmingham
University of California - Davis
University of Malmö - Sweden
University of Rochester
University of Texas – San Antonio
University of Texas Southwest Medical Center
Wake Forest University
Wayne State University / Karmanos Cancer Institute
Westmead / Breast Cancer Tissue Bank – Australia
18
Appendix: Caisis Privacy and Security


Limited access to patient data by job function
(role/permissions) and dataset
HIPAA compliant data export



IRB approval or de-identification required
Disclosures logged
Tracking / Logging



Who views which patient
Who performs what action
Nothing is overwritten (full audit trail)
19
Appendix: Dataset Production Algorithms
Automated variable selection and progression calculations
20
Appendix: Caisis Protocol Manager
21
Appendix: External Interfaces / caBIG
caBIG Grid
MSKCC
DMZ
MSKCC
Network
Catalog
caTISSUE
Suite
Tracking
JIT Annotation
caBIG
22
Appendix: Metrics
23