Shared Health Research Information Network
Andrew McMurry, MSSHRINE Architect
Harvard Medical School Center for BioMedical Informatics Children’s Hospital Informatics Program Harvard/MIT HST
Boston University Bioinformatics PHD Student
Andrew_McMurry(@) hms.harvard.edu
Current Status, Demo, and Open Source
History of the SHRINE world in 1 minute
Partners RPDR i2b2 Data Repository Shawn Murphy et all (wont steal their thunder here) Everyday patient encounters huge research cohorts
SPIN cross-hospital query model Extract, Transform <de-identifiy, autocode>, Load Federated query over multiple independent hospitals
Harvard CTSA funded
Recombinant deploying i2b2 across the country
Overly simplified: clinical data repository + federated query = SHRINE
SHRINE Technical Architecture
Bird’s Eye View Leverage local i2b2 deployments
Broadcast queries and aggregate responses across autonomous sites as if they were “one clinical data warehouse”
There is no central database
Connect sites in a peer-to-peer or hub-spoke fashion
Current Status : in BETA!
East Coast : Harvard Effort BETA system running at Harvard across BIDMC, Children’s,
and Partners representing both BWH and MGH.
Last AUG: Prototype 1 year of data with demographics and
diagnosis IRB approval for demonstration only
NOW Patient data from 2001+ Demographics, Diagnosis, Medications, some Labs Full IRB approvals
Current Status : in BETA!
Demo use cases, things to think about…..
Lung cancer cases requesting biospecimens
Irritable Bowel Syndrome requesting sized cohorts
Pediatric rheumatology Creating national registry
Current Status
National Registry : Pediatric Rheumatology
“Grand Opportunity” Grant
60 institutions, 2 years SHRINE has been selected as inter-institutional glue Consented patients, prospective data collection Shared Virtual Machine Hosting
SHRINE Open Source : brewing in the background
Fast recap of 2006 and 2009 Open Source proposals:
1. Write documentation as if we are writing code stubs Preconditions + Steps (Algorithm) + Post-conditions = Use Cases or “Scenarios”
2. Test driven development Write the test first Implement second Regression test everything all the time
Continuous Integration 1. Test everything every time you change any code
SHRINE Open Source Process Leveraging routine development activities
1. Public access to bug tracking system
2. Public access to latest source code
3. Public mailing lists• Shrine-users • Shrine-developers• Shrine-announce
4. Test Driven Development
5. Continuous integration (daily builds of the complete software)
Acknowledgements: Core SHRINE team
Zak Kohane (SHRINE Lead / HMS)Griffin Weber (HMS CTO / bidmc)Shawn Murphy (I2B2 CRC / partners) Dan Nigrin (Children’s CIO)Ken Mandl (Public Health Use Cases/ CHIP IHL)Sussane Churchill (I2B2 Executive director)Doug Macfadden (HMS CBMI IT Director)Matvey Palchuck (Ontology Lead / HMS)Andrew McMurry (Architect / HMS)
Could give an entire talk on all the collaborators, multi-institutional effort. Asking forgiveness from those not listed
Acknowledgements: Core SPIN team
Zak Kohane (SPIN PI / HMS)Frank Kuo (PSL Program Director / BWH)
John Gilbertson (PSL Pathologist / MGH)Mark Boguski (PSL Pathologist / BIDMC)Antonio Perez (PSL Pathologist / Children’s)Mike Banos (PSL Developer / BWH )Ken Mandl (Biosurviellance PI/ Children’s) Clint Gilbert (Biosurviellance Dev Lead / Children’s) Greg Polumbo (SPIN Developer/ HMS) Ricardo Delima (SPIN Developer / NCI at HMS) Britt Fitch (SPIN Developer / HMS
http://spin.chip.org/community.html