Post on 15-Jul-2020
transcript
1
LONGITUDINAL DATA ANALYSIS IN HEALTHCARE
HOW TO SET UP RELEVANT AND ACCESSIBLE DATABASES
Prof. Dr. Hans-Ulrich ProkoschPD Dr. med. Thomas Ganslandt, Dr. Martin Sedlmayr, Jan ChristophChair of Medical Informatics, Friedrich-Alexander Universität Erlangen-Nürnberg
02.05.2017
ZD.B Symposium: The Future of Healthcare – Big Data Driven Healthcare
Make EHR Data Accessible
Prokosch HU, Ganslandt T. Perspectives for medical informatics. Reusing the electronic medical record for clinical research. Methods Inf Med. 2009;48(1):38-44.
Data Warehousing
Make EHR Data Accessible:Instrumenting the health care enterprise for discovery research in the genomic era
Murphy S, Churchill S, Bry L, Chueh H, Weiss S, Lazarus R, Zeng Q, Dubey A, Gainer V, Mendis M, Glaser J, Kohane I. Instrumenting the health care enterprise for discovery research in the genomic era. Genome Res. 2009 Sep;19(9):1675-81.
McMurry AJ, Murphy SN, MacFadden D, Weber G, Simons WW, Orechia J, Bickel J, Wattanasin N, Gilbert C, Trevvett P, Churchill S, Kohane IS. SHRINE: enabling nationally scalable multi-site disease studies. PLoS One. 2013;8(3):e55811.
Kohane IS, Churchill SE, Murphy SN. A translational engine at the national scale: informatics for integrating biology and the bedside. J Am Med Inform Assoc. 2012 Mar-Apr;19(2):181-5.
FederatedQuery
Networks
4
Data Integration – international effortsPCORnet
33 Partner Networks• 13 Clinical Data Research Networks (CDRNs)
• based in healthcare systems such as hospitals, integrated delivery systems, and federally qualified health centers
• 20 Patient-Powered Research Networks (PPRNs)• operated and governed by groups of patients and their partners.
• identify the optimal doseof aspirin for secondary prevention in patients with atherosclerotic cardiovascular disease (ASCVD)
• identify patients who are at high risk for ischemic events (phenotyping)
• 20,000 patients are randomly assigned to receive an aspirin dose of 81 mg/day vs. 325 mg/day
5
Data Integration – international effortsPCORnet – Adaptable Trial
First pragmatic clinical trial based on the PCORnet networks
6
Data Integration – international effortsOHDSI
• multi-stakeholder, interdisciplinary collaborative to bring out the value of health data through large-scale analytics.
Observational Health Data Sciences and Informatics (OHDSI, pronounced "Odyssey")
7
Data Integration – international effortsOHDSI large scale analysis
Hripcsak G, Ryan PB, Duke JD, Shah NH, Park RW, Huser V, Suchard MA, Schuemie MJ, DeFalco FJ, Perotte A, Banda JM, Reich CG, Schilling LM, Matheny ME, Meeker D, Pratt N, Madigan D. Characterizing treatment pathways at scale using the OHDSI network. Proc Natl Acad Sci U S A. 2016 Jul 5;113(27):7329-36.
8
Data Integration – industrial effortsIBM Watson Health
9
Data Integration – industrial effortsIBM Watson Health
Core Elements• Data Integration Centres
o at University Hospital Sites
o within funded consortia
o across consortia (Germany wide cooperation)
• currently seven funded consortia(http://www.gesundheitsforschung-bmbf.de/de/6685.php)
o have submitted proposal for networking anddevelopment phase (28.4.17 1.1.2018)
• National Steering Committee (NSC), Working Groups
o Patient Consent / Data Use / Data Governance / Data Protection / Data Security
o Use and Access Policies / Committees
o Interoperability
10
Data Integration – national initiativeBMBF Medical Informatics Funding Scheme
https://www.bmbf.de/pub/Medical_Informatics_Funding_Scheme.pdf
Governance
Ethics Committee
Board of DirectorsUse- & Access
Committee
11
The MIRACUM example:Components of a Data Integration Center
Comm-server
Routine-Business
Reporting
Long Term Data
Archive
ID- und Consent-Manage-
ment
Federation
sourcesystem 1
sourcesystem 2
Data Warehouse
ETL
Research Queries / Analysis
ResearchData
Repository
Harmonisation
NLP
Phenotyping
Enrichment
Authentification
Audit-Trail
Security
MDR
Terminologyservice
Metadata
Project Proposal& Trial
Registry
Data Transfer Unit Software Test/ Deployment
Pipeline
Data Provenance
Visualisation…..
12
The challenge of data harmonization
OMOP Common Data Model(Observational Medical Outcomes Partnership)
https://www.ohdsi.org/data-standardization/
13
The challenge of data harmonization
http://www.pcornet.org/pcornet-common-data-model/
14
The MIRACUM consortiumFirst Pilot Results
The challenge of data harmonization
MIRACUM: towards a Learning Health SystemDistributed Data Analysis
Establishing a first DIC Infrasturcture atall 8 MIRACUM partner sites• Applying the i2b2-Plattform
(Informatics for Integrating Biology & the Bedside)
o Open Source, US-, European and German AUG
• Based on billing data for inpatient stays
o covers 5 of 7 data types basis module in thein the NSC core data set
o standardized data strutures across all sites
• Appication of this infrastructuer for
o distributed analysis
o federated cohort identification
16
MIRACUM DIC 0.9 (after nine months conceptual phase)
17
MIRACUM DIC 0.9Distributed Analysis
MIRACUM inpatient coverage MIRACUM disease category distribution
Cross-consortial cooperationMIRACUM – HD4CR• Charité, Ulm, Würzburg, Vivantes
• Provision of i2b2 instances
• Provision of ETL routine
Joint geovisualisation• Years 2015 - 2016
18
MIRACUM DIC 0.9Distributed Analysis
applies broker / connector concept• developed in previous projects (DKTK, DZL, GBN: German Biobank Node)
o based on i2b2 research data repository and -webclient
o connector polls the broker (IT security: no open ports for access from outside)
• supports federated cohort identification (aggregated patient counts)
• query on colorectal cancer patients across all participatingMIRACUM/HD4CR sites
19
MIRACUM DIC 0.9Feasibility Studies
Standort 1
Researchdata
repository
Standort 2
Researchdata
repository
Connector ConnectorcentraleQuery Broker
QueryFrontend
20
MIRACUM DIC 0.9federated cohort query across 10 hospitals
http://www.miracum.de/http://www.hd4cr.org/
21
MIRACUM DIC 1.0next steps – adding omics data
22
MIRACUM DIC 1.0next steps – adding omics data
Open API supports the development of new Analyse-Plugins, e.g. Kaplan-Meier Plots
23
MIRACUM DIC 1.0next steps – adding omics data
1Knell C. et al. Developing interactive plug-ins for for tranSMART using the SmartR fram ework workwork: The case of survival analysis. Stud Health Inform 2017
2Christoph J. et al. Two Years of tranSMART in a University Hospital for Translational Research and Education. Stud Health Inform 2017
24
MIRACUM DIC 1.0next steps – adding imaging data
Murphy SN, Herrick C, Wang Y, Wang TD, Sack D, Andriole KP, Wei J, Reynolds N, Plesniak W, Rosen BR, Pieper S, Gollub RL. High throughput tools to accessimages from clinical archives for research. J Digit Imaging. 2015 Apr;28(2):194-204.
Herrick R, Horton W, Olsen T, McKay M, Archie KA, Marcus DS. XNAT Central: Open sourcing imaging research data. Neuroimage. 2016 Jan 1;124(Pt B):1093-6.
Marcus DS, Olsen TR, Ramaratnam M, Buckner RL. The Extensible Neuroimaging Archive Toolkit: an informatics platform for managing, exploring, and sharingneuroimaging data. Neuroinformatics. 2007 Spring;5(1):11-34.
He S, Yong M, Matthews PM, Guo Y. tranSMART-XNAT Connector tranSMART-XNAT connector-image selection based on clinical phenotypes and genetic profiles. Bioinformatics. 2017 Mar 1;33(5):787-788.
25
Next Steps:Adding Patient Generated Data
Mandl KD, Kohane IS. Time for a Patient-Driven Health Information Economy? N Engl J Med. 2016 Jan 21;374(3):205-8.
https://blog.dacadoo.com/2012/11/06/sonntagsblick-quantified-self-das-handy-ist-mein-fitnesstrainer/
Quantified Self
WearablesSensor Data
26
The MIRACUM example:How to set up relevant and accessible Databases
Governance
Ethics Committee
Board of DirectorsUse- & Access
Committee
Comm-server
Routine-Business
Reporting
Long Term Data
Archive
ID- und Consent-Manage-
ment
Federation
sourcesystem 1
sourcesystem 2
Data Warehouse
ETL
Research Queries/ Analysis
ResearchData
Repository
Harmonisation
NLP
Phenotyping
Enrichment
Authentification
Audit-Trail
Security
MDR
Terminologyservice
Metadata
Project Proposal& Trial
Registry
Data Transfer UnitSoftware Test/
DeploymentPipeline
Data Provenance
Visualisation…..
i2b2
OMOP
XNAT
tranSMART
pEHR (Patient)
Thank you very much !
ulli.prokosch@uk-erlangen.de