EmergingTechnologies..

Report
Emerging Technologies
Semantic Web and Data Integration
This meeting will start at 5 min past the hour
As a reminder, please place your phone on mute unless you are speaking
06 Sep 2013
Emerging Technologies
Semantic Web and Data Integration
06 Sep 2013
Meeting Agenda
• New Projects: Call for Volunteers
• Semantic Technology Initiatives - Lee Feigenbaum
• Reusing Medical Summaries for Enabling Clinical
Research – Landen Bain
New Projects: Call for Volunteers
Group I (Sep 2013):
• Representing Regulations and Guidance in RDF – Mitra
• Representing CDISC Conformance Checks – Scott
• Representing SDTM, SEND, and ADaM Datasets in RDF –
Phil
• Toolsets to Access Clinical Trial Data Represented in RDF
(e.g. SAS, R, etc.) – Marc
Re-using medical summaries for enabling
Clinical research
Objective:
– Pre-Population of data collection sets from the medical
summaries of an eligible patient group through the use of an
ISO 11179 enabled semantic metadata repository through IHE
DEX profile.
Rationale:
– Demonstrating that a semantic metadata registry can bridge
the gaps between data elements used in clinical research and
clinical care for extracting the required data for clinical
research studies from the existing patient medical summaries
Re-using medical summaries for enabling
Clinical research
Background:
– IHE DEX Profile:
– EHRs contains data elements that can be used to populate research case report
forms
– DEX leverages the power of an ISO/IEC 11179 Metadata Registry standard to apply
mappings earlier in the process, at the point of form design
– DEX will create a customized map of the elements in a particular case report form to
the corresponding elements in the EHR export. The metadata registry maintains the
exact interactions between the research and healthcare data elements, and will
provide an exact map by which data can be extracted from the pre-population data
RetrieveDataElementList
set
[QRPH -43] 
Metadata Consumer
RetrieveMetadata
Metadata Source
[QRPH -44] 
DEX maps secondary use domain data to healthcare data
elements
Re-using medical summaries for enabling
Clinical research
Background:
– Semantic MDR Implementation from SALUS
•
•
•
•
•
•
An implementation of the ISO/IEC 11179 Metadata Registries standard
with Semantic Web technologies on top of Linked Data principles
Allows to link the data elements defined by different authorities (like the
HITSP C154 data sets and CDISC SDTM variables) through linked data
principles
Allows definition of extraction specifications between abstract data
element definitions to physical content models (like CCD documents)
Open source version of this MDR is already available at:
https://github.com/srdc/semanticMDR.
Screencasts available at :
http://www.srdc.com.tr/projects/salus/blog/?p=181
This MDR also implements the DEX Profile, hence it allows retrieving
extraction specifications of data elements for different content models.
Re-using medical summaries for enabling
Clinical research
Use case from SALUS Project:
• ROCHE use case in SALUS project
– a pilot scenario to demonstrate that, the semantic MDR
and DEX can enable ROCHE to semantically annotate
the data set collection required for an observational study
through CDISC SDTM variables, and then through
SALUS it can seamlessly collect the required datasets
from EHR Sources although the EHR Sources send
anonymized medical summaries in HL7 CCD formats
Re-using medical summaries for enabling
Clinical research
Use case from SALUS Project (ROCHE Data Collection set):
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Patient id (Pseudonym)
Sex
Date of birth (or year of birth)
Date of ACS event
Date of ACS +30 days (Start date)
History of type 2 diabetes (T2D) before
start date (Y/N)
Date of the first T2D diagnosis date ever
Average HbA1C over the 12 months
before start date (will be missing for most
non diabetic pts)
Average systolic BP over 12 months
before start date
Average diastolic BP over 12 months
before start date
History of hypertension before start date
(Y/N)
Last BMI before start date
Last weight before start date
Ever smoked before start date(Y/N)












Smoked within the last 3 months of start
date(Y/N)
Taken Sulfonylurea anytime within 3
months before start date (Y/N)
Taken metformin anytime within 3
months before start date (Y/N)
Taken insulin anytime within 3 months
before start date (Y/N)
Taken Thiazolidinediones (Glitazones)
anytime within 3 months before startdate
(Y/N)
Taken other oral anti-diabetic drugs
within 3 months before start date (Y/N)
Had a CHF before start date (Y/N)
Had a CHF after start date (Y/N)
Date of CHF after start date
Patient died any time after start
date(Y/N)
Date of Death
Re-using medical summaries for enabling
Clinical research
Use case from SALUS Project:
• ROCHE defines the data set definitions by annotating them with SDTM
variables (Through a GUI)
• SALUS MDR hosts HITSP Data sets, and also the mappings of these data sets
from ASTM/HL7 CCD medical summaries
– Through XPATHs
– If the medical summary is in RDF, SPARQL is also possible
•
•
•
SALUS MDR hosts the SDTM data sets and also maintains the links between
HITSP data sets and SDTM variables by referencing to the unique URIs of
CDISC2RDF datasets
SALUS MDR also implements IHE DEX, hence when it is asked about the
extraction specifications of an SDTM variable from an HL7 CCD document, it
automatically follows the link between SDTM datasets and HITSP datasets,
retrieves the extraction specification from the CCD document, and readily gives
this to the Tool used at ROCHE side
The tool uses these mapping specifications of individual data elements and
builds extraction specifications to automatically retrieve the required data from
the CCD document.
Re-using medical summaries for enabling
Clinical research
Deliverables:
• Importing SDTM to Semantic MDR
• Importing HITSP Data elements to Semantic MDR, and defining mapping
specifications to ASTM/HL7 CCD
• Defining the mappings between SDTM variables and HITSP data elements (
Through skos:exact match, skos:closeMatch) for a subset required in demo
• Demonstrate a Data Collection Tool for an observational study to use the DEX
profile and semantic MDR capabilities to collect data from medical summaries
– We can check whether we can produce and process the annotated data collection set as
“CDISC Protocol Representation Model (PRM) in RDF”
– We can check whether we can produce the collected data sets in “RDF” as it is going to
be done for “Trial Data representation in RDF”
• Publish working examples to GitHub
Timeline:
• October 2013- February 2014

similar documents