Electronic Healthcare Record For Clincial Research EHR 4

Electronic Healthcare Record For Clincial Research (EHR 4 CR) Semantic interoperability framework WP 4 C. Daniel INSERM UMRS 872 team 20 January 22, 2013 Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 1

Objective, scope Executive Summary • Objectives – EHR 4 CR platform for reusing EHR data in order to support medical research – Comprehensive business model for governance, acceptance, adoption and sustainability • Duration & budget: 4 years – 7 Mons Euros (+ in kind contribution) • Partners: 33 European academic and industrial partners – 11 Pharmaceutical Companies (members of EFPIA) – 22 Public Partners (Academia, Hospitals and SMEs) January 22, 2013 Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 2

Partners January 22, 2013 Convergence Meeting: Semantic Interoperability for 3 Clinical Research & Patient Safety in Europe

Use cases A loosely coupled SOA, which interconnects independent services implementing EHR 4 CR usage scenarios Protocol feasibility Patient recruitment Clinical trial execution Pharmacovigilance January 22, 2013 1 Leverage clinical data to design viable trial protocols and estimate recruitment 2 Detect patients elegible for trials & better utilize recruitment potential 3 Re-use routine clinical data to prepopulate trial e. CRFs 4 Detect adverse events & collect/transmit relevant information Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 4

Use case 1 1 Protocol feasibility Patient care (HL 7 -13606) Patient identification Hospital n° 1 EHR CDW Clinical research (CDISC) Central queries transformed into local queries distributed Eligibility criteria CRO n° 1 and executed overmanually formalized into central EHRs/CDWs queries CTMS CDMS (EHR 4 CR workbench – SHARE data elements) Hospital n° 2 CRO n° 2 EHR CDW CTMS CDMS Hospital n°X EHR CDW Protocol feasibility Number of patients matching eligibility criteria CRO n°x CTMS CDMS 5

Use case 2 2 Patient recruitment Patient care (HL 7 -13606) Clinical research (CDISC) Patient identification Hospital n° 1 EHR CDW PSM Protocol content CRO n° 1 criteria distributed over local Protocol content +Eligibility process step managers manually formalized into central queries CTMS (including central ->local queries) (EHR 4 CR workbench – SHARE data elements) CDMS Hospital n° 2 CRO n° 2 EHR CDW CTMS CDMS PSM Hospital n°X CRO n°x EHR CDW CTMS CDMS PSM 6

Use case 2 2 Patient recruitment Patient care (HL 7 -13606) Hospital n° 1 EHR CDW PSM Patient recruitment process Lists of eligible->screened ->recruited subjects (protocol specific consent) Clinical research (CDISC) CRO n° 1 CTMS CDMS Hospital n° 2 CRO n° 2 EHR CDW CTMS CDMS PSM Hospital n°X EHR CDW PSM Patient recruitment Number of recruited subjects CRO n°x CTMS CDMS 7

Use case 3 3 Data capture Patient care (HL 7 -13606) Hospital n° 1 EHR CDW PSM Clinical research (CDISC) Forms CRO n° 1 Items manually formalized (SHARE data elements) CTMS CDMS Hospital n° 2 CRO n° 2 EHR CDW CTMS CDMS PSM Hospital n°X CRO n°x EHR CDW CTMS CDMS PSM 8

Use case 3 3 Data capture Patient care (HL 7 -13606) Hospital n° 1 EHR CDW PSM Patient data capture process Clinical research (CDISC) Forms At each visit, retrieve query CRO n° 1 Items manually formalized extraction specification & (SHARE data elements) form(s) for data capture CTMS CDMS Hospital n° 2 CRO n° 2 EHR CDW CTMS CDMS PSM Hospital n°X CRO n°x EHR CDW CTMS CDMS PSM 9

Use case 3 3 Data capture Patient care (HL 7 -13606) Hospital n° 1 EHR CDW PSM Clinical research (CDISC) Data ( Recruited subjects protocol specific consent) CRO n° 1 CTMS CDMS Hospital n° 2 CRO n° 2 EHR CDW CTMS CDMS PSM Hospital n°X CRO n°x EHR CDW CTMS CDMS PSM 10

Semantic resources • Common EHR 4 CR clinical information model/meta data repository – A unique global as view schema of the heterogeneous EHRs/CDWs distributed over different pilot sites across Europe. – Shared core data elements • Common EHR 4 CR terminology – Integrate a range of clinical terminologies that are needed to collectively encode the variety of clinical entities (including observations , procedures, substance administration, etc) represented in the EHR 4 CR information model. – Query expansion and some degree of terminological reasoning • Query extraction specification for eligibility determination and data capture January 22, 2013 Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 11

Semantic resources EHR 4 CR pivot representation 12 PHARMA Formal representation of eligibility criteria & clinical data (Eligibility criteria, items) CDISC-SDM Clinical data/document Age Cardiovascular disease • HBP • Angor • Cardiac failure SBP (mm. HG) Weight (kg) Tabacco ? yes/no Hemoglobin A 1 c HL 7 CDA (e. g. CCD) (Clinical data/documents) “Model of meaning” CDISC-ODM HOSPITAL Eligibility criteria Between ages 18 and 56 Hemoglobin A 1 c value within the diabetic range At least 2/3 systolic blood pressure measurements ≥ 140 mm. Hg Items in e. CRFs Last known weight (kg) Systolic Blood Pressure (mm. HG) Hypertension? Yes/No Smoking / Non smoking ? Frankfurt WPG 2 - 20 october 2011

EHR 4 CR Clinical Information Model • Material : “source” models – BRIDG model & HL 7 v 3 models (HL 7 RCRIM WG) • «Study. Design» and «A_Supporting. Clinical. Statement. Universal» models – I 2 b 2 model • Method: Model-driven engineering – Transforming HL 7 v 3 models in UML models and adapt these models to the purpose and scope of the EHR 4 CR project • Multidimensional EHR 4 CR Information Model (ISO 21090 datatypes) • Standardized data elements (pre-processed eligibility criteria of 10 clinical trials) • Tooling: Open Medical Development Framework (OMDF) [Ouagne 10] EHR 4 CR WP 7 Meeting Thurs 13 th September 2012 13

EHR 4 CR Clinical Information Model EHR 4 CR WP 7 Meeting Thurs 13 th September 2012 14

EHR 4 CR meta data repository & terminology (top down) • Material – Reference document templates (e. g. CCD, CDA templates) – Reference terminologies – Patient care: clinical findings, test results, labs, or medications, etc. – Clinical research (Med. DRA, CDASH/ Ontology of Clinical Research) • Method & Tooling – We used Bioportal to upload terminologies – from UMLS (SNOMED CT, LOINC, ICD-10 codes, etc. ) – from other sources (e. g for ATC or Path. Lex) – We developed a data element/value set editor to build a core data element repository EHR 4 CR WP 7 Meeting Thurs 13 th September 2012 15

EHR 4 CR meta data repository & terminology (bottom up) Scope Execution (Core data elements) Eligibility criteria of clinical trials 4 CT 80 data elements Mappings 16 10 CT 300 data elements 100 CT Standardization

Semantic services Protocol feasibility January 22, 2013 Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 17

Evaluation of the semantic resources 11899 20050182 27919 BIO 111482 CENA 713 B 2315 COU-AA-301 D 3191 C 00009 D 4320 C 00015 EFC 11785 NC 25113 Total 18 Bayer Amgen Merck Cardiovascular Oncology Nervous system disorders Oncology Neurology onco CV/Arrhythmia s Oncology GSK Novartis Janssen Astra. Ze neca Sanofi Oncology Roche Cardiovascular and Metabolic X X X WW U Tota l uom Univ dun Uo. G MU W U 93 6 UCL KCL* Disease Area HUG EFPIA Partner APHP FAU Internal Study ID 10 clinical trials – 11 pilot sites 23 proofs of concepts X 3 2 3 X 2 1 X X X X X 3 2 1 1 2 2 X 3 1 4 3 23

Adoption of standards • IHE integration/content profiles • Standards – Reference information models (CDISC, HL 7) – ISO 11179 Meta data repository – ISO 21090 data types Boolean String Concept Descriptor Physical Quantity Coded Ordinal Real Integer <code ='271649006' display. Name=‘Systolic blood pressure' code. System='2. 16. 840. 1. 113883. 6. 96 ' code. System. Name='SNOMEDCT'/>

Semantic interoperability issues • Building/maintaining semantic resources – Sharing core data elements (CDISC SHARE) – Representing consistently complex clinical information models (clinical data structures templates, data elements) including clinical context • Antibiograms, anatomic pathology cancer checklists (histologic type, grade, TNM, tumor size, etc) – Supporting the mapping process between pivot/local information models & terminologies • Semantic services – Clinical research • Formal representation of eligibility criteria & e. CRF – Mapping medical concepts in eligibility criteria/e. CRFs including highly pre-coordinated terms to standard reference terminologies – Representing temporal constraints – Patient care • Mapping local data structures and/or interface terminologies to pivot models January 22, 2013 Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 20

Project Information • Project website : http: //www. ehr 4 cr. eu/ • Publications – Ouagne D, Hussain S, Sadou E, Jaulent MC, Daniel C. The Electronic Healthcare Record for Clinical Research (EHR 4 CR) information model and terminology. Stud Health Technol Inform. 2012; 180: 534 -8 – El Fadly A, Rance B, Lucas N, Mead C, Chatellier G, Lastic PY, Jaulent MC, Daniel C. Integrating clinical research with the Healthcare Enterprise: from the RE-USE project to the EHR 4 CR platform. J Biomed Inform. 2011 Dec; 44 Suppl 1: S 94 -102. • Contact information WP 4: Christel. daniel@crc. jussieu. fr January 22, 2013 Convergence Meeting: Semantic Interoperability for Clinical Research & Patient Safety in Europe 21