CDISC © 2010 The CDISC Protocol Representation Model and IHE’s RPE: A SHARP Solution SHARP Teleconference 16 August 2010 Rebecca Kush Landen Bain Value of a Structured Protocol Representation • Structured core protocol information to facilitate re-use of (trial registries, study management/tracking, study design, reporting) • Ensure compliance with IRB requirements • Facilitate study team comprehension of study requirements • Automation of Case Report Form creation (based on study design) and/or EHR configuration to support clinical research NOTE: NOT intended to inhibit creativity or innovation in study designs/protocol content 2 PR Group Approach • Development should concentrate on content first and implementation second • Elements must be defined in a glossary, since the industry uses multiple definitions for the majority of protocol elements – CDISC Glossary, Applied Clinical Trials, published yearly • Flexible enough to accommodate any protocol-based research • Identify core set of elements initially, expand with further details as needed • Initially based on ICH, EudraCT/clinicaltrials.gov/WHO requirements 3 CDISC Protocol Representation Standard - Development Protocol Representation Excel Spreadsheet Clinical Trial Tracking, Study Summary (SDTM) Clinical Trial Registry XML Schema BRIDG Mapping; PR V Study 1.0 Standard Clinical Trial Tracking Summary (STDM) Clinical Trial Development Harmonization Registry Documentation Eligibility Criteria 9most common) CDISC Trial Design Part I (aims, elements, visits) CDISC Trial Design PART II Planned assessments and interventions (NCI Study Calendar) CDISC Statistical Analysis Plan Other Protocol Template Sections and Attachments Eligibility Criteria (most common) PR V 1.0 Q1 2010 CDISC Trial Design Part I (arms, elements, visits) CDISC Trial Design Part II Planned assessments & interventions (NCI Study Calendar) CDISC Statistical Analysis Plan PR V 1.x (2011) Other Protocol Template Sections and Attachments Copyright CDISC 2009 4 The BRIDG Model* A clinical research domain analysis model (UML) initiated by CDISC, BRIDGing – Organizations (CDISC, HL7, FDA, NCI) – Standards – Research and Healthcare Towards semantic interoperability; a Portal to Healthcare Open source ; Collaborative Project – See BRIDG Model on CDISC website or www.bridgmodel.org *Biomedical Research Integrated Domain Group (BRIDG) Model 5 Global Biomedical Research Standards (Protocol Reporting) Integrated Standards (BRIDG Release 3) and Controlled Terminology/Vocabulary Protocol Case Report Forms Laboratory Data Tabulated CRF data (SDTM) Analysis Datasets FDA eSubmissions Analysis and Reporting Harmonized w/NCI caBIG CRFs CDISC ODM and/or HL7 Transport BRIDG = Biomedical Research Integrated Domain Group Model Integrated Standards (BRIDG Release 3) and Controlled Terminology/Vocabulary Protocol •Study Design •Eligibility •Registration •Schedule Case Report Forms (CDASH)** •Study Data FDA eSubmissions Analysis and Reporting Laboratory Data (LAB) Tabulated CRF data (SDTM) •Study Data •Lab Data •Study Design •Schedule Analysis Datasets * (ADaM) ** Harmonized w/ NCI caBIG CRFs * CDISC ODM and/or HL7 Transport BRIDG = Biomedical Research Integrated Domain Group Model BRIDG Accomplishments • Released BRIDG 3.0.1 January 2010 • ‘Balloted’ as CDISC Standard through the Joint Initiative Council (JIC) in 2009/2010 • Passed ISO and HL7 ballots in May 2010 7 Domain-Friendly, Subdomain-specific Business Model Layer 1 Domain-friendly, Subdomain-specific business models Layer 2 BRIDG Domain analysis model (DAM) Layer 3 RIM based GRIDG model 3 Layer Model 8 View PR Protocol Representation View PR Protocol Representation 9 CDISC Protocol Representation Model V Version 1.0 Now Available Protocol Section Info for Trial Registration and Tracking Eligibility Criteria Study Design: Arms, Epochs CRF Data Development Collection Data Analysis Report or eSubmission Study Summary Information Re-Use Improved Quality and Efficiency PR Version 1.0 SDTM Eligibility Criteria Study Design: Arms, Epochs Study Design: Study Design: Planned Events Planned Events CDASH CRFs Statistical Analysis Plan Data Collection Data Tabulation SDTM Tabulated Data Data Analysis ADaM Analysis Datasets Appendices, 10 etc. Appendices, etc. Clinical Trial Registration: CTRR Project • The intent of the CTRR project is to create a comprehensive and generic interchange standard for Clinical Trial Registries that includes all required and optional elements of most external clinical trial registries; including, but not limited to, EudraCT, clinicaltrials.gov, PDQ, and WHO. • Health Level Seven (HL7) Regulated Clinical Research Information Management (RCRIM) Technical Committee Project • The CDISC/HL7 clinical trial registries working group includes representatives from several pharmaceutical companies (eg., Novartis, J&J PRD, GSK), technology solution providers, CROs, NIH, clinicaltrials.gov, and observers from the FDA. 11 Eligibility Criteria • Each criterion is an Observation about a subject. • The observation can be complex, based on multiple other activities or results of other observations; the model has the capability to deal with such complexity. Eligibility Criteria – ‘pan-disease’ • • • • • • • Minimum Age Maximum Age Gestational Age Subject Gender Pregnancy Nursing Current Population Disease Condition • Past Population Disease Condition • Prior and Concomitant Medication(s) • Subject Ethnicity • Subject Race(s) • Special Populations • Ethical Considerations (e.g. informed consent) • BMI • Lifestyle Choices • Substance Use 13 Defined eligibility criteria Deferred Observation Defined inclusion Criterion Defined exclusion criterion 14 Coded Core Set of Eligibility Criteria Use Case: Joyce Niland, City of Hope • Minimal dataset to support the following use cases: – Locate potential protocols for a given individual • Tailor the query by matching core protocol eligibility against the subject characteristics – Locate potential subjects for a given protocol • Based on coded EHR data, to then contact (with permission) and screen for full eligibility – Evaluate the utility and prevalence of common core eligibility criteria • Assist in the initial design or refinement of protocols Sample Use Case: City of Hope Breast Cancer-Specific Protocol Search Filter City of Hope Breast CancerSpecific Protocol Search Filter Study Design: Two Views • The protocol includes two views of a study: – The experimental design • Often represented in a study schema – The schedule of activities • Often represented as a time and events table Source: Diane Wold 17 Schedule of Activities from ICH E3 annex Sample schedule of activities Source: Diane Wold 18 Experimental Design Example Experimental design example to include screening, induction, continuation and followScreening Epoch Induction Epoch up Chemotherapy & Radiation Arm Continuation Epoch Induction Treatment Continuation Chemotherapy Induction Treatment Surgery Follow-up Epoch Follow-up Screen Continuation Chemotherapy Follow-up Chemotherapy, Radiation & Surgery Arm Experimental Design in the Protocol Representation Model An activity in a global library of activities An activity in a global library of activities A planned instance of an activity at a particular point in a study CDISC Standards and Data Flow BRIDG model to include healthcare standards and CDISC clinical research standards CDISC Standards and Data Flow BRIDG model to include RPE healthcare standards and CDISC clinical research standards and where RPE fits in What is Retrieve Process for Execution (RPE)? • An IHE profile that leverages BPEL and BPM to automate collaborative workflow between healthcare and secondary use domains such as research. 23 How does RPE work? • Externally defined processes, such as a research protocol, are consumed and executed by EHR systems. • Web service ‘actors’ provide methods to import business process definitions (activities, tasks) as executable BPEL code. • Execution of the process is recorded by a state manager. 24 Retrieve Process for Execution Well-accepted IT standards do exist for business process management EHR (BPM) architecture. However, these areProcessActivityExecutor primarily focused on InitiateActivity ReceiveProcessStateAlert processes ReceiveProcessDefinitions managed InitiateProcess within an organizational RetrieveActivities RetrieveProcessDefinitions UpdateActivity boundary. RPE defines a Subscribe ReceiveProcessStateAlert RetrieveProcessDefinitions set of transactions for Subscribe collaboration acrossReceiveProcessDefinitions IT ProcessDefinitionManager ProcessStateManager DeployProcessDefinition system boundaries involving three main Research EDC Sponsor actors: the manager of process definitions, the manager of runtime RPE 2.0 Architecture • Key Goals: – Support collaborative process management – Be consistent with accepted BPM architecture BUT, – Accommodate existing variety of process engine/task processor implementations – Support arbitrary process representation formats Trial Design vs. Trial Execution • “The trial design model is just that – a model of the trial’s design. It is not a record of an individual patient’s route through the trial. An execution runtime would reference the design elements to decide how a patient should progress through the trial and record this path accordingly. The trial design incorporates aspects of: – Structure: Arms, Cells, Activities, etc. – Workflow: How a patient should progress through a trial – decision points, branches, etc. – Timing: When the activities should happen relative to other activities or decision points in the trial. Trial execution would need to capture at least the following: – – – – – The path that a patient took through the trial design. The activities that were performed on that path. The planned times for when the activities should have happened for that patient. The actual times that the activities and decision points were performed. References to the design elements for static information that is common. Trial execution is outside the scope of this initial proposal.” BPEL/BPEL4People provides an IT standard execution runtime solution What is needed? • Discussion and analysis of how high-throughput phenotyping fits into the protocol representation and RPE paradigm. • Hypothesis: phenotype definition is a ‘content specification’ that could be represented within Protocol Representation and executed in RPE. 28