|
|
 |
|
|
|
| As you read in the last edition of Harmony, the CDISC strategy for 2006 and beyond will focus on harmonization of the separate models into one cohesive standard via the CDISC Technical Roadmap.
Therefore, in subsequent issues of Harmony, the updates will focus on these efforts instead of individual standards. However, CDISC standards teams were so active in 2005, it was decided to devote an entire special edition to cover their many accomplishments and upcoming goals.
We would like to take this opportunity to extend great appreciation to all the volunteers who have been working on these teams. Through their hard work and dedication, much has been accomplished, as evidenced in the updates in this edition. Thank you, CDISC team members for all you have done! We look forward to continuing to work with you, and new volunteers, in this new and exciting phase of model harmonization and implementation.
Additional information can be found at
http://www.cdisc.org |
|
|
by W. Kubick
During 2005, the Submissions Data Standards (SDS) team released version 1.1 of the Study Data Tabulation Model and release 3.1.1 of the SDTM Implementation Guide for Clinical Trials. In addition to continued support for harmonization projects such as the Trial Design Model, Controlled Terminology, and Define.XML, the team also worked on developing new SDTM domain models for Pharmacokinetics, Microbiology, Pharmacogenomics and Imaging data. In the area of Controlled Terminology, the team identified terminology for more than 30 SDTM variables, and prepared a standardized list of commonly used Lab Test Codes. The team also continued development of additional implementation guides to help sponsors how to represent data that falls into multiple classes, such as Findings about events and Signs and Symptoms. SDS representatives also worked with the BRIDG team to represent the key components of the SDTM in the BRIDG model.
by W. Kubick
A new workshop demonstrating the End-to-end integrated use of the CDISC LAB, SDTM and ODM models was prepared by Dave Iberson-Hurst and Tony Friebel and presented initially at the CDISC interchange. Additional documentation on the harmonization of these models will be released this year as a CDISC Roadmap deliverable.
by S. Handu
SEND Consortium was split into three sub-teams focusing on Reproductive, Safety- Pharmacology and Pharmacokinetic domains. It was decided to finalize the Implementation Guide for general toxicology studies to include single-, repeat- dose and carcinogenicity studies.
SEND version 2.3 was released in November 2005. This version of SEND is ready for the industry to start sending toxicology data to the FDA. FDA is working on streamlining processes internally in order to start accepting SEND data.
by S. Cassells
During 2005, the ODM team has been finishing work on Version 1.3 of the ODM. This version introduces several important new features, including typed data transmission and support for conditional study events, forms, item groups and items. Look for the publication of the new version in March 2006 on the CDISC website. |
|
 |
|
|
by S. Kenny
In the beginning of 2005, the ADaM team updated the statistical models for Change from Baseline and Categorical Analysis to reflect recent changes in the SDTM Version 3.1 release. These models were released for comment in April.
A new model, the Subject-Level Analysis was released for comments in April 2005. This model was developed to illustrate the application of statistical methodology, specific to analyses that utilize one record per subject. Based on feedback from the joint SDS/ADaM meeting in September 2005, the Subject-Level Analysis model was incorporated into the General Considerations model as the key Analysis Dataset and will serve as the model for all other statistical analysis datasets. The model illustrates the use of SDTM tabulation datasets as input for the derivation of statistical variables. This model was completed in late 2005 and will be released for final comments in early 2006 as the Analysis Data Model Version 2.0
In mid-2005, work was initiated to develop an ADaM model for the analysis of adverse event data. This model is the first ADaM model that specifically addresses safety analysis and is built largely upon the SDTM AE domain. This model has been reviewed by the ADaM team and will be finalized and released for comments in early 2006. During 2005, progress was also made on developing the ADaM model for linear model and time to event analyses.
by C. Willoughby
The Protocol Representation (PR) Group invested a significant amount of time and effort during the first half of 2005 towards developing a “static” model of a clinical trial protocol within the BRIDG Model. This early model includes semantics from the CDISC Glossary and the PR Spreadsheet of Elements. Although it began as a representation of a clinical trial protocol document, the model quickly evolved to encompass the entire plan for a clinical trial, including draft models for trial design and statistics. The PR Group also piloted new software, model maintenance, and people processes for the BRIDG project.
As part of our outreach efforts, several PR Group members attended a public meeting at the FDA on April 20, 2005 to share feedback on ICH efficacy guidelines that influence clinical trial protocols. In May, we joined the SDS team for our first joint team meeting. We also conducted a comprehensive review and provided feedback on the SDS Trial Summary domain.
At our Face-to-Face meeting, we kicked off a new project, led by Lakshmi Grama (NCI), to gather requirements for the development of an HL7 message to support exchange of clinical trial registry information. By September, we identified a total of 64 information elements (~85% come from the protocol) and mapped them to the elements and code-lists of several key clinical trial registries (PDQ, clinicaltrials.gov, and WHO) and the SDTM Trial Summary domain. These clinical trial registry elements were shared with representatives from the European Agency for the Evaluation of Medicinal Products (EMEA) and the World Health Organization (WHO) International Clinical Trial Registry Platform during Q4.
In addition to championing the new projects outlined above, the PR Group provided substantial support during 2005 to update previous work products with new releases such as Version 2 of the PR Spreadsheet of Elements, posted to the CDISC Website in August; Part 2 of the Trial Design Model, released for comment early October, 2005; and Version 4 of the CDISC Glossary, published in Applied Clinical Trials, December, 2005.
by S. Raymond
At the DIA Annual Meeting in Washington in 2005, Steve Raymond agreed to lead the group, which Becky Kush had previously included in her voluminous portfolio of responsibilities. In July 2005, the CDISC Glossary Group membership was canvassed and made current. It was also important to establish a clear definition of the Glossary Group charter in order to ensure delivery of the updated glossary (Version 4.0) for publication in the December 2005 issue of Applied Clinical Trials (ACT). With a reasonably broad representation in the membership achieved, including medical writers, regulators, NIH, and international clinical trialists, the Glossary Group revisited the Version 3.0 Glossary.
In addition, comments were harvested from the CDISC website, and helpful input was provided to the Glossary Group by ADaM, the PR Group, and Trial Design Group. A core team met in person in August in Flemington NJ, courtesy of Beardsworth Consulting, and by phone afterwards several times to review the comments, proposed terms, and historical terms. This review was completed just in time for publication of Version 4.0 in ACT and posting for comment on the CDISC website. Art Gertel and Helle-Mai Gawrylewski contributed a lot of time and expertise to make this publication possible. Our group’s objective during 2006 is to ensure that additional key terms are defined, as needed, for the various CDISC initiatives and that the Glossary is integrated into the clinical trial protocol models and applications.
It should also be noted that the CDISC Version 4.0 Glossary of eClinical Trial Terms and a listing of Acronyms, Abbreviations and Initials are available for comment on the CDISC website. The CDISC version presently matches the content published in ACT in December. During 2006, members are invited to promote new terms or to clarify or criticize the definitions in Version 4.0. When suggesting that a new word or abbreviation be added, please be sure to include a draft definition and any appropriate reference to current usage. To suggest that a term be deleted, please explain your position. The Glossary Group will review all comments and suggestions, and will post incremental updates for comment during the year.
|
|
|
|
by B. Kisler
The Terminology Team, which includes key members of the Submission Data Standards Team, continues to make progress in the development and release of terminology standards to support the SDTM (Study Data Tabulation Model). CDISC has two sets of clinical trials terminology in the works, known as SDTM Package-1 (30 codelists) and Package-2 (40+ codelists). Package-1 covers several SDTM classes: Demographics, Interventions, Findings, Events and Trial Design. Working closely with staff of the National Cancer Institute Enterprise Vocabulary Service (NCI EVS), Package-1 terminology has been loaded into the NCI Thesaurus (NCIt), one of NCI’s premier vocabulary products. Throughout February and March 2006, CDISC and NCI will work together to define the Common Data Elements (CDEs) and metadata structure, incorporating the terms in the NCI data standards repository known as caDSR, an ISO 11179 metadata repository. Communication and training for the CDISC user community is also being developed. Package-1 terminology will be made open source and globally accessible by April 2006.
SDTM Package-2 terminology includes additional codelists for SDTM domain models: Interventions (CM – Concomitant Medications, EX – Exposure , SU – Substance Use); Findings (EG - Electrocardiogram, PE – Physical Examination, VS – Vital Signs, SC – Subject Characteristics); and Events (AE – Adverse Events). Throughout the past 6-months terminology for Package-2 has been developed across terminology sub-teams each with focus on specific SDTM domains. As these smaller teams complete their tasks, the full CDISC terminology team was brought back together with a 2006 kick-off call in early January. All Package-2 terminology will be coalesced and posted on the CDISC website for comment and review by May 2006. Once finalized, Package-2 will also be incorporated into the caDSR later in the year.
In addition, the SDS team has finalized a list of laboratory analytes for LB-Laboratory Test Results Domain. This codelist includes the most commonly used analyte names for LBTEST with the short names LBTESTCD. The LBTEST codelist will be posted for comment and review on the CDISC web site in the near future. Following this process, it will be incorporated in NCI EVS along with Package-1.
One of the key priorities of the CDISC terminology initiative is to ensure broad adoption of terminology standards through a harmonized approach to vocabulary development. In this regard, CDISC is working with both national and international industry stakeholders and collaborating with the US FDA, National Institute of Health, the World Health Organization, Health Level Seven, as well as international pharmaceutical companies and service providers. Furthermore, CDISC is collaborating with academic institutions and has begun working with Duke Clinical Research Institute on several NIH Roadmap grants to develop data standards and terminology for several therapeutic specialties, including cardiovascular medicine and tuberculosis.
If you have an interest in contributing to the CDISC terminology initiative, please contact Bron Kisler at bkisler@cdisc.org.
by J. Diefenbach
A data model is only as good as its semantics, or so the experts tell us. Well-formed semantics not only help humans to communicate accurately, but they support the interoperability of computer systems.
The SDTM presents a robust and flexible structural framework for storing the data of clinical trials. In developing this model, the SDS Team began to specify the semantics by displaying sample vocabulary in the Comments column and examples. The Vocabulary Initiative picked up the pace of this effort with a dedicated project team to specify permissible value sets, largely for context variables, and will release these soon as
Package 1.
There is one piece missing; clinical vocabulary (terminology) is specialized and depends on the therapeutic area, and industry standards are lacking in terminology. The NCI has made strides to standardize the vocabulary of oncology, and now two NIH Roadmap grants were awarded to Duke Clinical Research Institute (DCRI) and involve CDISC for standards aim to do the same thing for cardiology and tuberculosis.
The Roadmap projects support overlapping teams comprised of representatives from Duke Clinical Research Institute (DCRI), CDISC, and HL7 that are following an HL7 process to develop semantic standards for cardiology and tuberculosis. This process begins with efforts to involve the stakeholders, and a regular set of meetings of the stakeholders in cardiology and tuberculosis has been established for this. The team is working to compile a list of vocabulary specific to these therapeutic areas from stakeholder contributions. With this list, we hope to accumulate a common set of frequently used vocabulary, or data elements, with consensus on the meanings from the stakeholders. HL7 Working Group Meetings will serve as a forum for vetting the process and the results. The project has initiated a new HL7 Special Interest Group (SIG) for Cardiology, which will also work with HL7 RCRIM. For the tuberculosis vocabulary, the project group will work through the HL7 Public Health Technical Committee. This work will be ongoing through 2006. Contact Jane Diefenbach (jdiefenbach@pharmastat.com) or the Cardiology SIG co-chair, Brian McCourt (brian.mccourt@duke.edu) to join this effort.
|
by P. Pochon
The CDISC Lab team’s efforts in 2005 have focused on minor enhancements to the base LAB model that reflect experience gained during implementation of the production model and expanding the model to include microbiology and pharmacogenomics data.
In early 2005, release 1.1 of the base LAB model was posted on the CDISC website for public comment. Based upon comments provided, and upon experiences gained in pilot projects, this model has been slightly updated, and will be released for production use in Q1 of 2006. The major enhancements in this model are expanded support for delta flags, and support for separate Conventional and SI LOINC codes. To support implementation of LOINC codes, the CDISC Lab team, in cooperation with the Regenstrief Institute, has published on the CDISC website a list of LOINC codes for tests commonly used in clinical trials.
In connection with release 1.1 of the base LAB model, the CDISC LAB team posted for public comment a heavily revised and updated reference range model. This model provides support for the delta and exclusion flag definitions, as well as an XML implementation. This model will go into production alongside the base LAB model in Q1 of 2006.
In Q2 of 2005, a microbiology extension to the LAB model was posted on the CDISC website for public comment. Lab team members have now completed two pilot projects with this model, and after harmonization with the CDISC SDTM Microbiology domain, this model will become a production extension to the LAB model in its ASCII, SAS and XML formats.
In the second half of 2005, a sub-group of the CDISC LAB team with other participants developed a pharmacogenomics extension to the base LAB model in its HL7 implementation. This pharmacogenomics message is intended for end-to end, collection-to-submission use. This HL7 message was balloted with the HL7 Regulated Clinical Research Information Management (RCRIM) in December of 2005, and will become available as a draft standard for trial use in mid-2006.
For comments and submissions please write to Sheila Camhi,
Managing Editor at scamhi@cdisc.org. |
 |
|
|
|
|
|
