The CDISC Lab Model: Saving Users Time and Money

November 1, 2005

Applied Clinical Trials

Applied Clinical Trials, Applied Clinical Trials-11-01-2005, Volume 0, Issue 0

Central laboratories are realizing significant cost and time efficiencies by using standardized data.

The Clinical Data Interchange Standards Consortium (CDISC) released its first production version of the CDISC Laboratory (LAB) Model in 2002. The model was designed for the standardized representation of laboratory data generated during the conduct of clinical trials. Estimates on the impact of the use of the standardized model indicate that costs are reduced 30%–50%. Standards for laboratory data have a significant impact, as it has been estimated that 60%–80% of clinical trial data originates in labs.

Over the past two and a half years, the LAB model has been adopted by pharmaceutical and biotechnology companies to facilitate the acquisition and interchange of data between laboratories, CROs, and sponsors. A majority of CROs and central laboratories, many of whom have been involved in the development of the model, are familiar with the model and are able to assist companies in its implementation.

Initially, inquiries to CROs and central laboratories about the LAB model came from smaller companies. Often, these were relatively new companies that did not have a pre-existing standard format for laboratory data or companies that had recently undergone a merger and were seeking to utilize the CDISC format for harmonization between the merged companies. More recently, large pharmaceutical companies have been inquiring about and have been implementing the CDISC LAB model as they make the significant investment necessary to update their systems. Although CDISC is frequently asked for lists of those companies that have implemented its models, permission for use of company names is problematic. However, a small survey of just four central laboratories found that 44 sponsor companies are currently utilizing a CDISC LAB model format for receipt of clinical trial lab data. A significant number of additional companies are discussing future use of the model for their clinical trials involving central labs. A full survey of central laboratories is not available.

Useful URLs and Documents

Some centralized laboratories interested in promoting adoption of the LAB Model have offered preferential pricing for those clients interested in using the CDISC LAB format for data interchange. Use of the predefined CDISC format saves laboratories considerable time and effort that has been necessary in the past to sort out data interchange with formats that vary from sponsor to sponsor, and often within sponsors. One central laboratory estimated that it has been required to support over 1200 different data formats at one time, including multiple formats for a single sponsor. Efficiencies can be gained using the CDISC format in its standard representation, which includes all 98 fields of the model, both required (# = 18) and optional (# = 80) data fields. The receiver of the transmission can then determine which of the transmitted fields are to be read into its database. In accepting the standard format, the laboratory is not required to invest time and effort in understanding client formats and supporting test transfers of data. Thus, some central laboratories are willing to pass along the savings to the client.

As a service for organizations that are looking for assistance in implementing CDISC standards, CDISC has established a program for Registered CDISC Solution Providers (RCSP) who are qualified consultants, system integrators, and subject matter experts believed by CDISC to have sufficient knowledge and experience implementing the various CDISC standards. To become an RCSP, a service provider must complete a registration form and submit CVs of all contacts, which must be reviewed and approved by the CDISC Technical Coordinating Committee and endorsed by at least one member of the CDISC Board of Directors. In addition, applicants must meet the following minimum criteria: adhere to the CDISC Code of Ethics, have a positive overall reputation for services in some aspect of clinical research, be involved in CDISC as a sponsor, member, benefactor, associate, or active team participant, and demonstrate expertise in one or more CDISC models. Information on Registered Solution Providers can be viewed on the CDISC Web site.

Multiple formats for widespread implementation

The CDISC LAB Model is available for adoption in four different formats: bar-delimited ASCII, SAS, and XML schema as well as a Health Level 7 (HL7) Version 3 Reference Information Model (RIM) compliant message, which has been approved as an American National Standards Institute (ANSI) standard. This approach was taken in recognition of the varying implementation formats in use in the industry. Initially, most inquiries and implementations involved the ASCII or SAS formats. However, as the industry has become more fully acquainted with XML over the past two years, there has been more interest in the XML schema. Of the 44 companies who are working with the central laboratories surveyed, 31 have implemented the ASCII or SAS format, six have implemented an XML format, and seven formats were not specified.

The LAB model is distinct from the other CDISC models, including the Operational Data Model (ODM) and the Study Data Tabulations Model (SDTM). The LAB model was created to address a specific set of use cases for the transmission of data between lab (either central or reference lab) to lab, CRO or sponsor company. As such, it accommodates administrative data, quality control data, and comments at several different levels of the model. These are data that are not usually incorporated into archived study data or submissions data. Where data elements are the same across models, the elements map to each other.

The CDISC LAB Team's effort at standardization of clinical trial lab data has included recommendations for valid value lists or code lists for some of the model fields. Use of standardized codes facilitates the exchange and pooling of laboratory test results, improves the reliability and accuracy of data interchange, and should permit higher level analyses of databases. The most obvious coding system for lab test codes is LOINC (Logical Observation Identifier Names and Codes). LOINC has been recommended by the eGOV initiative Consolidated Health Informatics (CHI) as a standard for U.S. government agencies. LOINC codes allow users to merge clinical results from many sources into one database for patient care, clinical research or management.

LOINC codes are recommended in the LAB model as a supplement to internal test codes and can, in fact, be used as the primary test identifier. Some pharmaceutical companies have already moved from proprietary internal test codes to LOINC. The LOINC database provides a set of universal names and IDs for identifying laboratory and clinical tests. To assist users in the clinical trial arena, the CDISC LAB Team has worked with the Regenstrief Institute, the developers of LOINC, in the creation of a subset of the full 45,000-term LOINC database. The 700-term subset of the Most Commonly Ordered Tests in Clinical Trials can be accessed using a search tool available on the LOINC Web site: the Regenstrief LOINC Mapping Assistant (RELMA). RELMA can search the database or a subset of the database for codes and can create extracts for bulk mapping. Use of LOINC codes should increase the efficiency of lab data transfers as common terminology decreases the need to learn, load, utilize, and maintain proprietary code lists.

Other external vocabularies that are referenced in the LAB Model include code lists from the International Organization for Standardization (ISO) and Health Level 7 (HL7). The LAB Model recommends the use of ISO 8601 for date/time formats and HL7 vocabularies are recommended for sex and specimen material.

The CDISC LAB Team has also worked to define some additional code lists for areas such as "specimen condition" where no existing standardized vocabulary could be identified. CDISC is working with the Enterprise Vocabulary Services (EVS) of the National Cancer Institute (NCI) to define and support vocabularies for clinical research. The NCI has created a CDISC context in its Cancer Data Standards Repository (caDSR), a server that NCI has offered as a location for the consolidation and centralization of various vocabulary efforts that are ongoing in the clinical research arena. By posting our vocabulary for specimen condition, for example, we make that vocabulary available to others in our field and related fields who may have an application for it.

The CDISC LAB Team has worked on two different types of standardization that would benefit our industry: standardization of structure or file formats (the CDISC LAB model), which ensures that data is structured in a consistent format so that both sender and receiver know where the data is and how to access it, and standardization of semantics or vocabularies (LOINC, ISO, HL7, and CDISC code lists) so that data from a variety of sources can be brought together with enhanced reliability and accuracy. Early implementers are already reaping the benefits of a standardized approach to data collection and interchange. Others are encouraged to investigate the CDISC LAB model. Version 1.1.0 of the LAB model will be posted shortly on the CDISC Web site.

Susan Bassion, PhD, is leader with the CDISC LAB Team, CDISC, 11324 Savin Hill Lane, Austin, TX 78739, (512) 288-6430, fax (512) 301-5480, email:

Related Content: