Code Map Services

The Code Map Services project was in part funded through Office of the Assistant Secretary for Planning and Evaluation (ASPE) for the Patient-Centered Outcomes Research Trust (PCORTF). This project is phase III of the Common Data Model Harmonization (CDMH) project. Through the use of common data models (CDMs), scientists conduct analysis of real-world data which is critical for healthcare and clinical research. CDMs standardize the structure, format, and terminologies for data, so researchers can collaborate, share data, and conduct analyses across different datasets. Code Map Services provide researchers with an automated tool to transform (machine to machine) data collected and stored among CDMs and/or data standards:

• The National Patient-Centered Clinical Research Network (PCORnet)
• The Observational Medical Outcomes Partnership (OMOP)
• FDA Sentinel Initiative
• Health Level Seven International (HL7) Fast Health Interoperability Resources (FHIR) US Core Profiles V6 (based on United States Core Data for Interoperability USCDI)
• Informatics for Integrating Biology & Bedside / Accrual to Clinical Trials (i2b2/ACT)
• TriNetX
• Clinical Data Interchange Consortium (CDISC) Study Data Tabulation Model (SDTM)

Code Map Services establish a source of truth within the Cancer Data Standards Registry and Repository (caDSR, a shared metadata mapping repository) by using standards-based semantic annotations for establishing model-to-model equivalents. Mappings can evolve and be versioned over time, thereby improving and expanding the mapping functionality with each release of new models.

This project will harmonize semantics across multiple CDMs and data standards to provide researchers in federal agencies and academia with access to interoperable real-world data from various data sources (such as electronic health records and administrative claims) and from different research networks. The data elements in these models will be registered in caDSR as Common Data Elements (CDEs), which include semantic concept annotations. These annotations make the meaning of each data element more usable by computers.

Code Map Services advance the ability to conduct high-quality and efficient patient-centered outcomes research and other types of research studies by making data elements in different data models semantically interoperable. Along with publicly available transformation rules, caDSR will enable transformation tool vendors to provide services to the CDM communities to extend the accessibility of their data to other communities, vastly expanding the impact of their work.

The caDSR II features support a collaborative effort among experts to enable these advanced new features:

Public users can: 

• View models and their detail using Model View/Browse, as described in Viewing and Browsing caDSR Models. 
• View model mappings using Model Mapping View/Browse, as described in Viewing and Browsing caDSR Model Mappings. 
• Download/export content to Excel, as described in "Delivery Options Menu QR" (https://service.cancer.gov/cadsr-curation?id=nci_ext_kb_article&sysparm_article=NCI-KB0014777).

Model owners can:

• Register models, as described in Registering each Model. 
  • If the CDEs that you need do not already exist, request them by contacting the caDSR RA caDSR.RA@mail.nih.gov. 
  • Refresh the associated CDE, as described in Refreshing the CDE Associated with a Model.
• Maintain models and model versions, as described in Maintaining Models. 

Model experts can:

• Create model mappings, as described in Mapping Models. 
  • Maintain model mappings, as described in Maintaining Model Mappings. 
  • Manually add a narrative description of the rule, as described in Revising Rows: Derived From.  
  • Generate pseudocode, as described in Generating Pseudocode.   
  • Generate semantic mapping rules (transformation rules) between two models, such as semantically equivalent and semantically similar, as described in Generating a Semantic Map and a Value Map. 
  • Review and manually create rules where fields are not semantically aligned, as described in Reviewing Transformation Rules and Updating Transformation Rules. 
  • Generate value mappings based on the CDEs, as described in Generating Value Mappings.  
  • Review and manually adjust values not mapped, as described in Reviewing and Updating Value Mapping.  
• Enter provenance information into the model mapping about subject matter reviews of the rules and value mappings, as described in Capturing Provenance Information for a Model Mapping. 
• Capture provenance, as described in Capturing Provenance Information for a Model Mapping. 
• Generate code to be used by an ETL (extract, transform, load) tool and use a pre-parser to customize generated code. 

ETL programmers can implement the rules themselves from the following registered information, or use generated code depending on the available registered resources:

• Text based mapping rules are provided by model owners. 
  • You can use these to write your own data transformation program. 
• Generated semantic mapping rules
  • These are generated by the system based on the concept annotations for fields that share the same CDE or the concept annotations (share the same semantics). 
• Optional structured rules can be created by model mapping experts for fields that are not semantically aligned. 
  • Pseudocode can be generated for all structured rules making review of the rules for accuracy and approval easier for model owners. 
• Value mapping tables are generated for semantically similar fields that have enumerated variables or variables referencing external standard terminologies.
  • You can generate these based on CDE concept annotations that establish the meaning of each permitted data value.
  • The mappings can be edited by model mapping experts to map "orphaned" data values that had no matching value in the target model.  
• Limited NCI Metathesaurus / NCI Thesaurus terminology cross-walk mapping tables for data coded using a subset of selected standard terminologies
• The code can be generated as needed after the rules have been structured, as described in Generating Code. It is provided as-is for you to customize. Once you have accepted it and revised it, then we are no longer responsible for what you've done to it.
• Rules retrieved via:
  • caDSR graphical user interface for review using "Export to Excel" and the Download Collection feature
  • Application Programming Interfaces (APIs) and a File Transfer Protocol (FTP) site for retrieving model information and mappings

If you have questions, refer to the caDSR II Metadata Management Knowledge Base:

• Home page: https://service.cancer.gov/cadsr-curation?id=nci_ext_kb_homepage
• Curation Glossary: https://service.cancer.gov/cadsr-curation?id=nci_ext_kb_article&sysparm_article=NCI-KB0015054
• caDSR II and ISO 11179: https://service.cancer.gov/cadsr-curation?id=nci_ext_kb_article&sys_kb_id=a04edf101b5f9910932da8efe54bcb96