Creation of LOINC Equivalence Classes

Creating the infrastructure to allow disparate systems to interpret and exchange patient information.
Agency
  • National Institutes of Health (NIH)
Start Date
  • 8/31/2016
Functionality
  • Standardized Collection of Standardized Clinical Data

 

STATUS: Completed Project

BACKGROUND

Interoperable health information exchange relies in part on common vocabulary standards that allow disparate systems to interpret and exchange information with the same clinical meanings. Logical Observation Identifiers Names and Codes (LOINC), developed by the Regenstrief Institute, Inc., is a universal coding system for laboratory tests and other clinical observations. However, mapping local terms to LOINC codes can be complex and resource intensive.

PROJECT PURPOSE & GOALS

The goal of this project was to create a flexible, extensible, and computable mechanism for rolling up LOINC codes into clinically relevant equivalence groups. Under the NIH, this project enables more efficient processing and aggregation of laboratory and other data from diverse health information technology (IT) systems.

Project Objectives:

  • Identify high priority and clinically relevant content subsets for representation in the new LOINC hierarchy. To support this work, the team developed an enhanced software tool that searches the LOINC database, and stores the subset for later retrieval. Input on the high priority subsets were gathered from an expert advisory group and the global LOINC user community.

  • Develop a clinically relevant roll-up hierarchy for LOINC terms. The LOINC team reviewed the candidate subsets to identify “equivalence groups” and define the term attributes necessary to create the group.

  • Disseminate the aggregation hierarchy within the main LOINC release distribution. The new hierarchy was published with each bi-annual LOINC release and is available publicly at no cost.

PROJECT ACHIEVEMENTS & HIGHLIGHTS

  • The team created several releases of the LOINC Equivalence Class artifact. The first version, LOINC Group File Alpha 1, was released June 2017 and contains 12 Parent Groups and 2,178 Groups that aggregate a total 6,438 unique LOINC terms. LOINC Group File Alpha 2 was released December 2017 and includes the revised format and ancillary data such as molecular weight. It contained 21 Parent Groups, 4,100+ equivalence groups that aggregate more than 17,500 unique LOINC terms. The final version was released June 2018 as LOINC Group File Beta 1 and contains 36 parent groups, containing 5,650 groups that organize 24,075 unique LOINC terms. It also includes groups in several new areas, including radiology terms by region imaged, document ontology groups based on the setting, several groups based on a broader clinical concept, such as social determinants of health and physical activity. Inside the LOINC Group File are Usage Notes and component-level molecular weights from the PubChem database for nearly 900 Groups.

  • To support this initiative, the team created a webpage/portal for submitting LOINC equivalence groups that allows users to share groups of LOINC codes they are using, and to submit new groups for addition to LOINC. The portal was demonstrated to the LOINC Committee in December 2017 and is publically available.

  • The NIH project team met with representatives from the Fast Healthcare Interoperability Resources (FHIR) community to create FHIR value set representations of the LOINC Groups. Making the LOINC Groups available as FHIR value sets supports their uptake and availability within the broader health IT community.

  • To foster ongoing input from the LOINC community, the team developed a portal for users to recommend and submit groups. The team has received substantial contributions of additional content, which were incorporated in subsequent LOINC releases, including additional groups of microbiology, drug testing, smoking, pregnancy, and pain terms.

PUBLICATIONS, PRESENTATIONS, AND OTHER PUBLICALLY AVAILABLE RESOURCES

Resources:

Presentations:

  • The third release of the equivalence class artifact and example groups was presented during the June 2018 Laboratory LOINC Conference. Handouts from the LOINC Committee Meeting are available here: https://loinc.org/conference/summer-2018/meeting/

  • A presentation at the FHIR® DevDays Conference focused on the API capabilities for accessing LOINC content, including LOINC groups. The full educational session slides and a detailed tutorial of exercises is available here: https://danielvreeman.com/fhir

  • A presentation for the Clinical Laboratory Improvement Advisory Committee (CLIAC) in April 2018 focused on improving lab data interoperability including use of the LOINC equivalence groups. Slides from this presentation are available here: https://danielvreeman.com/presentation/advancing-lab-data-interoperability-with-loinc/

  • The project team presented the systems demonstration “LOINC Groups: A novel tool for data aggregation” at the AMIA 2018 Annual Symposium in San Francisco, California in November 2018.

RELATED PROJECTS

Below is a list of ASPE-funded PCORTF projects that are related to this project

Development of Data Infrastructure for Use of EHRs in Comparative Effectiveness Research (ONC/NLM CDEs) The CDE project of ONC and NLM was developed to standardize the collection of standard data with its focus on creating standardized terms for the collection and exchange of data, i.e., CDEs. For the electronic case report forms (eCRF), the project developed the standards for the eCRF structure/template as well as the CDEs that were used with them. These forms were ‘extensible forms’ which means that they were developed in a high-level computer language. This computer language facilitates defining the form as a single, stand-alone object. Taken together, these two ONC/NLM CDE products make evident that the project fits the functionality of Standardized Collection of Standardized Data.

Harmonization of Clinical Data Element Definitions for Outcome Measures in Registries The Agency for Health Research and Quality (AHRQ) administered this project to convene a series of clinical topic-specific working groups to discuss the various definitions for outcome measures currently in use and how these definitions can be harmonized into a common set to be used across data collection and reporting systems. This project aimed to standardize clinical data collection; by adopting the core common outcome measures/data element sets, the data available for research was standardized and comparable across different registries and health care organizations that collect the information. This project developed a consensus set of clinical data element definitions that can be consistently used to represent specific outcome measures for each of five clinical topic areas: atrial fibrillation (AF), lumbar spondylisthesis (lower back pain), lung cancer, asthma, and depression.