Semantic Interoperability - a key component of data sharing
Electronic health record (EHR) adoption, Meaningful Use, the universal adoption of the ICD-10 coding system, the Physician Quality Reporting System and other federal initiatives all have something in common. They all seek to increase the efficiency and reliability of data sharing across the healthcare industry.
By making use of data transport and syntactic standards such as HL7 which define messaging structure, todays HIT maturity has gone a long way in establishing the foundational elements required for interoperability amongst disparate IT systems. Healthcare organizations are now turning their attention to semantic interoperability — the ability for IT systems to understand the meaning of the data that is being shared. Two local drug codes, for instance, may describe the same drug in different terms. An information model that normalizes clinical terminologies to standards is a critical requirement in enabling semantic interoperability across a health system. Modern HIT operations are beginning to use data dictionary’s as their ‘single source of terminology truth’.
What is a Data Dictionary?
A data dictionary is a set of information describing what type of data is collected within a database, its format, structure, and how the data is used. In many respects, a data dictionary can be thought of as the rules in which all the data within your system need to abide by. If all of your systems are producing data that follow the same rules - you achieve semantic interoperability.
The dictionary can provide a list of names, definitions and data elements to be captured in the system and includes metadata—or additional information—about each of those elements. Metadata is a way to organize data at its most basic level and helps in distilling large amounts of data for specific purposes. The use of metadata will become increasingly important as large volumes of information become available from the increased use of HIE systems like EHRs. So much new information would have little value if it couldn’t be processed and analyzed dependably.
Data dictionaries should be created with federal standards to support HIE with Meaningful Use in mind. As HIE use increases, AHIMA warns that healthcare organizations will need to properly identify data elements for appropriate reporting and transmission.
A successful data dictionary can improve the reliability and dependability of an organization’s data, reduce redundancy, improve documentation and control, and make it easier to analyze data and use it to make evidence-based care decisions like those common in accountable care organizations.
Requirements for Semantic Interoperability
In addition to simply modeling the terminologies that will be used within your health system, an enterprise class data dictionary must also include various "translation" rules that disparate IT systems can access if they need to normalize their data to ensure compliance. When thinking of data dictionary requirements in the context of semantic interoperability, the following questions should be asked:
- Does your data dictionary properly model all relevant standard terminologies used within your health system? (e.g. ICD-10, ICD-10, SNOMED CT, RxNorm, LOINC, CPT)
Does your data dictionary contain maps between various terminology code sets? (e.g. ICD-9/ICD-10 to SNOMED CT, ICD-9/ICD-10 to CPT)
Does your data dictionary define code groups which can be used to represent a single clinical concept? (e.g. the ICD-9-CM, ICD-10-CM, and SNOMED CT codes defining patients who have a history of myocardial infarction)
Does your data dictionary contain a synonym library of provider (e.g. “ank fx”) and consumer friendly (e.g. “nosebleed”) terminologies and their mappings to standards.
A Single Source of Terminology Truth - data governance and semantic translation
The data dictionary can be managed within a spreadsheet or table. However, healthcare providers, payers and vendors quickly find that to be scalable, purpose-built data dictionary tools are needed even for simple mapping - including mapping algorithms for taking unstructured text and ‘translating’ it to abide by the rules within your data dictionary. In addition, the complexity of managing updates from the standards bodies can be a daunting task. Depending on the depth of your data dictionary, aggregating and updating information via spreadsheets can be time consuming and very difficult to maintain.
Are you using a data dictionary to enforce semantic interoperability? Are you using a ‘home-grown’ solution or a terminology server?