The management and use of the growing volume of clinical and claims data to navigate evolving regulatory initiatives such as the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) and the 21st Century Cures Act is leading many healthcare organizations to turn to reference data management (RDM). This technology provides the infrastructure needed to establish a single source of truth, enable interoperability, and optimize analytics for regulatory and value-based programs. RDM plays a vital role in normalizing your data, achieving semantic interoperability, and accurately representing a patient population for accurate quality measures reporting and analytics.
Recently, Dr. Brian Levy, vice president, Global Clinical Operations and Product Management at Health Language, took time to talk with Health IT Outcomes about the many benefits of RDM, the challenges you may encounter — as well as how to overcome them — when implementing it, and why it is important.
Levy has 20 years of experience in medical informatics, with particular expertise in development of terminologies and clinical content, and the use of the Internet by patients and physicians to improve care delivery. He leads a team of physicians, nurses, PhDs, and professional medical coders who maintain the terminologies, mappings, and other content in the Language Engine.
The management and use of the growing volume of clinical and claims data to navigate evolving regulatory initiatives such as the Medicare Access and CHIP Reauthorization Act of 2015 (MACRA) and the 21st Century Cures Act is leading many healthcare organizations to turn to reference data management (RDM). This technology provides the infrastructure needed to establish a single source of truth, enable interoperability, and optimize analytics for regulatory and value-based programs. RDM plays a vital role in normalizing your data, achieving semantic interoperability, and accurately representing a patient population for accurate quality measures reporting and analytics.
Recently, Dr. Brian Levy, vice president, Global Clinical Operations and Product Management at Health Language, took time to talk with Health IT Outcomes about the many benefits of RDM, the challenges you may encounter — as well as how to overcome them — when implementing it, and why it is important.
Levy has 20 years of experience in medical informatics, with particular expertise in development of terminologies and clinical content, and the use of the Internet by patients and physicians to improve care delivery. He leads a team of physicians, nurses, PhDs, and professional medical coders who maintain the terminologies, mappings, and other content in the Language Engine.
Q: What is reference data management and why should healthcare organizations care about it?
Levy: Reference data management (RDM) refers to data such as code sets used throughout an organization. In the healthcare domain, RDM includes standards such as SNOMED, RxNorm, and LOINC, as well as local code sets such as lab catalogs, patient cohort code lists, domain value lists, and more. RDM is the first step to establishing a master data management solution, providing the data management infrastructure needed to establish a single source of truth, improve interoperability, and optimize analytics for regulatory and value-based programs. Healthcare organization should care about this because RDM helps with complying with regulations such as MACRA and the 21st Century Cures Act. A cloud-based approach for managing reference data provides an easier path to enhancing existing analytics and data warehouse investments.
Q: What are some of the data sources that are necessary for healthcare reference?
Levy: Necessary data sources are clinical and claims data including standard code sets, as well as locally created domain values. Some examples of reference data include:
- Code sets such as ICD-10-CM, CPT, HCPCS, Zip Codes, country codes, gender, and marital status needed for the revenue cycle.
- Standard terminology codes used in the clinical domain to support Meaningful Use, including SNOMED, LOINC, medications (proprietary and RxNorm), and Medical Devices.
- Maps between terminologies to support analytics and billing, such as SNOMED to ICD-10-CM, SNOMED to CPT, and CPT to LOINC.
- Code groups, or domain value lists, to address quality measures, analytics, and population health. These code groups are subsets of the standard and local sets mentioned above.
- Synonyms to facilitate searching, indexing, and NLP. For example, clinicians use abbreviations and synonyms to record patient problem lists.
- Read-only data provided to operational and analytical systems.
- Departmental data such as the names of departments.
Q: What are the steps to achieving semantic interoperability?
Levy: Interoperability in healthcare starts with being able to send and receive patient information. Semantic interoperability involves actually understanding information such as problem lists, medications, and allergies. To understand this data it must be converted into the standards and domain values managed in the RDM solution. So, for example, local lab codes would need to be converted into LOINC, disparate medication codes into RxNorm, and problems into SNOMED. If the data being received from another health system is already using the standards, RDM can help store all of the code sets needed to understand and validate this receiving data. To achieve semantic interoperability, you must be able to gather all the necessary data you need, then normalize and make sense of it. This can be a long-term, tedious task without the right tools and expertise. The latest versions of each code set should be available, but previous versions also must be used to reconcile patient information documented in the past.
Q: What are some obstacles to overcome in attempting to normalize data?
Levy: Data can be codified, structured but un-codified, or completely unstructured. Eighty percent of healthcare data is locked away in unstructured data. Healthcare organizations need the ability to capture all this data and encode it to help them make more informed business decisions for better overall patient care. All this data are coming from disparate sources — EMRs, claims, clinician notes, lab results, etc. Being able to reconcile where the data is coming from, its context, and the needed normalization presents significant obstacles.
Q: Can you give us an example of why data normalization/interoperability is so important?
Levy: MIPS will have significant impacts on provider reimbursement — up to 9 percent plus or minus in a couple of years. MIPS fundamentally changes the payment model to enforce value-based care and hospitals, provider networks, and payers will be affected. A fundamental aspect to MIPS is quality measure reporting which will be required for the Advanced Payment Models and the standard MIPS. Proper identification and reporting of patients for the appropriate quality measures will require normalized data from structured and unstructured sources.
Q: What are key benefits — clinical and financial — to embracing RDM?
Levy: There are myriad standard and local terminologies and code groups that will need to be accessed as part of the RDM solution. The even more challenging aspect to RDM is these data sets constantly change over time. So a strong RDM solution can handle the updating and versioning of these data to ensure access to the most recent codes as well as historical codes that were previously recorded. RDM feeds many other applications within the enterprise, including Master Data Management systems, data warehouses, and reporting/analytics tools.
Q: Why are healthcare organizations struggling to deploy infrastructures that support complete, accurate capture of data and ongoing management of those assets and how does RDM help overcome those struggles?
Levy: I believe this is a struggle because there is so much data now captured in the healthcare world, and it has become so difficult to make sense of it all. Plus, there are continuous changes and updates to all the standards and healthcare terminologies. An RDM solution ensures you normalize all of your data and that it is the most up-to-date and accurate data.