Jan09

Healthcare Terminology

Monday, 09 January 2012

Healthcare Terminology: A Year in Review

Initiatives funded through the Office of the National Coordinator on Health Information Technology (ONC)1 in 2011 have put a bright light on healthcare IT, primarily through the specifications and standards underpinning ONC's stage I and II (and beyond) meaningful use (MU) requirements. 2011 also saw attention on the requirements to migrate from the decades-old ICD-9-CM to ICD-10 (CM and PCS). Both MU and the ICD migration highlight the need for better processes with improved terminology management.

Healthcare facilities and vendors studying MU with an eye to Stages beyond Stage I encounter incomplete standards -- especially standards defining the use of healthcare terminology. Work done by entities such as the Health Information Technology Standards Panel (HITSP)2 that was charged with harmonizing multiple standards, and the National Quality Foundation (NQF)3 that has launched a quality measure re-tooling initiative to represent key quality measures in a format consistent with information found in an EHR, have helped reduce standards gaps and provide core elements for terminology-specific MU requirements. While MU requirements focused on certification of health IT tools4 have moved forward in 2011, true usability has remained a work in progress.

During 2011 there was considerable focus on getting MU Stage I certified Health IT systems into the hands of providers and it became clear that installation was necessary, but not sufficient, to garner use. In fact the Institute of Medicine (IOM) recent report "Health IT and Patient Safety: Building Safer Systems for Better Care"5 notes that "safer use of health IT involve the people and clinical implementation as much as the technology." The "use" in MU means safe and effective use of systems to improve patient outcomes and enhance the clinical environment. This means providing systems that address activities of patient care, that are important to health care and doing so in a way that integrates into standard clinical workflow processes.

Current Steps to Interoperability
The past year also brought renewed focus on consistency in data collection. Every drop-down list and every set of available test, procedure, diagnosis or billing code that is available for use in HIT systems is driven by code or word lists. To date, most installations of HIT systems create those lists during the implementation process. More recently the code lists arrive with the install as part of the typical start-up set and are modified based on local needs. While appearing "standardized", these modified lists result in very little consistency and where consistency does exist, is usually restricted to organizationally-related installs.

Now that the industry is far enough along the "hype cycle"6 to begin to understand that "doing terminology right" is an important factor in getting value from the HIT investment, there is growing interest in standardized terminology. The looming terminology-centric activities required to move to ICD-10-CM (and HIPAA 5010) have resulted in resources focused on terminology and coding. Add to that HITECH funding and Meaningful Use requirements that identify specific additional terminology expectations, and you have unique convergence that requires "terminologic awareness." MU functionality requirements in stage II and beyond make clear that defined sets of specific standardized vocabulary elements will be required for data capture and patient reporting. This will become even more critical when systems must integrate additional requirements for data consent and privacy.

A better solution to the proliferation of dissimilar code lists is the use of value sets. Value Sets7 are sets of defined vocabulary concepts8 that are specifically tailored to represent the information needed for patient care (and often also reported out for quality measures and billing). For instance, pneumococcal pneumonia is the English-language term for the idea of this disease, which can be exchanged between computers using the SNOMED CT concept identifier "350052012". Requirements, such as meaningful use requirements, decision support systems, quality measure assessments, billing requirements, and even localized "requirements" such as drug formulary lists, referral lists and available procedure choices, all exist as value sets. Sharing and aggregating patient data is dependent on using common value sets, but where can users get access to these value sets?

Part of the problem is that many of the descriptions of "expected codes" for quality measures, reporting requirements, and decision support systems exist only in written documents, and extracting the information to create a formal value set is a substantial undertaking. The National Quality Forum (NQF) is working to "re-tool" endorsed quality measures into a format that supports direct access of data from an EHR.9 It is clear that this type of work will take time and once accomplished, the resulting specific value sets will need to be reviewed, maintained, and updated over time. This level of work effort suggests that a trusted, centralized organization be responsible.

Recognizing the importance of getting to the details and then subsequently determining the gaps, ONC and CMS have begun to define specifications and then try to apply those specifications to the work at hand.10 Yet despite major pushes from standards groups, including the HIT Policy and Standards Committees and the HITSP Vocabulary Work Group, little has happened to increase the availability to implementers and providers of value sets needed for health care interchange and reports standards.

Even so, some bright spots do exist. The United States Health Information Knowledge Base (USHIK),11 has created discrete value sets to support healthcare standards-primarily those for HITSP and HL7. The National Cancer Institute (NCI) Thesaurus continues to make available via the Enterprise Vocabulary Services (EVS)12 sharable vocabularies primarily focused on cancer research, but NCI has made the infrastructure available for many other kinds of value sets. So examples do exist, but the critical need for commercially viable, explicitly-defined value sets remains.

While few "end user" health care practitioners will search the Web for sources of the latest standardized terminologies and value sets, they will certainly be interacting with them through the EHR interfaces they employ. In 2011 health care providers began to realize that MU requirements-from care coordination and collaboration, to information exchange and population health management-call for a common, consistent language/vocabulary that allows for efficient, ongoing communication and information exchange among individual professionals and healthcare organizations (HCOs). Implementers began to understand the importance of vetted, up-to-date, reliable value sets tuned to support national requirements (standards, quality measures, reporting requirements, billing expectations) that allows them to "document once, and use many times". In the concluding article of this series, we will preview next year's expected actions from national healthcare groups and suggest what steps provider organizations should take to prepare for them.