Healthcare Terminologies and Classifications: An Action Agenda for the United States

Although there has been significant progress in the United States to better understand the role terminologies play in our health information systems, and to make terminologies more broadly available in machinable forms, more must be done to ensure that this progress serves as a robust foundation for the information content of the EHR. This report describes some of the significant efforts that have led to this progress and some challenges that require action. It proposes the formation of a Centralized Terminology Authority, and other steps, to address these challenges.

Signs of Progress

Over the past several years the national health IT agenda has addressed this critical component to a successful system. Major positive steps have been taken, resulting in foundational work, momentum, and a platform for additional progress.

For example, in July 2003 the National Library of Medicine (NLM) licensed SNOMED CT for five years, making it available without charge throughout the US via the Unified Medical Language (UMLS).2 This NLM agreement presented a new opportunity to resolve issues in a more open environment. The NLM also supports the development and ongoing maintenance of RxNorm, a standard medication terminology system that may promote interoperability of medication-related services.3

A more recent development changing the landscape significantly and holding great promise is the evolution of SNOMED CT ownership from the College of American Pathologists to an International SNOMED Standards Development Organization (SDO). Already having a SNOMED license has placed the US in a favorable position in the talks regarding the structure of the SDO. Plans are also under way for the SNOMED SDO to have at the very least a standing group working on the coherence between SNOMED CT and international classifications, a requirement for ensuring that terminologies and classifications work together.

One final example of a positive step and the formation of foundational work for additional progress is the Healthcare Information Technology Standards Panel (HITSP) established by the American Standards Institute. Under contract with the Office of the National Coordinator, HITSP is working on harmonizing the standards required to enable the secure exchange of patient data across the system.4

More to Be Done

The American Medical Informatics Association (AMIA) and the American Health Information Management Association (AHIMA) are pleased with the progress made thus far and are committed to the development of the emerging nationwide health information system. Both organizations are taking this opportunity to bring forth yet unresolved issues to offer recommendations for the major challenges that still remain and to create a process that is sustainable over time and results in interoperability. Additional resources and appropriate funding are needed to build on the foundation established for the healthcare terminologies and classifications strategy, governance, and development and maintenance processes.

There is still much to be done to take advantage of opportunities begun by the NLM, HITSP, and the soon to be formed SNOMED SDO. Fragmented governance, proprietary licensing, uncoordinated release cycles, and the lack of available standards currently stand in the way of advancing terminologies and classifications to meet our nation’s future healthcare needs. For example, tighter integration between terminology and classification is needed to drive processes. There also needs to be ongoing investment by the federal government in mapping and maintenance.

To address the remaining issues, AHIMA and AMIA convened a Terminology and Classification Policy Task Force composed of medical and nursing informaticians, nosologists, and health information management professionals and educators well-known in the terminology and classification field. The Task Force has formulated a vision and associated goals and recommendations that it hopes will be used to frame a public-private dialogue about how to redesign the US approach to healthcare terminologies and classifications against a backdrop of international approaches and achievements.

The vision consists of the following:

  • US governance occurs from a national perspective against a backdrop of international agendas.
  • US policy coordinates with other countries, and the US actively collaborates and shares costs.
  • Coordination and collaboration occurs with international terminology and classification development and maintenance initiatives.
  • Terminologies, classifications, and maps form a coherent set of policies and procedures for openness and assured performance.
  • There is transparency of process even when the development organization maintains the system within its own organization.
  • The infrastructure for development and maintenance of the terminology is subject to an open process.
  • Business process automation is implemented, allowing organizations to participate and track the terminology and classification development processes, reducing cost, and automating many aspects of the system release cycles.

To implement this vision, the Task Force recommends that the healthcare industry—government, public and private institutions, and professional organizations—collaboratively undertake the following tasks:

  • Create a publicly funded research and development project to prepare specifications for coordinated solutions and where possible, consolidate terminology. The critical elements of the plan relate to policy, governance, standards adoption, legal issues including licensing, technology, data integrity requirements, maintenance, education, and conformance testing, and a roadmap for change so public and private industry sectors understand the goals, target, and required actions.
  • Secure funding for the planning and development of a centralized authority, representing both public and private stakeholders, to manage the funding and be responsible for overseeing US terminology and classification development and maintenance, including the supporting systems. The public/private central authority would define the governance structure for the US role in the SNOMED Standards Development Organization (SDO) and would oversee the SNOMED national release center in the US. Part of these funds would be used to support healthcare standards in ways that do not rely on revenue generated by selling the standards themselves. The authority will identify the certification standards for terminologies and classifications in the EHRs and EHR systems and oversee the development of implementation guides. It must have sufficient authority to coordinate efforts among the various agencies and stakeholders to assure congruency and collaboration, not competition.
  • Develop a governance model for the central authority that is accountable to the needs of the end users and implementers, and also has accountability for the funding of the central authority. It will pursue recommendations that emerge from the research and development project outlined above. Australia’s model, where the central authority is funded by the state, territorial, and national governments, and where the organization is governed by chief executives from health departments within these jurisdictions, has a number of desirable features. However, governance must also formally include some care providers and other end users. It need not be a total government responsibility; a quasi-governmental entity might meet this need quite adequately if permanence of the structure can be assured.
  • The entity(ies) will commit to the adoption of sound principles for operation of a terminology and classification standards development organization. Primary among the principles is an infrastructure for development and maintenance that includes a transparent process with policies and procedures that can support open standards and assure that those who build content are engaged in any consolidation or integration of their products.

The Task Force believes that action on this agenda is an urgent need. To this end, the two organizations are ready to lead in concert with the appropriate government agencies the US effort for terminologies and classification reform. AMIA and AHIMA are willing to assist with the ongoing work that will be needed, including discussions within the federal government about US representation and involvement in the international SDO. It recognizes the US agenda for health information reform will require strong public and private collaboration and continuity of resolve. In the months to come, AHIMA and AMIA will convene stakeholders to build a broader understanding of the current problems, generate wide support for, and begin to construct a roadmap for change.

Background

Clinical terminologies and classification systems have been in use for many years—for example, the International Classification of Disease (ICD) was created in the 19th century as a way to identify the causes of death. The Systematized Nomenclature of Medicine (SNOMED) began as an enhancement to the Standard Nomenclature of Diseases and Operations, first published in 1965 and used to organize information from surgical pathology reports.

In addition to their lengthy history, terminologies and classifications have different use cases. Certain systems are appropriate for specific applications. For example, terminologies are used primarily to capture clinical information. As such, they are highly detailed and have substantial granularity but at the same time they lack reporting rules and guidelines.

Classification systems are intended for secondary data use, including quality of care measurement, reimbursement, statistical and public health reporting, operational and strategic planning, and other administrative functions. While reporting rules and guidelines for administrative code sets exist, compliance is not at an acceptable level.

Systems such as ICD-9-CM and CPT are used to organize specific diseases and procedures in a general classification schema. This allows the specific diseases and procedures to be grouped into more broad-based categories and then used for reimbursement (ICD-9-CM codes to DRGs), quality of care measures (ICD-9-CM codes to patient satisfaction rankings), or resource utilization measures (CPT codes to cost of care). Because classification systems are considered broader ways to classify specific diseases and procedures, they are not the most appropriate system to use to annotate and aggregate the clinical aspects of an episode of care.

In ICD-9-CM and ICD-10-CM, for example, a hospital stay for a newborn is represented by codes identifying the type of birth (i.e., live-born infant, single or multiple births, born in hospital or outside of hospital, and vaginal versus cesarean delivery) and medical conditions or risk factors the infant may have. In contrast, SNOMED CT® codes provide the complete clinical detail for a healthcare encounter. To represent the APGAR score, a commonly used scoring system for neonatal health and viability, alone would be represented by separate SNOMED CT codes for the cardiac score, respiratory score, muscle tone, reflex response, and color.5

Together, terminologies, such as SNOMED CT, and classification systems, such as ICD-9-CM, ICD-10-CM, and ICD-10-PCS, provide the common medical language necessary for the EHR and for population health reporting, quality reporting, personal health records, safety, clinical trials, biosurveillance, and reimbursement.

Terminologies, Classifications, and the EHR

A robust universe of health-related terminologies and classifications identifying the clinical data is required, and simply having a list of relevant words is not the same as representing all the relevant data in a computable form. Thus, while coherent systems of terminologies and classifications are essential they are not sufficient, especially in the context of data and information representation. The process of data or information representation begins with information models, not just the terms and codes alone.

For clinical data representation what is needed is a library of shared models that are linked to value sets drawn from standard coded terminologies and classifications.

Both the National Committee on Vital and Health Statistics (NCVHS) and the Institute of Medicine in their reports noted that standard representation of the full meaning of patient medical data requires integrating terminology models with models of context and other structural relationships.6, 7

Terminologies, Classifications, and Mapping

The NCVHS also concluded in its report when patient medical record information standards do not include a comprehensive information model and terminology model, mapping is made more difficult. Mapping is a process that links the content from one terminology or classification scheme to another.8

There are many factors within the healthcare industry driving mapping technology, including the movement to adopt EHRs and create a nationwide health information network (NHIN).9 The trend of increased administrative costs within healthcare is also a factor. The manual coding process is expensive and inefficient, and when combined with a shortage of qualified coders the use of computer-assisted coding and mapping technologies is sought for gains both in productivity and consistency. Today, computer-based coding and mapping systems are lacking but are an achievable vision for the future.

An EHR that can map from a reference terminology to a classification system is rare. There are currently no nationally recognized standards for map development and validation, and there is a lack of coordination among terminology and classification development organizations which inhibits the development of maps. There is also no standard definition of a “valid” map. To arrive at these objectives, iterative research and development is needed along with testing in real environments to achieve a usable map.

The NLM has taken a lead in this important area, and has made maps from SNOMED CT to ICD-9-CM, SNOMED CT to CPT, and a LOINC to CPT preliminary draft are available from the NLM. These maps can serve as a starting point for the development of standards and for testing the utility of different mapping approaches.

Terminologies, Classifications, and the NHIN

 

In addition to the EHR, clinical terminologies and classifications are important to the NHIN. An NHIN and the interoperable exchange of health information between standard electronic health records of healthcare data depend on:

  • Clinical content identified with clinical terminologies, which maintain highly specific and descriptive patient care data and classifications for aggregate reporting that can be reliably captured, encoded, exchanged, and interpreted
  • Common information models
  • Management and coordination of the various systems in a synchronized fashion for a streamlined life cycle

 

Magnitude of the Problem

The terminologies and classifications life cycle—in particular, development, distribution, and maintenance—is complex. This complexity causes problems to all who deal in some form or another with them. There is virtually no interoperability between commercial environments and the limited market size makes commercial offerings challenging and creates a high risk of vendor lock in10 and significant susceptibility to vendor failure. In addition, no comprehensive open source development environment exists, resulting in authoring environments that constrain responsiveness. The outcome is the creation of content in the system based on a defined need rather than desirable characteristics of a controlled clinical terminology.11 Current terminology efforts typically do not meet rudimentary development practices considered routine for software development in other industries or subject domains. Terminology systems also probably do not meet Food and Drug Administration standards for embedded software—yet many should.12

Another facet making the current situation difficult is the lack of business process automation across organizations or departments for updating or maintaining the terminology.13 Even a single organization is required to use multiple tools if it is to function today. Appendix B illustrates the complexity for four healthcare organizations.

Adoption and implementation are in themselves complicated issues with many governing bodies determining what system should be adopted and implemented. In addition, while there have been discussions about the basic principles for maintenance and what the ideal process should look like, attributes and standards for terminology development and maintenance processes have not been brought forth, discussed, and agreed to by the principal stakeholders. Simply finding out what systems already exist, what they are used for, when they were adopted and implemented, who maintains them, and what the release cycles are can be an onerous task.

Today’s proprietary standard development models hamper development and maintenance of terminology standards by prohibiting open publication of content. Terminology systems are significantly lacking in any principled approach to validation and quality assurance.

Despite its essential need, insufficient funding has also plagued the development and maintenance of clinical terminologies and classifications for years. The bulk of the work is done on volunteer time.

Other problems widely acknowledged include:

  • the lack of sufficient incentives to cooperate
  • systems are seen as independent when they are highly interdependent
  • a general lack of understanding of how encoded data can support initiatives such as quality or patient safety in the healthcare environment

The Vision and Goals for the US

The Task Force’s determination of the ideal state for terminologies and classifications takes into account key issues including governance, licensing, and the methods used in the terminologies and classifications life cycle—specifically, development, distribution, and maintenance.

Under the auspices of international sharing of solutions to the terminology and classification issue, the vision for terminologies and classifications consists of the following:

  • US governance occurs from a national perspective against a backdrop of international agendas.
  • US policy coordinates with other countries, and the US should actively collaborate and share costs.
  • Coordination and collaboration occurs with international terminology and classification development and maintenance initiatives.
  • Terminologies, classifications, and maps form a coherent set of policies and procedures for openness and assured performance.
  • There is transparency of process even if the development organization maintains the system within its own organization.
  • The infrastructure for development and maintenance of the terminology is subject to an open process.

 

The characteristics of this process include:

  • Any classification algorithms required to transform terminology stated forms into inferred forms are publicly described and free of patent royalties so that any party may implement a terminology-capable classifier unencumbered.
  • Interfaces to the terminology development infrastructure (automated term submission process, automated update services, terminology configuration management services, and terminology authoring services) are publicly described and licensed under an open source agreement such as the Apache 2 license.14 An open source license is meant to ensure that any party can freely develop software that can interact with or replace components of the terminology development infrastructure unencumbered by licensure fees, patent royalties, or from other licensing restrictions that may require commercial organizations to disclose their proprietary source code. A license that requires a vendor to open their source code, such as the GNU style licenses, is inappropriate.
  • Any fees associated with mandated infrastructure necessary to implement the standard are under regulatory control to assure that fees to support the infrastructure are not diverted.
  • Business process automation that allows organizations to participate and track the terminology development process, reducing cost and automating many aspects of the terminology release cycle is put in place.

The goals of a coordinated policy for the US that enables the Task Force vision builds upon lessons from national and international efforts in three areas:

  1. governance
  2. development and maintenance, and
  3. acquisition, licensing, and implementation issues.

Governance

The Task Force identified the following goal for terminologies and classifications governance:

  • Establishment of a centralized oversight authority of the various terminologies and classification systems to ensure consistency in processes, system coordination, responsiveness to end users, and the availability of robust, valid maps. Centralized oversight is also necessary for adoption of uniform rules, regulations, and guidelines for standardized terminology and up-to-date classification systems across the country. A central authority would also ensure that the organizations authorized to develop, deploy, and maintain such standards and guidelines assume ongoing responsibility to provide clarity with a specific standard or guideline as required, publish and disseminate the standards or guidelines in a manner that is generally understood, and respond in a timely manner to all requests for clarification of standards or guidelines.

Simplifying Governance

Meeting this goal will require changes in the way in which the US governs terminologies and classifications. In the US, there is no centralized oversight of terminologies and classifications. Terminologies and classifications have their own governance process dictated by the organization who “owns” the system. Table 1 provides a summary of the owners and governance processes for various terminology systems currently used in the US. In many cases, the governance process is strictly an internal function of the terminology development organization with little detail available publicly. In a few cases, the governance process includes oversight by a committee composed of users and other stakeholders. The fragmentation illustrated in Table 1 prohibits interoperability. The number of systems in existence and the resulting acronym alphabet soup make apparent the existing complexity of governance.

An important consideration when choosing to integrate a component such as a terminology is to understand the governance model of the organization providing the component either via license or through contribution to the public domain. For example, the problem with third-party control is particularly challenging for SNOMED CT because the governance of SNOMED CT development and licensing policies has been subject to approval by the CAP board of governors, not by the end users or licensees of the SNOMED CT terminology standard. In contrast, HL7’s board of directors is elected by the HL7 members, and is fiducially bound to serve their interests. The proposed formation of a new SNOMED SDO outside of CAP most likely will resolve this important governance issue.

Fragmentation of the governance process is also illustrated by the current situation with ICD. While the National Center for Health Statistics (NCHS) maintains the US clinical modification if ICD-9-CM and ICD-10-CM for diagnosis coding, the Centers for Medicare and Medicaid Services are responsible for development and maintenance of the procedure coding systems. While subject to a single Coordination and Maintenance Committee, coordination of these two essential dimensions could and should be strengthened.

Recognizing that the US could learn much from other countries that have already redesigned their governance and support processes for terminologies and classifications, the approaches taken by Canada, Australia, and the United Kingdom were studied. Details for each country are found in Appendix A.

Development and Maintenance

The Task Force identified the following goals for terminologies and classifications development and maintenance:

  • The infrastructure for development and maintenance of the terminology is subject to an open process. The characteristics of this process include:
  • Any classification algorithms required to transform terminology stated forms into inferred forms are publicly described and free of patent royalties so that any party may implement a terminology-capable classifier unencumbered.
  • Interfaces to the terminology development infrastructure (automated term submission process, automated update services, terminology configuration management services, and terminology authoring services) are publicly described and licensed under an agreement at least as open as the Apache 2 license.
  • Any fees associated with mandated infrastructure necessary to implement the standard are under regulatory control to assure that fees to support the infrastructure are not diverted.
  • Robust methods for qualifying terminology systems for use in patient-care applications are developed. These robust methods are patterned after those developed for other software components, whose failure may result in loss of life, such as the DO-178B “Software Considerations in Airborne Systems and Equipment Certification.” Adapting that specification for terminology requires the terminology development organization to provide at least the following:
    • Certification Plan
    • Development Plan
    • Verification Plan
    • Configuration Management Plan
    • Requirements Standards
    • Design Standards
    • Verification Cases and Procedures
    • Verification Results
    • Problem Reports
    • Configuration Management Records
    • Quality Assurance Records

Need for a Standard Process

In comparison to the above goals, presently the healthcare industry has no standard process for the development of terminologies and classifications. Typically, the processes for terminology and classification development vary, generally depending on the size of the development organization and the purpose of the standard. This variability is evident in Table 2.

Many systems are developed by professional associations and academic institutions, while others are developed by government agencies. Some organizations follow relatively formal sets of processes for developing and refining concepts, such as reliance on an editorial board for approval, whereas other organizations’ processes may be informal. In addition, developers may create new terms at will without a formal or open process with input from the users of the system.

Regarding maintenance, each organization or developer determines their own maintenance process, and update schedules vary considerably. The level of resource commitment to ongoing system maintenance also varies, which in turn, influences the maintenance process and the frequency of updates. In some cases, maintenance is handled entirely by the system developer, while in other cases, a committee of industry experts oversees the maintenance process.

Meetings to address maintenance issues may be open to the public or may be private. For those processes that include open meetings, final decisions regarding system modifications may be made privately rather than during the open meeting. Some processes allow submission of requests for modifications from any source, whereas others place restrictions on requesters. As shown by the variability in the level of detail outlined in Table 2, terminologies even vary in the level of transparency regarding their maintenance processes.

There is also considerable variability in the timing of the release of terminology or classification system modifications and their effective dates. As Table 2 shows, some terminologies do not have an established update schedule. And variability in the method of publicizing the modifications is an issue as well. For example, some systems are updated annually, others biannually, and still others are updated quarterly or even more frequently. Even for those systems that are updated annually or biannually, the effective dates still differ. The variability in maintenance schedules is not purposeful or useful, as it increases the cost and complexity of using terminology and classification systems and keeping them up-to-date.

In some cases, a change in the system update schedule would require legislation. For instance, federal law mandates the implementation of ICD-9-CM updates on October 1 and CPT updates on January 1 for the use of these systems in the Medicare program. There is no central Web site for accessing system updates or an established, standardized process for notifying the healthcare industry of an update release.

The “official” source of clarification for using a particular terminology or classification and the process for obtaining an “official” answer are not always clear. The integrity of coded data and the ability to turn it into functional information require the use of uniform standards, including consistent application of standard codes, code definitions, and reporting requirements. In Table 2, the system developer is generally listed by default as the official source for clarification because there is no other identified source. In many cases, there is no clearly defined process for obtaining clarification. As a result, obtaining guidance on the proper use of a terminology or classification is sometimes difficult, time-consuming, and expensive. To add to the complexity, there may be more than one official source for a system—for example, the AMA is recognized as the source for information on the proper use of CPT codes for physician providers; however, the AHA is the official clearinghouse for hospital providers.15

Acquisition, Licensing, and Implementation

The Task Force identified the following goals for terminologies and classifications acquisition and implementation:

  • A licensing model that embraces the following:
    • More inventiveness, not the promulgation of monopolies (even nonprofit or regulatory ones) that stifle innovation and hamper collaboration,
    • Open standards to overcome the challenges, and
    • Standards development organizations which better serve the healthcare community needs by providing open standards
  • Business process automation that allows organizations to participate and track the terminology development process, reducing cost, and automating many aspects of the terminology release cycle is put in place.
  • Establishment of funds for healthcare standards in ways that do not rely on revenue generated by selling the standards themselves.

Acquisition and Licensing

The way in which licensing is done today is a far cry from the goal. Similar to the systems discussed in the previous section, healthcare standards that structure terminologies and classifications are developed by various organizations. These organizations are typically not for profit; however, the standards they produce are almost always released under proprietary licenses along with fees—even though the efforts to produce those standards are often provided by resources external to the standards organization itself.16

Take the case of HL7. It has placed all intellectual property of their standards into the public domain; however, HL7 retains the copyrights to published standards and sells the documents as a revenue-generating mechanism to help sustain the organization. The HL7 license requires that the specification documents must either be purchased or be available only to members who have limited ability to redistribute them to their customers. This inability to redistribute specifications has had a direct influence on the development and distribution of open source implementations of HL7 specifications, since the specifications of those implementations cannot be distributed to a group of potential implementers who are not HL7 members. Open posting of development content is thereby not allowed.

Even LOINC, which charges no license fee for allowed uses, has specific use restrictions.17 The LOINC Committee has retained the copyright to its content so that it maintains editorial control of the content. It prohibits the LOINC content to be used to make a competing coding system for observations.18 These LOINC restrictions prevent development and/or enhancement of other products in ways that may promote interoperability because the other products may be regarded as “promulgating a different standard” and therefore is prohibited by the LOINC license. For example, it would be a violation of the LOINC license to map LOINC into an existing terminology by assigning a new identifier to the LOINC code and creating the appropriate linkages to the concepts in the existing terminology. The inability to create linkages in this way is a barrier to creating true interoperability between LOINC and other important clinical terminologies.

The SNOMED CT license is not standardized, and different organizations have different rights and responsibilities depending upon their particular license. The US National Library of Medicine (NLM) version of the SNOMED license prohibits distribution of SNOMED content outside the US, and it also prohibits modification of SNOMED “core content.” These restrictions make open collaboration around SNOMED-based content challenging for some and untenable for others. Many organizations are reluctant to engage in such collaboration when a third party is able to exercise unilateral control over content they may develop a dependency upon—even when that content is based upon their own work. The new SDO may remove some of the barriers to broader community participation in the enhancement and expansion of SNOMED CT.

There are two types of CPT licenses, depending upon whether CPT codes will be used in a product that will be sold or distributed to others or whether CPT will only be used within a company and will not be redistributed. In both types, the license fee structure is per user. A user is defined as an individual who directly accesses CPT data in a product or, in the case where CPT is embedded in a product and not directly accessible, relies on embedded CPT data to perform his or her intended function with the product or its output. These license requirements increase the cost of products containing CPT content and inhibit widespread use of CPT. Also, determining the number of CPT users for the purposes of license fee calculation can be difficult.

Even the NLM’s Unified Medical Language System (UMLS) Metathesaurus, which serves as an official distribution vehicle for HIPAA standard code sets and US government target clinical vocabulary standards, requires a license. While content may be used with few restrictions, some vocabularies require separate agreements, which may involve fees, with the individual vocabulary producers.19

There are various justifications for resorting to proprietary licenses. These justifications range from “we must get revenue to support development of the standard” to “we must prevent fragmentation of the standard.” Unfortunately, consequences of relying on proprietary licenses include deliberate or inadvertent creation of monopoly positions by standards development organizations and licenses that stifle opportunities for innovation.

Experience with open source software and the World Wide Web development have shown that licensing restrictions are not helpful—and are also not necessary to prevent fragmentation of the standard. The HTTP protocol and the success of the World Wide Web consortium (W3C) are successful and compelling demonstrations of the potential of alternative approaches. All W3C documents are under an open license that provides permission to copy or distribute, in any medium, for any purpose, without fee or royalty.20 W3C operations are supported by a combination of member dues, research grants, and other sources of public and private funding. W3C operations do not rely on revenue from the specifications of the standard itself.

Eben Moglen, professor of law at Columbia Law School, states that open source software provides a “freedom to invent, not reinvent—not invent over again something someone else had invented and locked up, but invent in the way that inventing was done in the great spurt of 19th-century inventiveness.”21

The recent ASTM/HL7 dispute over ownership of ideas contained within the ASTM Continuity of Care Record (CCR) and the HL7 Clinical Document Architecture (CDA) is an example of challenges that are created when standards organizations do not publish their works under open licenses. The focus should be on innovation and building on the foundations provided by others. Standards organizations should not be permitted to compete with each other for proprietary advantage.22

Implementation

Implementing terminologies and classifications is not an easy task. Distribution is part of what makes it difficult. Terminologies are usually distributed through the developer or a middle-ware software or system application vendor. With more than 100 terminology and classification systems, the communication of releases and updates affects the ease with which vendors can incorporate them into products for distribution.

For example, the NLM’s UMLS has been designated as the central coordinating body within HHS for patient medical record terminologies.23 Coordination means the following:

  • Uniform distribution of designated standard vocabularies through the UMLS Metathesaurus
  • Reducing peripheral overlap and establishing explicit relationships between standard clinical vocabularies (e.g., SNOMED, LOINC, RxNorm)
  • Aligning standard clinical vocabularies with standard record and message formats
  • Mapping between standard clinical vocabularies and administrative code sets and/or other important vocabularies24

While the UMLS was envisioned as a distribution mechanism for clinical terminologies and classifications, it is not a major distributor because of publication delay and difficulty in extracting data formats. In addition, the UMLS is not governed by a mechanism that is directly accountable to implementers and EHR developers, which may help explain why they are going elsewhere for terminology content. Implementation therefore is made complicated as most venders go to the source, or owner of the system, because of delay and complexity of user formats.

Recommendations

The Task Force recommends the following action items:

  1. Create a publicly funded research and development project to prepare specifications for coordinated solutions and where possible, consolidate terminology. The critical elements for the plan include policy, governance, standards adoption, legal issues including licensing, technology, data integrity requirements, maintenance, education and conformance testing, and a roadmap for change so public and private industry sectors understand the goals, target, and the required actions.
  2. Secure funding for the planning and development of an entity that would serve the public interest, for example, a centralized authority representing both public and private stakeholders, to manage funding and be responsible for overseeing US terminology and classification development and maintenance policies.25 The central authority would define the governance structure for the US role in the SNOMED SDO and would oversee the SNOMED national release center in the US. Part of these funds would be used to support healthcare standards in ways that do not rely on revenue generated by selling the standards themselves. The authority will identify the certification standards for terminologies and classifications in the EHRs and EHR systems and oversee the development of implementation guides. It must have sufficient authority to coordinate efforts among the various agencies and stakeholders to assure congruency and collaboration, not competition.
  3. Develop a governance model for the central authority that is accountable to the needs of the end users and implementers, and also has accountability for the funding of the central authority. It will pursue recommendations that emerge from the research and development project outlined above. Australia’s model has a number of desirable features. However, governance must also formally include care providers and other end users. It need not be a totally government responsibility—a quasi-governmental entity might meet this need quite adequately if permanence of the structure can be assured.
  4. The entity(ies) will commit to the adoption of sound principles for operation of a terminology and classification standards development organization. Primary among the principles is an infrastructure for development and maintenance that includes a transparent process with policies and procedures that can support open standards and assure that those who build content are engaged in any consolidation or integration of their products.

Objectives of the Planning Grant

The main objectives of the planning grant are to continue the work of the Task Force in partnership with the Office of the National Coordinator (ONC). AHIMA and AMIA will also coordinate with Healthcare Information Technology Standards Panel (HITSP) to harmonize work. The work will initially focus on the task of helping determine how to handle US participation in the SNOMED SDO followed by further research to identify a governance model and funding sources for the central authority.

In addition, moving from the current ineffective and fragmented strategy for adoption and maintenance of terminologies and classifications to one capable of meeting the requirements outlined herein will require a series of actions and a long-range change strategy. Therefore, an additional outcome of the work from the planning grant will be a roadmap for change so public and private industry sectors understand the goals, target, and required actions.

The Centralized Authority and the SNOMED SDO

The Task Force applauds the recent global progress on the formation of the SNOMED SDO and is fully supportive of this development. Having US involvement as a charter member of the SDO will require an entity to be responsible for overseeing US interests. The Task Force believes the central authority would best serve this role. The central authority would define the governance structure for the US role in the SDO and would oversee the SNOMED national release center in the US, where questions regarding the process for requesting of new terms or establishing regional extensions would be determined. With the aid of planning grant funds mentioned above, the US participation process in the SNOMED SDO will be defined.

The Centralized Authority and Healthcare Information Technology Standards Panel

A centralized authority responsible for overseeing US terminology and classification development and maintenance, including the supporting systems, is critical. The Task Force supports the Healthcare Information Technology Standards Panel (HITSP) basic standards readiness criteria as a starting point for the work of this centralized authority. Tier 2 criteria take into consideration issues the Task Force identified as necessary for ridding the development, implementation, and maintenance process of its complexity and to structure governance in a coordinated fashion.

These criteria were developed to screen the potential candidate standards, which include terminologies and code sets. They are used by the HITSP and its technical committees for evaluating standards for possible use in the HITSP Interoperability Specifications. Tier 2 criteria are placed into the following categories: 26

  • Suitability: the standard is named at a proper level of specificity and meets technical and business criteria of use case(s)
  • Compatibility: the standard shares common context, information exchange structures, content or data elements, security, and processes with other HITSP harmonized standards or adopted frameworks as appropriate
  • Preferred Standards Characteristics: approved standards, widely used, readily available, technology neutral, supporting uniformity, demonstrating flexibility and international usage are preferred
  • Standards Development Organization and Process: meet selected criteria, including balance, transparency, developer due process, stewardship, and others
  • Total Costs and Ease of Implementation

Challenge to the Stakeholders

There are many challenges for the healthcare industry—government, public and private institutions, and professional organizations—with regard to moving forward. A collaborative undertaking will be the key to success along with the following:

  • A Terminology and Classification Interoperability Fund. The fund estimated to be from $5-50 million will be used to structure a public-private effort to adjudicate and set policy for various stakeholders to follow. These funds should be placed in one location to do ongoing monitoring, research, and evaluation results to address:
    • Quality of encoded data
    • Quality of encoding systems
    • Value of encoding data
    • Patient safety issues pertaining to encoded data
    • Reproducibility of encoded data
    • Quality, reproducibility, value, and patient safety pertaining to mapping and to different versions of encoded data in the same system
    • Development and recommendations of standards for coded data (i.e., for computer-assisted coding)
    • Biosurveillance
    • Use health informatics data to improve public health practices as well as medical readiness

Responsibility for the fund will fall to the central authority that will coordinate the funding of research, standards, grant review, and contracts. This authoritative body will focus on processes for governing, maintaining, and distributing the terminologies and classifications. Its responsibility will be to align funding with policy and guide the contracts/grants process toward long-term goals.

  • A terminology group made up of both public and private organizations to assist the central authority. This group would be contracted and therefore not dependent on “voluntary” commitment. It will be responsible for establishing the standards used for certification of terminologies and tools.
  • Simplified coding guidelines and reimbursement regulations so that mapping rules can be more readily developed and maintained.
  • Use case maps from reference terminology to administrative code sets (for instance, SNOMED-CT to ICD-9-CM, SNOMED-CT to CPT, and LOINC to CPT) that are validated with an easy, no-cost distribution mechanism.
  • Educational resources to train healthcare professionals on the use and interpretation of coded data and its relationship with clinical terminologies, classification systems, and mapping technologies. Since much of this relationship is still misunderstood and somewhat difficult to fully grasp, educational resources that clearly and succinctly describe the mapping process and the importance of reference terminologies and classification systems is truly needed.

Role of AHIMA and AMIA

AMIA and AHIMA are seriously committed to a global approach to health-related terminologies and classifications and seeing the recommendations in this report move forward. Both organizations are prepared to engage with others to accomplish the following:

  • Develop a planning grant. This proposal will be broad based and be in partnership with the Office of the National Coordinator (ONC). AHIMA and AMIA will also coordinate with HITSP to harmonize their work with the plans outlined in the grant.
  • Engage informatics and information management leaders in increasing the awareness of the interdependencies of encoding data with terminologies, classifications, and with issues of reimbursement. It is important to understand how encoded data can support other initiatives such as quality or patient safety in the healthcare environment.
  • Enlist the support of other key organizations in these efforts, such as the HHS/Centers for Medicare and Medicaid Services (CMS), including the ONC, large integrated healthcare organizations like Kaiser Permanente, Veterans Health Administration, National Institutes of Health, including the National Cancer Institute and the National Library of Medicine, foundations such as the California Healthcare Foundation, large insurance organizations such as Blue Cross/Blue Shield, and others such as the Certification Commission for Healthcare Information Technology.
  • Create educational resources to train healthcare professionals on the use and interpretation of coded data and its relationship with clinical terminologies, classification systems, and mapping technologies.
  • Broadly publicize these recommendations to those whose support is needed to adopt the roadmap.

Conclusion

The members of the Task Force believe the state of terminologies and classifications in the US is ineffective and in disarray and necessitates immediate action. The healthcare industry – including the government, professional organizations, public and private institutions, and health informatics and information management professionals – must respond to the chaos and address the issues before the US healthcare system is marginalized. To not do so would mean continued reliance on poor-quality data for decision making and the spending of dollars to retrofit a system that is obviously broken. If the changes outlined in this paper are not made, it is unlikely that the goals of the nationwide health information agenda can be achieved.

Moreover, without question there is a need for worldwide comparability of healthcare data to improve the effectiveness of global public health policies and programs. To meet this need a variety of health-related terminologies and classifications identifying the clinical data is required. However, just having these systems available does not mean the data can be used and exchanged. The various systems must be coordinated and managed in a synchronized fashion. Processes for developing, releasing, implementing, and maintaining terminologies and classifications must be streamlined. They must have an effective life cycle infrastructure to improve quality and simplify reliable and maintainable health data. The recent global progress with SNOMED gives special urgency to this matter for the US. We must get our collective house in order; the lack of a coherent healthcare system in the US and the lack of prospects anytime soon for a coherent system cannot be allowed to prevent progress in this extremely important domain.

Achieving the goal of a well-organized, coordinated health-related terminologies and classifications governance structure and efficient development, implementation, and maintenance processes will take time. The industry is challenged to continue forward while new approaches are being designed, tested, and funded. An industry dedicated to prompt response and collaborative work can build solutions for today’s turmoil and ensure better health information management for the nation and the world.

The American Health Information Management Association (AHIMA) is the premier association of health information management (HIM) professionals. AHIMA’s 50,000 members are dedicated to the effective management of personal health information needed to deliver quality healthcare to the public. Founded in 1928 to improve the quality of medical records, AHIMA is committed to advancing the HIM profession in an increasingly electronic and global environment through leadership in advocacy, education, certification, and lifelong learning. To learn more, go to http://www.ahima.org/.

The American Medical Informatics Association (AMIA) is the premier organization in the United States dedicated to the development and application of medical informatics in the support of patient care, teaching, research, and healthcare administration. AMIA links developers and users of health information technology, creating an environment which fosters advances that revolutionize healthcare. To learn more, go to http://www.amia.org/.

©2006

Notes

1. For purposes of this report, an EHR is defined as an information system designed to provide access to complete and accurate clinical data, practitioner alerts and reminders, clinical decision support systems, and links to medical knowledge. Giannangelo, K. (Ed). Healthcare Code Sets, Clinical Terminologies, and Classification Systems. AHIMA. [Note: Within this definition, a personal health record (PHR) only qualifies as an EHR if it is part of an EHR having the identified capabilities.]
2. See http://aspe.hhs.gov/sp/NHII/News/NHIIJul1_03.htm.
3. See http://www.nlm.nih.gov/research/umls/rxnorm/index.html.
4. Available at http://www.ansi.org/news_publications/news_story.aspx?menuid=7&articleid=1284.
5. Dougherty, M. “Standard Terminology Helps Advance EHR.” Journal of AHIMA 74, no. 10 (2003): 59-60.
6. National Committee on Vital and Health Statistics. Patient medical record information terminology analysis reports. 2002.
7. Institute of Medicine. “Key Capabilities of an Electronic Health Record System.” Letter Report of the Committee on Data Standards for Patient Safety, Board on Health Care Services, Institute of Medicine, 2003.
8. Available at http://www.nlm.nih.gov/healthit.html.
9. Available at http://www.nlm.nih.gov/healthit.html.
10. Wikipedia states vendor lock-in is a situation in which a customer is so dependent on a vendor for products and services that he or she cannot move to another vendor without substantial switching costs, real and/or perceived.
11. Cimino, J. “Desiderata for Controlled Medical Vocabularies in the Twenty-First Century.” Methods of Information in Medicine 37 (1998): 394-403.
12. Many of the regulations enforced by the Food and Drug Administration (FDA) with regard to medical devices can be found in Title 21 Code of Federal Regulations (CFR) Part 800 to Part 1299. The information required in a 510(k) submission is defined 21 CFR 807.92. Among the list of 510(k) submission requirements is software development, verification and validation information.
13. Wikipedia states business process automation is the process of integrating enterprise applications, reducing human intervention wherever possible, and assembling software services into end-to-end process flows.
14. Available at http://www.apache.org/licenses/LICENSE-2.0.html.
15. Available at http://www.ahacentraloffice.org/ahacentraloffice/html/coding_advice/hcpcs/index.html.
16. Wikipedia states that proprietary indicates that a party, or proprietor, exercises private ownership, control or use over an item of property, usually to the exclusion of other parties.
17. See http://www.regenstrief.org/medinformatics/loinc/.
18. Available at http://www.regenstrief.org/loinc/license/. To prevent the dilution of the purpose of the LOINC codes and LOINC table of providing a definitive standard for identifying clinical information in electronic reports, users shall not use the RELMA program, RELMA Users’ Manual, RELMA database, LOINC Users’ Guide, LOINC database, LOINC table or related files, and/or the LOINC codes for the purpose of developing or promulgating a different standard for identifying laboratory test results, diagnostic study reports, clinical measurements and observations or orders for these entities in electronic reports and messages.
19. Available at http://www.nlm.nih.gov/pubs/factsheets/umlsmeta.html.
20. Available at http://www.w3.org/Consortium/Legal/2002/copyright-documents-20021231.
21. Available at http://news.com.com/Open-source+politics+are+American+as+apple+pie/2100-7344_3-6079401.html.
22. Available at http://www.healthcareitnews.com/story.cms?id=1292 and http://www.healthcareitnews.com/story.cms?id=3076.
23. Available at http://www.ncvhs.hhs.gov/040922lt.pdf.
24. Available at http://www.nlm.nih.gov/nichsr/2005_NCVHS_July.ppt.
25. Available at http://www.jamia.org/cgi/content/abstract/5/6/503.
26. Standards Readiness Criteria, Tier 2, Version 1.0
27. National E-Health Transition Authority, http://www.nehta.gov.au/content/view/1/103.
28. Available at http://www.govhealthit.com/article94797-06-12-06-Print).
29. Connecting for Health, http://www.connectingforhealth.nhs.uk/technical/standards/.
30. Ibid.

AHIMA and AMIA Terminology and Classification Policy Task Force Members

Keith E. Campbell, MD, PhD
Chair, AHIMA and AMIA Terminologies and Classifications Policy Task Force
Chief Technology Officer, Informatics, Inc., and Assistant Clinical Professor; Department of Medical Informatics and Clinical Epidemiology, Oregon Health and Science University

Suzanne Bakken, RN, DNSc, FAAN
Alumni Professor of Nursing and Professor of Biomedical Informatics School of Nursing and Department of Medical Informatics, Columbia University

Sue Bowman, RHIA, CCS
Director of Coding Policy and Compliance, American Health Information Management Association

Christopher Chute, MD, PhD
Professor and Chair of Biomedical Informatics, Mayo Foundation

Don Detmer, MD, MA
President and Chief Executive Officer, American Medical Informatics Association

Jennifer Hornung Garvin, PhD, RHIA, CPHQ, CCS, CTR, FAHIMA
Medical Informatics Postdoctoral Fellow Center for Health Equity Research and Promotion, Philadelphia Veterans Administration Medical Center

Kathy Giannangelo, MA, RHIA, CCS
Manager, Practice Leadership, AHIMA

Gail Graham, RHIA
Director, Health Data and Informatics Department of Veterans Affairs

Stanley M. Huff, MD
Professor of Medical Informatics, University of Utah Senior Medical Informaticist, Intermountain Healthcare

Linda Kloss, MA, RHIA, FAHIMA
Chief Executive Officer, AHIMA

Margaret Skurka, MS, RHIA, CCS
Professor and Director of Health Information Programs, Indiana University Northwest

Mary Stanfill, RHIA, CCS, CCS-P
Vice President, Practice Resources, AHIMA

Valerie Watzlaf, PhD, RHIA, FAHIMA
Associate Professor, Department of Health Information Management, School of Health and Rehabilitation Sciences, University of Pittsburgh

 

Article citation: Perspectives in Health Information Management, Summer 2007

Printer friendly versino of this article

Posted in:

Leave a Reply