Abstract
Objective: The objective of the 1999 Nursing Vocabulary Summit Conference was to seek consensus on and a common approach to the development of nursing terminology standards for use in information systems.
Methods: A four-day invitational conference brought together authors and representatives of responsible organizations concerned with the nursing terminologies recognized or under consideration by the American Nurses Association, along with experts on language and standards and representatives of professional organizations, federal agencies, and the health informatics industry.
Results: Participants distinguished between colloquial terminologies and reference terminologies, and between information models and terminology models. They agreed that most recognized nursing terminologies were colloquial terminologies and that a reference terminology was needed. They formed task forces to develop and test aspects of a reference terminology model prior to a second meeting in June 2000, at which they would determine readiness to collaborate on a single international standard.
Discussion: The 1999 Nursing Vocabulary Summit Conference changed the level of discussion about nursing vocabulary standards from a debate about the relative merits of the various terminologies recognized in the United States to an examination of methods for developing and testing a reference terminology model and, eventually, a reference terminology that could serve as an international standard.
From the early days of medical informatics, nurses have recognized the need for standardized language. Werley, an advocate for the use of clinical data to study and improve care processes, wrote, “…before you can computerize nursing information systems, the variables must be identified”1 In 1974, Bartoszek decried the fragmentation of computer-based patient records by problems of language, design, and style.2 She urged nurses to contribute to the development of the Systematized Nomenclature of Medicine (snomed).
Challenged by Werley to create a nursing minimum data set,3 because “medical” terminologies did not adequately represent nursing practice, a number of nurses developed classification systems and sets of terms to represent diagnoses, interventions, outcomes, and goals. By 1999, these sets were competing with one another for incorporation into health care information systems, but none had been accepted as a standard. In June 1999, Vanderbilt University convened the Nursing Vocabulary Summit Conference with the objective of reaching consensus on the required characteristics of a nursing terminology standard and on the means of developing such a standard. This paper gives an overview of the conference process and conclusions. It provides a context for the other papers in this section, which have been developed in the follow-up to the conference.
Background and Significance
Nurses provide hands-on and supervisory care to sustain life, enable recovery, alleviate suffering, manage symptoms, facilitate self-care, and promote health. These services, vital to the well-being of the public, are an important component of ambulatory care and are the reason-for-being of hospitals, home care agencies, and long-term care facilities. Yet institutional and public health databases contain hardly a trace of data about these services, other than the costs of salaries and benefits. In the absence of data about the nature and effectiveness of nursing services, quality of care is assessed by negative measures, such as frequency of adverse events, and by indirect measures, such as patient satisfaction. Both are important, but they are not adequate. Nursing is more than a matter of averting harm, and few patients have the training to appraise sophisticated interventions.
Without data about the problems and goals nurses address or the actions they take, nurses' contributions to individual and collective outcomes are easily overlooked. With little available evidence, positive or negative, of the impact of nursing, administrators can decrease resources allocated to nursing with no apparent ill effects—because no effects are measured.
The invisibility of nursing in health care databases results not from ill will toward nursing but from the lack of standardized data about nursing in patient records. The variety of services that nurses perform—including fundamental physical care, sophisticated therapeutics, exquisitely nuanced monitoring and adjustment of care, individual and family counseling and education, psychiatric care, and community development—has mitigated against easy definitions of nursing practice or nursing phenomena. As a result, documentation of nursing care in patient records has been idiosyncratic and unstandardized. Although nurses spend from 25 to 60 percent of their time putting information into the patient record,4,5 the names they give problems, goals, interventions, and outcomes vary among patients, nurses, times, and settings. Missing data, unreliable data, and incomparable data make it impossible to include nursing information in databases for quality improvement and health services research.
Yet the decisions of policy-makers, managers, and clinicians need to be guided by knowledge of the nature and effectiveness of nursing care. Such knowledge could be discovered from data abstracted from the patient record if the record contained valid, reliable, standardized data about problems, goals, interventions, and outcomes. Since the advent of electronic information systems in the 1970s, nurses have proposed a number of sets of standardized terms to represent these phenomena.6–12 The American Nurses Association (ANA) has recognized ten of these sets to date, and others are under consideration. Still other standardized terminologies for nursing are under development internationally.
For a while it was thought that one of these terminologies might emerge as a de facto standard or that together the recognized terminologies might become a unified standard.13 Neither of these events has come to pass. The nursing terminologies have been criticized for lacking a number of the desiderata for controlled medical vocabularies, such as concept orientation and concept permanence.
Moreover, these terminologies were developed for different purposes, with different levels of understanding of what is required of a language system if it is to represent concepts and relationships in intelligent computer-based patient records.14–16 With the exception of snomed-rt,17 the standardized languages recognized by the ANA have not been constructed from formal models that clarify and make explicit the semantic relationships among concepts. Thus, although standardized data about nursing practice could facilitate the creation of knowledge to improve care and enhance outcomes, efforts to create a standard had not, as of June 1999, succeeded.
Methods
The 1999 Nursing Vocabulary Summit conference was an effort to increase shared understanding of problems of standardized language for nursing and to seek convergence on a common goal for a standard that would transcend any single terminology.
Participants
The composition of the invitation list was critical. The conference was small—fewer than 40 participants—but all those who were invited had a vital role to play. Every competing U.S. terminology was represented by its primary author or by the head of the sponsoring organization. Experts on language and standards, many affiliated with key standards-developing organizations, brought their knowledge and experience to the discussions. Leaders of professional organizations made sure that the results would serve their members. Representatives of federal agencies (the National Library of Medicine and the Division of Nursing of the Health Resources and Services Administration) spoke about the role and the limitations of government in the development, maintenance, and use of standards. Members of health care agencies reminded others of the requirements and constraints affecting their acquisition and use of clinical information. Finally, representatives of health care information system vendors spoke about the needs of their industry and their clients.
The varied expertise and perspectives informed all discussions. Moreover, because participants were highly placed in the organizations they represented, they could speak with authority. When questions arose, for example, about the consistency of a proposed approach with ongoing work at snomed or Health Level 7, a member of the editorial board or the chair of the relevant committee was present and could respond immediately.
Process
Before the conference, the organizers sent participants a briefing book containing information about the various nursing vocabularies and about the development of terminology standards. Work at the conference began with vocabulary authors providing updates on their recent work. Then experts on standards development introduced principles to guide subsequent work. Participants grappled with typologies of terminologies and approaches to modeling. On the basis of those shared understandings, participants discussed positions they were willing to accept as assumptions. With consensus on those fundamental points, participants broke into small groups to focus on modeling nursing diagnoses, nursing interventions, and assessment data or findings. Finally, participants reconvened in plenary session to report on the work of the groups and to decide on subsequent actions.
Results
Participants distinguished between the colloquial or interface terminologies with which users enter and retrieve information and the reference terminologies that contain all the allowable terms to refer to concepts and the semantic relationships among the concepts. The recognized nursing terminologies, participants agreed, were for the most part colloquial terminologies. They considered the impracticality of continuing to map from each colloquial terminology to every other, as in the Unified Medical Language System (UMLS). As each colloquial terminology evolves, the work of mapping to every other (evolving) terminology will grow exponentially.
Variation in colloquial terminologies is necessary from site to site and from specialty to specialty so that clinicians can use familiar terms that say exactly what they mean. To get comparable data for aggregation and analysis, the terms of the colloquial terminologies must be linked to the underlying concepts contained in a reference terminology. The reference terminology can then serve as an interlingua to trace the identical concept from one colloquial terminology to another. Those responsible for each colloquial terminology need maintain the mapping only to the reference terminology.
Creating the reference terminology is not merely a matter of collecting all available terms, however. Rather, it is first necessary to construct and test the adequacy of a model that depicts the core concepts of each type (i.e., Problem, Intervention, Goal, Outcome), the defining attributes of those core concepts, and the semantic relationships among them. When the model is ready, the terms associated with each concept may be added.
The participants in the Nursing Vocabulary Summit Conference agreed that the development of useful standards for nursing terminology required diverse expertise and collaboration. They agreed to continue meeting and working together, as the Nursing Terminology Summit Group (NTSG), toward the development of a reference terminology model for nursing.
Participants organized task forces with defined responsibilities. The task force on nursing diagnosis developed a high-level conceptual definition of diagnosis that would apply equally well to nursing, medical, or other disciplinary diagnoses. They tentatively defined the definitional elements of a diagnosis and agreed to develop and test a terminology model of nursing diagnosis before the 2000 meeting.
In contrast, the task force on nursing interventions had more difficulty identifying the relevant definitional elements. Their review of existing intervention representations yielded a number of candidate elements for inclusion in information or terminology models. Members recognized the need for more clarity in differentiating types of models and in specifying criteria for inclusion of elements. The group therefore undertook to achieve this clarity and then to identify candidate concepts for inclusion in a reference terminology model of nursing interventions.
The third task force recognized that outcomes, like goals, were assessment data representing observations at different points in time. Opting to focus broadly on observations, this group too struggled to distinguish information models from terminology models and to determine which aspects of an observation were definitional and therefore appropriate for inclusion in a terminology model. A subgroup (Bakken, Cimino, Haskell, and Huff)* offered to investigate the adequacy of the Clinical LOINC (Logical Observation Identifiers, Names, and Codes) semantic structure as a terminology model for standardized assessment measures, including those contained in recognized nursing terminologies as outcomes. Ozbolt proposed to test the goal statements of the Patient Care Data Set against the LOINC semantic structure as a terminology model and against the components of goal messages as defined by the Reference Information Model of Health Level 7.
The work of the task forces continues, and progress reports on some of the efforts are included in the papers in this Focus section. In addition, conference participants have applied what they learned at the conference in their reflections on other issues in nursing terminology. The other papers in this Focus section report this complementary work on terminology.
The work of the summit has also extended internationally. On reading a report of the conference,18 the chair of the Nursing Informatics Interest Group of the International Medical Informatics Association (IMIA) invited members of the NTSG to join international colleagues in a standards development effort. Collaborating via the Internet, these persons drafted and revised a proposal for a new work item to submit to Technical Committee 215 (TC 215) of the International Standards Organization (ISO). As amended and approved, this work item calls for the development and integration of a reference terminology model for nursing. The IMIA Nursing Informatics Interest Group, in collaboration with the International Council of Nurses, will prepare the draft standard for submission in 2001.
In addition, at the invitation of European colleagues, some members of the NTSG have collaborated via the Internet to develop a prestandard for a categorial system of nursing concepts to be proposed to the Health Informatics Technical Committee (TC251) of the European Committee for Standardization (CEN). Finally, the Nursing Terminology Summit Conference 2000, held June 8–11 at Vanderbilt University, brought together members of the NTSG with colleagues from Europe, Asia, Australia, and the Americas. Participants in the 2000 conference agreed to test emerging models, to support the IMIA effort for ISO, and to work toward a single international standard. Papers describing that work will be forthcoming.
Discussion
The Nursing Vocabulary Summit Conference 1999 was remarkable in its composition and its accomplishments. Although the restricted size of the conference meant that some leaders in the nursing vocabulary and health informatics standards efforts did not attend, those who did attend were in the forefront of these efforts. The small size of the conference promoted open discussion, and the evening social events helped participants relax and become better acquainted. Participants' expertise made discussions deep and thoughtful. Their professional seniority enabled them to make decisions and take actions.
As a result of the conference, participants came to recognize the utility and the limits of colloquial nursing terminologies and the need for a standard reference terminology. Even more, they experienced the necessity for and the rewards of collaboration to develop the standard. The tenor and the content of discussions about nursing terminology standards are now radically different. No longer do we ask “Which nursing vocabulary is the best candidate for the standard?” but rather “How can we work together to develop a reference terminology that will eventually serve all nurses everywhere?” and “How can nurses' views inform the development of more general health care terminology standards?” Ideas and understandings springing from the conference have grown into collaborations with more comprehensive standards groups such as LOINC (see Bakken19) and HL7 and into international work with IMIA, CEN, and ISO. After more than three decades of diverse efforts, collaboration at the summit shows promise of delivering terminology standards for nursing.
Judy Ozbolt, PHD, RN
Affiliation of the author: Vanderbilt University, Nashville, Tennessee.
Baca selengkapnya...
