comprehension of subtleties in the terminology (eg, the distinction between an anatomical part and an anatomical structure).

Transcription

1 Page 25 of 42 comprehension of subtleties in the terminology (eg, the distinction between an anatomical part and an anatomical structure). This understanding will be crucial in the design of computer-based applications to ensure the accurate collection and retrieval of data, and to enable such systems to utilise optimally (and appropriately) the full potential of SNOMED CT. Knowledge of SNOMED CT and how coded data are organised in a computer application will significantly improve the quality of information extracted from a database. HIM professionals are poised to take the lead in this new era of data management. Additional information concerning SNOMED CT can be accessed from the following websites: SNOMED CT technical documents: < the CAP s SNOMED International division: < the NHS Information Authority s Clinical Terminology Service: < References Campbell JR, Carpenter P, Sneiderman C, et al (1997). Phase II evaluation of clinical coding schemes: Completeness, taxonomy, mapping, definitions, and clarity. Journal of the American Medical Informatics Association, 4: Chute CG, Cohn SP, Campbell KE, et al (1996). The content coverage of clinical classifications. Journal of the American Medical Informatics Association, 3: Diane J. Aschman MS, RPh Chief Operating Officer SNOMED International [1] SNOMED Clinical Terms Consultation Document: Proposals for the new terminology, version 5, October HIMJ: Classifications and terminologies Nursing classification and terminology systems Evelyn J S Hovenga Abstract A number of terminologies exist that represent concepts of relevance to nurses, although none of these is in use by Australian nurses. Without consensus, nursing language and definitions incorporated in clinical information systems now being implemented will continue to vary considerably. The result will be an inability to compare nursing practice, or to aggregate data for research purposes, or to collect national statistical data to demonstrate the significance of nurses contributions to health care. This article provides an international historical overview of nursing terminology developments relative to what is happening in Australia, brief reviews of the many available nursing terminologies, an update of this work relative to activities being undertaken towards the development and adoption of standards, and a discussion about desirable future research and development activities. Introduction A number of terminologies exist that represent concepts of relevance to nurses, although none are in use by Australian nurses. There is an urgent need for these nurses to decide which terminologies they would prefer to use nationally. Without such consensus, nursing language and definitions incorporated in clinical information systems now being implemented will continue to vary considerably. The result will be an inability to compare nursing practice, or to aggregate data for research purposes or to collect national statistical data to demonstrate the significance of nurses contribution to health care. Smallwood (2000) indicated that:

2 Page 26 of 42 Widely accepted and implemented standards are absolutely critical to underpin information activities in the health sector, including electronic business transactions (e-commerce), the development of a national approach to electronic health records and exchange of information between different parts of the health sector. Australian nurses have a long way to go before the objective of implementing standard nursing terminologies is achievable. This article provides an international historical overview of nursing terminology developments relative to what is happening in Australia, brief reviews of the many available nursing terminologies, an update of this work relative to activities being undertaken towards the development and adoption of standards, and a discussion about desirable future research and development activities. Data representing nursing concepts A review of the literature revealed that there is some blurring in what is meant by a nursing classification system. On the one hand, it is about the classification of terms in a defined data set of nursing terms, and on the other it is about the classification of nursing concepts for the purpose of measuring nursing work or the quality of nursing care provided, also referred to as patient outcomes. One could perhaps argue that the latter is a minimum data set consisting of the key indicators reflecting either of these concepts. Alternatively, these latter systems could be referred to as administrative terminologies. [1] This paper includes reference to some of these, although the focus is very much on the classification of nursing data. The nursing concepts that need to be represented by a standard nursing terminology are most commonly classified into three categories: (i) nursing diagnoses or problems; (ii) nursing interventions or actions; and (iii) nursing outcomes or assessment. These represent the nursing process, a widely used framework guiding nursing actions and documentation. Every time assessment takes place it has the potential to lead to further actions. Despite the existence of many nursing terminologies, none is recognised as being able to represent completely the nursing domain in terms of breadth or granularity. Most existing nursing terminologies were developed to suit a specific purpose. They may be appropriate for use as a computer interface or as an administrative terminology, but they do not necessarily represent the complete domain of nursing concepts, nor are they well suited for multiple purposes, as is expected from clinical systems such as decision support systems or electronic health records. Such systems are referred to as reference terminologies. A brief historical overview of available nursing terminologies and classification systems demonstrates the advances made over time. Historical overview of nursing classification systems The development of standardised nursing classification systems began in earnest in the early 1970s, mostly in the United States of America (USA). This type of work was also undertaken in a number of European countries. Overall, national nursing organisations have played a significant part in these developments. Little has been done in Australia in this regard. Indeed, in relation to terminology developments, Australian nurses are very much behind their overseas counterparts. North American Nursing Diagnosis Association (NANDA) taxonomy The first USA taskforce to name and classify nursing diagnosis was established in 1973, together with a clearing-house for nursing diagnoses at St. Louis University, Missouri, that same year. This taskforce evolved into the North American Nursing Diagnosis Association and the development of their standard NANDA terminology. NANDA s Taxonomy I was first published in 1973 as an alphabetical listing of nursing diagnoses (Gordon 1998). The taxonomy expanded over several revisions, and in 1987 consisted of 128 nursing diagnoses grouped into nine axes representing human response patterns (that is, a patient s reaction to their disease, for example exchanging, communicating, relating and perceiving). Each axis could be divided further into subcategories for greater specificity (for example 8 Knowing; 8.3 Altered thought processes; Impaired memory). NANDA s Taxonomy II (4th edition), approved in 2000 and released in 2001, underwent a major restructure to accommodate computerisation needs and to improve coder inter-

3 Page 27 of 42 reliability. This multi-axial system describes 155 nursing diagnoses and includes atomic-level elements. ( NANDA s Taxonomy I is used by some Australian university schools of nursing, as it is a very useful teaching tool, but it has not been embraced enthusiastically by Australian nurses, primarily because the terms adopted do not reflect their everyday nursing language. Gordon s Functional Health Patterns Functional health patterns were first identified around 1974 as a means of classifying nursing concepts for the purpose of teaching nursing assessment and diagnosis (Gordon 1987, p.93). These 11 functional health patterns came about in consequence of Gordon s observation that a nursing diagnosis is essentially a conceptual model for interpreting a set of observations. Functional health patterns, such as nutritional-metabolic pattern or activity-exercise pattern, provide an alternative to human response patterns used first to classify NANDA diagnostic concepts, and may be used to classify various nursing concepts, however defined. Home Health Care Classification (HHCC) The Home Health Care Classification (HHCC) is a nomenclature consisting of two interrelated standardised vocabularies: (i) the HHCC of nursing diagnoses; and (ii) the HHCC of nursing interventions. The HHCC is the result of a large national empirical research study and was completed in Its primary purpose is to provide a structured and consistent method to assess and classify patients in order to determine the resources required to provide home health and ambulatory care services to the US Medicare population, including their outcomes of care (Saba 1992; The HHCC of nursing diagnoses includes 145 categories and subcategories. The HHCC of nursing interventions includes 160 categories and subcategories. Modifiers are used to expand both vocabularies: expected outcome (such as improved, stabilized or deteriorated) for each nursing diagnosis, and action type (such as assess/monitor or teach/instruct) for each nursing intervention. The coding structure is flexible and expandable and the code itself is a five-character alphanumeric code based on the format of the Tenth Revision of the International Statistical Classification of Diseases and Related Health Problems (ICD-10). The coding system has been designed to link and map the six steps of the nursing process and facilitates the design of critical care protocols and/or pathways to make this a very useful nomenclature. The HHCC is also widely used: it has been registered as a Health Level 7 (HL7) language and integrated into SNOMED RT and the Unified Medical Language System (UMLS). It has been used as the basis for the International Classification of Nursing Practice and translated into Dutch, Chinese, Portuguese, Spanish, Finnish, German and Korean (Saba 1994, p.9; Saba 1997, p. S69; Saba 2001). Visiting Nurses Association of Omaha (OMAHA system) The OMAHA classification system was developed in the early 1970s specifically for the Visiting Nurses Association of Omaha for classifying clients problems in community and home health care settings (Martin 1982). The system consists of three axes: (i) problem classification scheme; (ii) intervention scheme; and (iii) problem rating scale for outcomes. The problem classification scheme classifies 40 client problems or nursing diagnoses into four domains, with two sets of modifiers. The intervention scheme classifies 62 nursing actions or activities into four categories of interventions. The problem rating scale for outcomes is an evaluation tool designed to measure outcomes in relation to specific problems or nursing diagnoses. A five-point Likert-type scale is used to capture three concepts of knowledge, behaviour and status (Martin & Bowles 1997). The classification system was field tested in other community health agencies in 1976 and in 1980 was computerised and implemented for daily use. It is widely used in home care, public health, outpatient, case management, school, hospital and other practice settings. It is also used internationally and has been translated into Danish, Dutch, Chinese, Japanese and Swedish (Martin & Bowles 1997). It is now touted as being comprehensive, well tested, and based on the language of practising nurses (Martin 1992; < >).

4 Page 28 of 42 Outcome and Assessment Information Set (OASIS) The Outcome and Assessment Information Set (OASIS) is a set of data items developed during the 1980s to measure patient outcomes in home health care. The use of a standard, reliable assessment instrument is mandated by legislation in the United States for home health care agencies participating in the US Medicare and Medicaid programs (Kang 1999). OASIS consists of 79 data elements that measure patients health and functional status, health service use, living conditions and social support (Kang 1999). Most data elements were derived in the context of a Health Care Financing Administration (HCFA)-funded national research program. The data items have been subject to clinical review during the last ten years and modified accordingly < Nursing Interventions Classification (NIC) The Center for Nursing Classification was established by the University of Iowa in 1995 to facilitate the ongoing research of two classification systems developed by that University: (i) the Nursing Interventions Classification (NIC), and (ii) the Nursing Outcomes Classification (NOC). The NIC, first developed in 1987, is a proprietary, standardised language of both nurse- and doctor-initiated nursing treatments and includes both direct and indirect care interventions. It may be used in all health care settings and by all nursing specialties. In the third edition, there are 486 interventions, which are grouped into 30 classes under seven domains. Each intervention has a unique identifier to facilitate computerisation, a definition, and a detailed set of activities that describes what it is a nurse does to implement the intervention. NIC interventions have been linked with NANDA nursing diagnoses, the NOC, the OMAHA system problems, long-term care resident assessment protocols (RAPS), and the OASIS outcome measures < Nursing Outcomes Classification (NOC) The NOC, first developed in 1991, is a comprehensive, standardised classification of patient and client outcomes. Its purpose is to evaluate the effects of nursing interventions. It may be used in all health care settings and by all nursing specialties. In the second edition, there are 260 nursing outcomes listed in alphabetical order and grouped within a coded, conceptual framework into 29 classes under seven domains. Each outcome has a unique identifier to facilitate computerisation, a definition, a list of indicators that can be used to evaluate patient status in relation to the outcome, and a five-point Likert scale to measure patient status < The NOC has been linked to other classification systems (as per the NIC), but has been linked also to Gordon's functional patterns; it is officially registered as an HL7 terminology < Nursing Intervention Lexicon and Taxonomy (NILT) In 1992, Grobe (1992, p.981) reported on her research that resulted in the Nursing Intervention Lexicon and Taxonomy (NILT). This study focused on nurses natural language expressions of clinical care activities, adopted a scientific method to build dictionaries and thesauri of nursing intervention clinical terms, and noted that once a lexicon of clinical terms is available the design of automated data extraction from text notes and classification becomes a possibility. This work has progressed. NILT may be used to characterise two processes of care variables: intervention intensity (frequency), and focus. Grobe et al (1997 p.13) noted that the process of care variables measurement is important, as it characterises the nature of nursing care delivered and it quantifies the amount of that care. Patient Care Data Set (PCDS) During 1994 and 1995, the University Health System Consortium funded the compilation and testing of a Patient Care Data Set (PCDS) consisting of pre-coordinated phrases representing statements of patient problems (363 terms), patient care goals (311 terms), and patient care actions (orders: 1357 terms) (Ozbolt 1999). This was probably the first terminology developed to serve as a set of standard terms to represent and capture clinical data for inclusion in patient care information systems.

5 Page 29 of 42 Terms classified were derived from 29 standards of care, 76 care protocols and 30 patient education plans. The terms have been organised into 22 components, in accordance with the HHCC components and with minor modifications to suit the acute care setting. There are three axes within each component: problems, goals and orders (Ozbolt, Fruchtnight, Hayden 1994, p.181; Ozbolt 1999). The fourth version of the PCDS underwent a major revision to incorporate characteristics better suited for use in HL7 messages and for software development. The pre-coordinated phrases have been parsed into atomic-level concepts and rules established for combining the atomic-level concepts into complex concepts (Ozbolt 1997; Ozbolt 1999). Perioperative Nursing Data Set (PNDS) The Perioperative Nursing Data Set (PNDS) has been developed by the Association of Peri- Operative Registered Nurses (AORN), Denver, USA. It is a standardised nursing vocabulary consisting of 64 nursing diagnoses, 127 nursing interventions and 29 nurse-sensitive patient outcomes which together describe the peri-operative patient experience from pre-admission to discharge < < Systematized Nomenclature of Medicine (SNOMED) and nursing terms The feasibility of using SNOMED III to represent nursing terms was first tested by Bakken et al (1994 p.61). They found that 69% of the nursing terms included in their test dataset were represented by one or more SNOMED III terms. The SNOMED International Editorial Board asked the 2000 Convergent Terminology Group for Nursing < to develop principles, processes and strategies for enhancing the coverage of the nursing domain in SNOMED RT. The group comprises nursing domain experts whose primary role is to advise the SNOMED Editorial Board regarding scope of coverage, creation of hierarchies, semantic terminology definitions and scientific accuracy of the concepts and terms within a specific clinical domain. Their work has been extended to include the new collaborative work of combining SNOMED RT with the UK National Health Service s (NHS) Clinical Terms Version 3 (also known as the Read Codes) into a single reference terminology for health care: SNOMED Clinical Terms (SNOMED -CT) (Bakken et al 2001, p.151). This work began in 1999 and is touted as creating the most comprehensive language of health to support the computerised patient record, worldwide. Collaborative efforts between the College of American Pathologists, which is the owner of SNOMED, and the ANA, support nursing terminology efforts worldwide. (See this issue's article by Aschman.) Clinical Terms Version 3 (also known as the Read Codes) and nursing terms In the United Kingdom, the Strategic Advisory Group for Nursing Information (SAGNIS) was set up in late 1991 to advise the Information Management Group of the NHS Management Executive on information issues related to nursing. A Nursing, Midwifery and Health Visiting Terms project was undertaken as a contribution to the development of Version 3 of the Clinical Terms (also known as the Read Codes V3 and released in 1998). International Classification of Nursing Practice (ICNP ) Another seminal initiative was undertaken by the International Council of Nurses (ICN) in 1989, in response to a proposal by the Canadian and American nurses associations. The associations had expressed concern that it was not possible to name nursing s patient problems and to describe nursing s distinctive contributions to solving or alleviating them. Their resolution resulted in a feasibility study for the development of classification systems for nursing care, nursing information management systems and nursing data sets in order to provide tools that nurses in all countries could use to describe nursing and its contributions to health. An often-referenced quote says it all (Clark & Lang 1992, p.109): If we cannot name it, we cannot control it, finance it, teach it, research it, or put it into public policy. The alpha version of the International Classification of Nursing Practice (ICNP ), containing a classification of nursing phenomena and a classification of nursing interventions, was released

6 Page 30 of 42 by the ICN in Its structure was monoaxial, with concepts arranged in a generic hierarchy (each subordinate term is related to a superordinate term by principle of division) (Clark 2000). The Telenurse project [2] essentially funded the research and development of the ICNP in Europe and effectively facilitated the establishment of an extensive European network of nursing working with the ICNP. This resulted in the development of the ICNP Beta 2 version < The Beta 2 version, released in 1999, is a multiaxial classification of nursing phenomena, actions and outcomes. The unifying framework is adaptable to computerisation and enables existing nursing vocabularies and classifications to be cross-mapped to enable comparison of nursing data. There are eight axes for both nursing phenomena and nursing actions. Terms from separate axes can be combined to create complex concepts. Each term is accompanied by a definition. Nursing outcomes are measured by the change in the nursing diagnoses over time (International Council of Nurses 1999). The ICNP Beta 2 version has been translated in many languages and enjoys a lot of international support. The Telenurse ID-ENTITY (Integration and Dissemination of European Nursing Terminology in Information Technology) project promotes the use of the ICNP among nurses in central and eastern European countries (Mortensen & Neilsen 2002, p.46). The ICNP Program has established formal evaluation and review processes to advance the ongoing maintenance and advancement of the ICNP. The ICNP Program includes plans to release ICNP Version 1 in The Royal College of Nursing Australia (RCNA) is Australia s national organisation with ICN membership. Conrick and O Connell s (1998) discussion paper represented one effort to stimulate Australian nurses to consider the ICNP classification system. However funding is yet to be obtained to test and evaluate the ICNP within the Australian nursing environment. Swedish VIPS (Välbefinnande [well-being], Integritet [integrity], Prevention, & Säkerhet [safety]) Doctoral nursing research in Sweden set out to develop criteria and guidelines for nursing documentation in patient records. This resulted in a list of key words with explanations, now known as VIPS. This model for nursing documentation in patient records was promoted by the Swedish Nurses Association, adopted nationally and tested for validity and reliability (Ehnfors et al 1991; Ehrenberg et al 1996; Ahlfeldt et al 1999). Modifications for different nursing practice areas have been developed, such as for primary health care, psychiatric nursing and operating room nursing, and translations have been made for the other Nordic languages (Norwegian, Danish and Finnish), as well as for Baltic languages and some African languages. It is also used in different computer applications < Danish Nursing Intervention Classification In 1992, the Danish National Board of Health began the development of a Nursing Intervention Classification as part of the development of a common health care classification system in Denmark. The Danish Health Classification System aims to combine all the existing official classifications into a joint hierarchical structure, which will be the basis for all future classification development. Common rules for the structure of classification systems have been applied to ensure that all new classifications share common principles (Madsen & Burgaard 1997). The Nursing Intervention Classification is divided into 12 domains and has a total of 569 interventions. The classification system was released for use early Since then development of nursing diagnosis and nursing assessments classification systems has begun (Madsen & Burgaard 1997). Dutch nursing terminology classification system The Dutch nurses are testing the WHO International Classification of Functioning, Disability and Health (ICF) < although they do have a Dutch

7 Page 31 of 42 nursing terminology classification system (Goossen et al 2000, p.541). They also participate in the work of the 'Coordinatiepunt Standaardisatie Informatievoorziening in de Zorgsector' (CSIZ) and the National Normalisation Institute (NEN), which facilitates national discussions and their contribution to CEN (Comité Européen de Normalisation [European Committee for Standardization]) and ISO (International Organization for Standardization) standardisation work. The CEN prenv 14032, about nursing concepts, has been translated into Dutch to advance developments in this area. Nursing Minimum Data Sets A minimum data set is the minimum number of data element types that are needed to enable knowledge discovery to suit a specific purpose. With the significant increase in the use of information systems in all areas of the health sector it has become possible to collect a greater range of data to suit a variety of health-planning and decision-making purposes. As a result, many countries are now undertaking the development of new minimum data sets. This development requires: the identification of key data items which underpin central information requirements to suit various aspects within both the health and welfare industries clear and precise definitions for every data item. Until recently, minimum data sets were developed primarily to suit administrative purposes; this is changing. The development of many standard minimum data sets is essential so that clinical information systems such as electronic health records (EHRs) may be used to collect evidence of all clinical practice, as well as to provide clinicians with the knowledge required to effectively adopt evidence-based practice. Standardisation in healthcare terminology is a prerequisite for the ultimate improvement of professional patient care. Since patient care is recognised as being multidisciplinary (that is, not belonging to one specific professional discipline), there is an increasing need for flexible usage of existing terminology systems. One example is the minimum diabetes dataset for New Zealand (Simmons et al 2000). Although the dataset consists primarily of data elements describing a diabetic person s health status from a medical perspective, it needs to be able to be linked with standard nursing data, as nurses will be expected to provide care based on how these patients are responding to such health status conditions. This example also demonstrates the need for a common health terminology. It is also imperative that health professionals identify the minimum data sets of key indicators required to reflect evidence of practice for all types of practices, diseases and injuries. This will ensure that these standard key indicators can be made available on a per patient basis, enabling outcomes research and general practice evaluation. This requires the linkage of process data with outcomes data, as well as the use of unique patient identification and unique provider identification. There is a need also to resolve issues of ownership, access and custodianship of EHRs. Data will be required on a per episode, per provider, or per patient basis over any time period incorporating data provided by any number of providers and organisations. Nursing classification or terminology systems need to be able to support such usage. The following is a brief historical overview of nursing minimum data set development and current usage. North American Nursing Minimum Data Set (NMDS) The North American Nursing Minimum Data Set (NMDS) was built during the 1980s on the concept of the Uniform Hospital Discharge Data Set that was first adopted in Devine and Werley (1988, p.97) indicated that its purposes were to: establish comparability of nursing data across clinical populations, settings, geographic areas, and time describe the nursing care of patients or clients and their families in a variety of settings, both institutionally and non-institutionally demonstrate or project trends regarding care needs and allocation of nursing resources to patients or clients according to their health problems or nursing diagnoses stimulate nursing research through links to the detailed data existing in nursing information systems (NISs) and other health care information systems.

8 Page 32 of 42 There are four specific nursing care elements included in the NMDS: nursing diagnosis; intervention; outcome; and intensity of nursing care. The remaining elements pertain to demographics and service elements. Nurses will need to adopt a standard terminology incorporating data for each of these four nursing care elements to make the best possible use of the NMDS. Nursing Management Minimum Data Set (NMMDS) Huber, Schumacher and Delaney (1997) noted that progress had been made towards the creation of structures and methods for collecting NMDS data. However, they were of the view that nurse administrators and managers required another minimum data set to describe, analyse and compare contextual co-variates of patient outcomes. It took seven years to arrive at a data set of 17 variables complete with definitions and measures that cross the continuum of nursing care. The resultant nursing management minimum data set (NMMDS) was recognised by the ANA in 1999 < Welton and Halloran (1999) noted that, despite the wide acceptance of the NMMDS within nursing informatics, there has been sparse empirical testing of the concept to date. Their study demonstrated that nursing diagnoses are an independent predictor of patient outcome. One can therefore conclude that the national collection of such data is likely to be very beneficial. Community Nursing Minimum Data Set Australia (CNMDSA) The Community Nursing Minimum Data Set Australia (CNMDSA) was developed in 1990 and released in 1991 by the Australian Council of Community Nursing Services (ACCNS) (Turley 1991). The data set contained 28 data elements and, although it is not a compulsory data collection among community nursing services, it has been adopted by several agencies for use in their information systems. Version 2.0 resulted from a two-year review project undertaken by the ACCNS and Coopers and Lybrand, with funding from the then Commonwealth Department of Health and Family Services (under the National Home and Community Care Programme) (Australian Institute of Health and Welfare [AIHW] 1998; 2000). The number of data elements was increased to 35; they include items relating to service provision, demographic characteristics of patients, and clinical items. The format of the data dictionary and guidelines were changed to align with the International Standards Organization (ISO) data guidelines standard (AIHW 1998). One result of this project was the inclusion of eight CNMDSA data elements in the AIHW s National Health Data Dictionary (NHDD). The remaining data elements have been proposed for inclusion and are under review by the National Health Data Committee. They have been incorporated in the AIHW s knowledge base, awaiting formal endorsement for inclusion in the NHDD (AIHW 2000). The data elements include several specific nursing data items, such as nursing diagnoses or nursing intervention, for which one or more coding schemes have been recommended, although none has stood out as the definitive classification for nursing. The CNMDSA v2.0 publication, which details these possibilities, is available from the website of the ACCNS < It is unlikely that the CNMDSA will be developed further. It has been overtaken by the Health and Community Care minimum data set (HACC MDS). There is considerable client overlap between the two data sets and as a result many of the CNMDSA definitions (with some modification) have been incorporated in the HACC MDS (AIHW 1998; 2000). Residential Care Manual Australia In the aged care residential sector, the Commonwealth has developed the Resident Classification Scale, which is essentially a nursing minimum data set consisting of 20 care indicators. Its primary use is to provide a basis for funding nursing homes < contents/5classi2.htm#5.8>. Resident Assessment Protocols (RAPs) The US Nursing Home Reform Act mandates that nursing homes use the Resident Assessment Instrument (RAI), a clinical assessment tool to identify residents strengths, weaknesses, preferences and needs in key areas of functioning. This tool comprises three key

9 Page 33 of 42 components: the Minimum Data Set (MDS), Triggers and Resident Assessment Protocols (RAPS), and Utilisation guidelines. Version Two, now in use, was introduced in If one or a combination of MDS elements identifies conditions or the presence of specific clinical factors (triggers) for additional assessment and review, then RAPs are used for additional assessment and review. The 18 RAPs assist in the development of plans of care and include ADL function/rehabilitation, nutritional status, dehydration/fluid maintenance, pressure ulcers and 14 other clinical care indicators very similar to those used, for example, in patient dependency/acuity or nursing workload systems (Gibbs Brown 2001). Belgian Nursing Minimum Data Set The value of collecting nursing minimum data has been demonstrated in Belgium. This project began in 1983 and a national collection of nursing minimum data from hospitals became law in Sermeus et al (1994) define this process as the... systematic registration of the smallest number possible of unequivocally coded data, with respect to or for the purpose of nursing practice, making information available to the largest group possible of users according to a broad range of information requirements. Its main purpose was to assist with the allocation of hospital budgets. Nursing care is visualised by means of 23 scientifically selected care items. Selection was based on: validity of the minimum list, for example 20% of original data (and also 20% of the time) results in the preservation of 80% of the information agreement with four main lines of nursing activity: basic, technical, intensive, and psychosocial nursing care typification of certain activities enabling the link activity with a medical specialty frequency of occurrence (not too rare or too common). relevance of nursing activity based on professional grounds ability to check clinical relevance based on documented evidence quality of the description: universal understanding of concepts included. These data are of a greater level of granularity than that proposed for either the USA (NMDS) or Australia (CNMDS-A). The data reflect nursing resource usage, as they indicate the activities nurses were required to undertake for each patient, such as the provision of mobility or hygiene or elimination or feeding assistance, as well as the more technical activities, such as care of a tracheostomy or endotracheal tube or decubitus ulcer preventive care. Most may well relate to any nursing intervention data set (Sermeus et al 1994). This is one instance where the data collected by nurses are being used, not only by health services researchers, but also by policy makers. Belgium now has more than 12 years of national nursing data. It is collected four times per year during a 15-day sampling period in which a systematic random sample is taken. Data are aggregated by hospital, nursing unit, medical diagnosis, patient and patient day (Vandewal et al 1996). Nursing Minimum Data Set for the Netherlands (NMDSN) The Netherlands established a national initiative group whose members undertook a research project to develop a method and to define the data necessary to describe the diversity and complexity of different patient populations, and the variability of patient-related nursing care activities. As part of the project, an evaluation of the presence of proposed data items in each hospital s nursing records was made. The resulting nursing minimum data set for the Netherlands (NMDSN) is expected to be used for trend analysis, budget negotiations and policy making. The proposed NMDSN consists of: five setting and provider items six patient demographic items seven items describing the medical condition ten nursing process items 24 patient problems 32 nursing interventions

10 Page 34 of 42 four items that reflect outcomes of nursing care three complexity of nursing care items. A major difference between this data set, the USA NMDS and CNMDS-A is that the previously described systems use unlimited sets of patient problems/diagnoses, interventions and outcomes. Goossen et al (2000) noted that such data collection only becomes feasible when the data can be extracted from EHRs. Even so, they concluded that the collection of their NMDSN would need to overcome a number of practical difficulties. The NMDSN was used to test its usefulness for six purposes, as well as data collection methods, data analysis and reporting. Goossen (2001, p.1337) concluded that it offers a maximum of information from a minimum amount of data and effort. Discussion on nursing terminologies, classifications and minimum data sets The review of the literature regarding the development of nursing terminologies or classification systems reveals that most were derived by means of consensus, from a selection of various nursing documentation, the use of Delphi methods, and general peer review and feedback. Furthermore, they tended to be first-generation hierarchical systems, [3] although a number were further developed to become multi-axial terminologies. By and large, their coding systems were not specifically designed to optimise computer use. The exception is the HHCC, which was empirically developed from a very large national, randomised sample of nursing documentation and its coding system was specifically designed to optimise computer use. The ICNP Beta is the most comprehensive nursing terminology available. The developers made every effort to ensure that the needs being met by other nursing classification systems could also be met by the ICNP. Its coding system leaves much to be desired (see Box below) and there continues to be some controversy over its structure. Moreover, where the ICNP is in use, nurses tend to use terms other than those used by this standardised terminology. An example of the ICNP Beta version coding system 1 Nursing Phenomenon 1A.1 Human being 1A.1.1 Individual 1A Function 1A Respiration 1A Ventilation 1A Hyperventilation 1A Hypoventilation 1A Dyspnea It is noteworthy that the major care components as defined in the HHCC and the PCDS nursing classification systems are similar to the categories identified in the Belgian and Netherlands NMDS. These, in turn, are very similar to the data collected using patient dependency/acuity systems such as the Patient Assessment and Information System (PAIS) [4] or the Trend Care system < in use in Australia. Indeed, ExcelCare < another system used for the same purpose, contains a lot of nursing data relevant to nursing diagnosis, interventions and outcomes, but does not subscribe to any particular nursing classification system or terminology. Similarly most systems that are in use or in the process of implementation, in Australia as well as elsewhere, use their own data sets, unless requested by the user to use a structured terminology. There has been considerable discussion among the international experts regarding the need to identify nursing outcomes. Some see this simply as another nursing assessment at a different point in time, and others are of the view that there are specific outcome concepts

11 Page 35 of 42 that need to be captured, as evidenced by a number of terminologies providing outcome data elements. Nursing reference terminologies Most, if not all, existing nursing terminologies: do not suitably represent all nursing concepts are unable to be used for a variety of purposes are not concept oriented, or do not have a coding system that is suited to information systems supporting EHRs. Bakken (2000) noted that the level of granularity of data required varies depending on the purpose of their use and... that no existing standardised nursing terminology meets the evolving criteria for terminologies related to suitability for implementation in and manipulation by computer based systems. It has become apparent that, in order to meet these criteria, terminologies must be concept oriented. True concept-oriented terminologies are now being referred to as reference terminologies. Concepts may be associated with multiple terms but a term should represent only one concept. Bakken (2000) defines a concept-oriented terminology as one where the concepts are formally defined and the semantic relationships among them identified in a manner that renders them suitable for computer processing, and notes that considerable progress is being made towards their development. Activities being undertaken towards the development and adoption of standards The nursing profession has been able to establish an international network of nursing terminology developers and experts in this field. As a consequence, a number of individual nursing initiatives in various countries have been influenced so that progress is underway towards greater convergence. Furthermore, as a result of the many activities being undertaken towards standards development in an effort to achieve a global health information society, a number of these nursing initiatives have been linked also to related standards activities. It is very encouraging to see the extent of collaboration between a number of disciplines now taking place. One regular event is the invitational Vanderbilt Nursing Terminology Summit, held in June or July in Nashville, USA, which began in The Summit provides a forum for key stakeholders such as terminology, standards and software developers to meet with a small number of nursing informatics experts who have a strong interest in nursing terminologies. Many participants are affiliated with key standards-developing organisations. They are now known as the Nursing Terminology Summit Group (NTSG). The first objective of the Summit was to seek consensus on, and a common approach to, the development of nursing terminology standards for use in information systems (Ozbolt 2000). The exchange of information about the latest developments permits a pushing of the boundaries to take place and it facilitates further cutting edge development to solve identified problems. One such example was the investigation of 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 (Bakken et al 2000). Ozbolt (2000, p.520) 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. Collaboration with others is continuing. The Nursing Terminology Summit Conferences, held in 2000 and 2001, brought together members of the NTSG with colleagues from Europe, Asia, Australia, and the Americas. Nursing Information and Data Set Evaluation Centre (NIDSEC) The American Nurses Association (ANA) has recognised the importance of the adoption of standards pertaining to nursing data and information systems for some time. In 1986, the ANA noted that

12 Page 36 of 42 It is essential that nurses have appropriate nursing data collected from a variety of health care settings where nurses provide care, for purposes of decision making in clinical nursing practice; management of nursing care and nursing resources; and nursing administration planning, monitoring, and control. The ANA established the Nursing Information and Data Set Evaluation Centre (NIDSEC) in 1996 to develop standards relating to automated information systems that support nursing documentation of clinical practice. The standards developed to date assess the four dimensions of nursing data sets and the systems that contain them: nomenclature; clinical content associations; clinical data repository; and general system characteristics (Bakken Henry et al 1998). Vendors voluntarily submit their products for review and any that meet the standards are ANA recognised. The accreditation lasts for three years, after which time the developer must reapply for recognition status (Bakken Henry et al 1998). The ANA hopes that the widespread use of information systems that meet NIDSEC standards will lead to its long-standing goal of achieving large, retrievable pools of patient data that reflect the nature, costs and effects of nursing practice (Zielstorff et al 1997). Eight terminologies have obtained ANA recognition and are incorporated into the Unified Medical Language System (UMLS). These are: NANDA Nursing Interventions Classification/Nursing Outcomes Classification Nursing Management Minimum Data Set Home Health Care Classification OMAHA system Patient Care Data Set Perioperative Nursing Dataset SNOMED RT It is noteworthy that the NILT has not been recognised by the ANA. The research that produced this terminology differed from all others in that it adopted scientific language (linguistical) methods. More recently, NIDSEC has begun to recognise software products as well < Nursing Information Reference Terminology Model The many vocabulary-related activities that are underway internationally have implications for the nursing profession. The absence of a mechanism to ensure that nurses were able to influence these standards development activities was of concern to the author. [5] Members of the International Medical Informatics Association s Nursing Informatics (IMIA NI) special interest group, some of whom were at the Vanderbilt summit, were alerted to this concern via their electronic mailing list. As a consequence, the members of the nursing terminology summit group (NTSG) were invited to join their 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 the ISO Technical Committee 215 (TC 215) via the American Standards National Institute Technical Advisory Group (TAG) on behalf of the IMIA NI and the International Council of Nurses (ICN). The work item related to the development of a reference terminology model for nursing. At the invitation of European colleagues, some members of the NTSG have collaborated via the internet to develop a pre-standard for a categorical system of nursing concepts for the Health Informatics Technical Committee (TC251) of CEN. This work became the foundation for the ISO work item. The resultant reference terminology model for nursing has been submitted recently as an ISO committee draft. Health Level 7 and nursing vocabularies Health Level 7 (HL7) is a standards development organisation accredited by the American National Standards Institute. Its work focuses on messaging standards to facilitate system interoperability. Its standards are used widely in Australia. Huff (1998) noted that between 75% and 90% of the work and expense of implementing a computer-to-computer interface is consumed in aligning vocabulary between the two

13 Page 37 of 42 interacting systems. As a result, a vocabulary technical committee was established with the objective of identifying, organising and maintaining terms used in coded fields for HL7 messages (Bakken et al 2000; < This committee has developed a set of principles for HL7-compliant terminologies. HL7 uses conceptual information models representing each sanctioned terminology and the terminology needs to have been submitted to an HL7-sanctioned terminology integration organization such as the National Library of Medicine s Unified Medical Language System (UMLS). As a consequence, only nursing terminologies that are approved by the ANA and incorporated in the UMLS can be HL7 compliant and registered. Association for Common European Nursing Diagnoses, Interventions and Outcomes (ACENDIO) The European nurses established an Association for Common European Nursing Diagnoses, Interventions and Outcomes (ACENDIO) in 1995 to promote the development of a nursing professional language. They recognised the need to be able to express, in their own language, what it is they do for patients and clients, why, and with what outcomes. They want to be able to share and compare information about practice, and noted that this was possible only by means of a standardised, common terminology which ensures that like is compared with like < An international nursing minimum data set The need for multiple minimum data sets to suit specific purposes was confirmed by those attending IMIA s International Nursing Informatics Symposium [6] pre-conference tutorial on the conceptualisation and feasibility of an International Nursing Minimum Data Set. Participants agreed that work should progress towards establishing synergy among several international communities to identify a common core of elements and their definitions that cross national boundaries. The IMIA NI nursing concept representation working group has secured ICN involvement with this project. Although the focus is nursing, a number of the concepts covered are equally applicable to other health professional disciplines. Desirable future research and development activities for Australian nurses The main initiatives undertaken by Australia s Commonwealth Government began during 1999, when the National Health Information Management Advisory Council first released its Health Online document, and, later that year, its publication Health Online: A Health Information Action Plan for Australia. As a result, a National Electronic Health Records Taskforce was established. The Taskforce released first an issues paper and later a publication titled A Health Information Network for Australia in July This is now referred to as the HealthConnect initiative. The second edition of Health Online was released in September The following statement provides the rationale for the HealthConnect vision and associated initiatives (Smallwood 2001, p.iii): Australia is embarking on great changes in the way health care is delivered. New and evolving information and communications technologies provide a powerful tool to improve the delivery of health care and achieve better quality care and health outcomes for individuals and communities. Such technologies also have the potential to deliver information that will empower individuals by helping them better manage their own health. One publication arising out of this initiative is titled A National Health Information Standards Plan for Australia < This initiative is essentially driving most health informatics standards developments in Australia. The adoption of terminology standards is fundamental to this project s success. The adoption of EHRs requires the capture of data at every point of care in a manner that is useful to others so that manual record keeping is no longer required. It is imperative that EHRs are useful to all healthcare providers and facilitate sharing and exchanging various aspects of these records. The draft requirements for an EHR reference architecture framework version 6.1 (Schloeffel 2001) document the primary and secondary uses of EHRs, as well as a comprehensive overview of the associated health terminology requirements. Collectively, they provide a solid