The development of classification schemes is the focus of several recently completed or ongoing nursing research projects including the North American Nursing Diagnosis Association Taxonomy 1 (NANDA, 1992), the Nursing Intervention Classification (McCloskey & Bulechek, 1992), the Georgetown Home Health Care Project (Saba, 1991), the Omaha Community Health System (Martin & Scheet, 1992a), and the Nursing Intervention Lexicon and Taxonomy (Grobe, 1992). Classification schemes are an essential prerequisite for data bases, knowledge bases, health care information systems, and expert systems. In addition, the development of unified language for clinical terms is pivotal in the development of an outcomes infrastructure to examine the linkages among patient problems, health care interventions, patient outcomes, and health care costs (McCormick, 1991). The lack of data elements related to nursing care in large federal data bases has highlighted the """"""""invisibility"""""""" of nursing and emphasized the need for additional research aimed at testing existing classification schemes for clinical terms related to nursing care (Lange & Jacox, 1993).
The aim of this study is to compare selected classification systems for their ability to represent the terms used to describe the patient problems or nursing diagnoses and nursing interventions in the patient record. The study will utilize an existing data set comprising more than 1,000 patient encounters for persons living with AIDS (PLWAs) receiving nursing care in three clinical settings. The data set includes a broad array of biopsychosocial problems and a diverse variety of nursing interventions (Janson-Bjerklie, Holzemer, & Henry, 1992). Two types of terms will be evaluated using a semiautomated lexical matching approach: 1) the natural language terms charted by the nurse in the care plan and progress note/flowsheet, and 2) the base concepts of the natural language terms which will be derived by stripping the modifiers from the terms and will include synonyms and lexical variants of the terms. The quality of the representation will be measured by a concept match score based on hierarchical classification relationships. The Stuart extension of the McNemar test for correlated proportions will be used to compare the systems on concept match scores. The study findings have the potential to refine and extend the Unified Nursing Language System proposed by the American Nurses Association (ANA, 1993a) and the International Council of Nurses (Clark & Lang, 1992) and to influence the inclusion of data elements related to nursing care into federal data bases and the computer- based patient record.
Showing the most recent 10 out of 14 publications
