the Nursing Citation Index

Transcription

1 z~~~~~~~~~~~~~~~~~~~~~ m In search of applications of nursing theories: the Nursing Citation Index By E. Diane Johnson, M.A.L.S. Head, Information Services J. Otto Lottes Health Sciences Library University of Missouri-Columbia Columbia, Missouri Cited author searches were conducted in Nursing Citation Index to determine its utility in locating clinical studies that apply the conceptual frameworks of Dorothea Orem, Callista Roy, Martha Rogers, Betty Neuman, and Dorothy Johnson. Fully 75 percent of the relevant papers would have been missed by a conventional subject/ textword search in the MEDLINE or CINAHL databases. Had Nursing Citation Index not been available, only 22 percent of relevant papers could have been retrieved by cited author searches of Social Science Citation Index. In summary, Nursing Citation Index provides an important indexing link between nursing theory and nursing research. THE PROBLEM Several years ago, the staff of the J. Otto Lottes Health Sciences Library at the University of Missouri-Columbia noticed that they were receiving a large number of queries from nursing students seeking applications of the work of various nursing theorists. Traditional bibliographic search techniques-both computerized and manual-proved to uncover only a small fraction of the universe of relevant literature. Many appropriate articles failed to mention their connection to the theorist in the title or, for that matter, in the abstract if one was available online. In addition the controlled vocabularies used by the AJN International Nursing Index (INI) and the Cumulative Index to Nursing and Allied Health Literature (CINAHL) did not permit direct subject access for theory application. After attempting to locate applications of Orem's Self Care Deficit Theory (SCDT), one nursing scholar lamented: In reviewing the literature related to Orem's SCDT, two major problems were noted. The term 'self care' is used in a variety of ways, so that literature searches are not exclusive to SCDT and not all reports based on SCDT are so referenced and retrievable [1]. Other authors share her concern [2-3]. In fact, it is not unheard of for nurse researchers to report paging 176 through ten to thirty years worth of Nursing Research and similar titles to locate research related to nursing models [4-5]. An alternative to a conventional subject/textword search for this type of information was a cited author search of the theorist in question. Frequently, however, these searches proved unsatisfactory because of scanty nursing journal coverage. As of 1985, Social Science Citation Index (SSCI) covered only eleven of the major nursing journals, while Science Citation Index (SCI) included only two. With the introduction of a Nursing Citation Index (NCI) section of INI in 1986, cited author access to the nursing literature was greatly expanded. It now appears that the researcher will have a powerful aid in tracking down much of this elusive literature. This index permits cited author searching for 157 of the major nursing journals worldwide. This study began with the working hypothesis that NCI would prove useful in locating research that applies the work of the various nursing theorists, and further, that many articles so retrieved would not be accessible by other methods. WHY A CITED AUTHOR APPROACH? Librarians have long recognized the usefulness of a cited author search to enhance the results of a conventional subject or keyword approach [6]. In theory,

2 Nursing Citation Index each time an author's work is cited by another, a conceptual link is established between the two. This link is not affected by changes in the terminology of a discipline over time; neither is it dependent on the use of a controlled vocabulary term as applied by an indexer [7]. With citation indexing, each cited reference serves as an access point. On the average, a scientific journal article has a bibliography containing fifteen references to previous works [8]. Therefore, using the cited references of a document as "content indicators" [9] provides a greater depth of subject access than is usually afforded through conventional indexing, since few conventional indexes provide as many as fifteen descriptors. There have been some attempts to codify the relationship between a citing paper and the items in its bibliography, that is, to classify cited papers by the reasons they were referenced [10-11]. Garfield has mentioned fifteen possible reasons for citing, two of which are relevant to this investigation: "Identifying original publications in which an idea or concept is discussed," and "Identifying original publication or other work describing an eponymic concept or term as, e.g., Hodgkin's disease, Pareto's Law, Friedel-Crafts Reaction, etc." [12]. Each time an author's work is cited by another, a conceptual link is established between the two. Herein we find the relationship between application and theory that we are seeking. When a nurse conducts research by using the concepts, assumptions, and theoretical framework of a given nursing model, it seems safe to assume that credit will be given to the model and its creator through a reference to the original work. Hence, the cited author approach would seem to have two added advantages over a conventional approach for identifying applications of nursing theories. First, by virtue of the number of citations typically found in the bibliography of an article, we are afforded additional access points, or what amounts to a greater depth of indexing. Second, cited references can serve to establish a conceptual or eponymic link between two sources that is either elusive or absent when a search is conducted through conventional access points. DEFINITIONS Many authors in nursing have attempted to distinguish among the terms "theory," "model," and "conceptual framework," but much confusion and dissent remains as to the meaning of these terms. Most nursing scholars agree that models differ from theories in their "level of abstraction" [13]. Models are not amenable to testing [14]; however, the theories which are derived from them are. Models provide the basic assumptions to guide research and aid in the creation of testable theories [15-16]. Despite this basic consensus, the terms continue to be used almost interchangeably in the literature [17-18]. To add further fuel to the fire, the creators of these models, theories, and frameworks are universally referred to as nursing "theorists." For the purposes of this paper, "theory," "model," and "framework" will be used interchangeably to refer to the written descriptions of the overall view of the nursing process set forth by each of the nursing theorists. How, then, are models applied in nursing research, and what constitutes a clinical application of a model? Nurse researchers tend to utilize models in three ways: 1) studies in which the model is named and described, but not actually integrated into the research itself; 2) descriptive studies in which the model is used as an organizing framework; and 3) research in which a theory is tested through scientific method [19]. Distinguishing among these three uses is difficult [20], and perhaps is best left to nursing scholars. For the purposes of this investigation, any of these three uses of models was considered an "application" if the article so stated, and to be "clinical" if it dealt with some aspect of patient care. BACKGROUND The quest for a model to conceptualize the nursing process actually dates back to Florence Nightingale's Notes on Nursing [21]. However, it was with the formation of the Nursing Development Conference Group in 1973 [22] that the nursing profession first began an effort to identify an organized framework of nursing practice and label it as such. Since then, nursing models have proliferated, and they have had an ever-increasing impact on research, education, and practice. Many nursing schools now base their entire curriculum on one or more models [23-24], and students are therefore required not only to learn about the various models, but also to "examine others' use" [25] of them; that is, to see how they are applied to nursing research and practice. As time passes, there will be more and more uses of the models to examine. Beck, in a 1985 study of theoretical frameworks used in the articles published in the journal Nursing Research from January 1974 to June 1985, noted a recent trend away from borrowing models from related disciplines in favor of using models unique to nursing [26]. Riehl and Roy concur that theory-based nursing research "will undoubtedly continue," because "the ultimate criteria of a theory's usefulness, and hence a model's, are whether 177

3 Johnson it stimulates new observations and insights, and generates predictions of relevant events that are subsequently confirmed" [27]. RESEARCH QUESTIONS The study was designed to answer the following questions: 1. How many source papers can be located by conducting cited author searches in NCI using the key papers of five prominent nursing theorists? Of these, what percentage of the source papers identified in this manner actually describe clinical applications of the various models? 2. How many of the citing papers identified in the NCI search could have been retrieved through "traditional" textword searches of MEDLINE and NAHL (Nursing and Allied Health Literature)? Stated another way, how many unique references were retrieved by the NCI cited author searches that would have been missed had an NCI cited author search not been conducted? 3. SCI and SSCI have provided cited author access to thirteen of the main nursing journals for a number of years. Do clinical applications appear primarily in the source journals already covered by SCI and SSCI? If the clinical applications appear only in this core group of thirteen journals, then NCI, its source coverage of 157 nursing journals notwithstanding, is contributing nothing unique toward locating applications studies. Does NCI provide unique cited author access to applications papers published in source journals outside the purview of SCI and SSCI? 4. How complete a list of applications papers can be compiled with an NCI cited author search? Does a traditional subject/textword search yield relevant items which a cited author search does not? A closely related question would be: do authors of clinical applications papers routinely cite the frameworks they are applying in their bibliographies? Finally, if a "traditional" search is undertaken, is a textword search for the name of the theory and its creator sufficient, or should broader controlled vocabulary terms that approximate the content of the theory (such as "self care" or "adaptation, psychological") be incorporated into the search to enhance recall? 5. What are the strengths and weaknesses of a cited author search versus conventional subject and textword searches with regard to precision and comprehensiveness of retrieval? 6. Finally, which of the theories is generating the most clinical research today? METHODS 1. NCI cited author search; percentage of clinical applications. The first three 1986 quarterly issues from 178 NCI were searched for any authors citing the key papers (Appendix 1) of the following nursing theorists: Dorothea Orem Betty Neuman Dorothy Johnson Martha Rogers Callista Roy Self Care Deficit Theory Neuman's Health Care Systems Model Johnson's Behavioral Systems Model Roger's Life Process Model/Science of Unitary Human Beings Roy Adaptation Model The key papers of each theorist were identified with the aid of a textbook on nursing theory, chosen because of its extensive bibliographies [28]. These particular theorists were selected for two reasons. First, their models are widely used in various nursing curricula [29-30]. Second, they are the ones most heavily used in clinical nursing research today [31]. After the cited author searches were completed, source journals in foreign languages were immediately eliminated, because the time and expense of obtaining full-text translations in order to determine which among them reported clinical applications would have been prohibitive. The article titles for the remaining references were obtained from the author section in the main body of INI and inspected. With five of the references, it was obvious from the title that the article was not a clinical study. The full text of the remaining seventy-nine articles was obtained and examined to judge if the paper described a clinical application. As stated earlier, even nursing scholars disagree on what qualifies as a clinical application study. Since no objective criteria have yet been developed that could be used here to identify clinical applications, an operational definition had to be devised. For the purpose of this investigation, the text of the article had to state explicitly that the study was drawing on the model for its theoretical framework in order for it to be counted as an "application" of the model. 2. Analysis of NAHL and MEDLINE access points in the clinical applications located with NCI. Citations that passed the relevancy test were then searched in the MEDLINE and NAHL databases. Titles, subject headings, and abstracts (where available) were printed for each. Each reference was then analyzed to determine if it contained the name of the model or its creator, making it accessible through a "traditional" textword approach. 3. Remaining unique NCI coverage after comparison with SSCI and SCI source journal lists. Cited

4 Nursing Citation Index author access to some of the NCI journal titles has been available for a number of years through SCI and SSCI. To determine how many of the relevant applications could have been located through a cited author search of those indexes, the journal titles which contained application articles were compared to the SSCI and SCI source lists for Next, these items were compared against the NCI applications retrievable in MEDLINE and NAHL to calculate coverage overlap among SCI, SSCI, MEDLINE, and NAHL relative to NCI. 4. Test for recall on NCI cited author search. Next, an attempt was made to assess the comprehensiveness of the cited author search. After all, if a relevant paper did not include one of the theorist's key papers in its bibliography, it would be overlooked in a cited author search even though it might show up on a traditional subject/textword search. A broad MEDLINE search for English language citations indexed from January through October 1986 yielded forty-five references. The search strategy incorporated a free-text search for the names of the theorists and their models, and also included MeSH terms such as "nursing theory" and "models, theoretical", as well as MeSH terms that approximate the content of the theories where available, such as "adaptation, psychological" (Roy) and "self care" (Orem) (Appendix 2). Citations that did not appear in the three 1986 quarterly INI indexes under study were then eliminated; this left thirtyeight references for review. The full text for each of these items was examined to determine which qualified as clinical applications. Then the citation for each relevant application was analyzed to determine if it could be retrieved with a narrow textword search for the model name or its creator, or whether it was retrieved solely by the use of the broader controlled vocabulary terms such as "self care" or "adaptation, psychological." 5. Comparison of comprehensiveness and precision: cited author vs. conventional methods. Combining all relevant applications studies identified by the broad MEDLINE and NCI cited author searches made it possible to calculate comprehensiveness for the broad and narrow MEDLINE approaches and for the NCI cited author search. Comprehensiveness is defined here as the percentage of all relevant application studies retrieved by a cited author search, a broad conventional search (as conducted in MEDLINE using a combination of subject headings and textwords), and a narrow conventional search (as determined by examining access points in the broad MEDLINE results to identify a subset that could be retrieved solely by searching on textwords). It was arrived at by using the formula: # of relevant applications retrieved by the method (# of NCI relevant applications + additional applications found in MEDLINE) Precision for each retrieval method was calculated in the following manner: 4 of relevant applications retrieved by the method total # of references retrieved by the method A table was prepared to permit comparison of comprehensiveness and precision for all three retrieval methods. 6. Percentage of clinical applications using each theory. The applications papers were then grouped by theory used, and the number of times each theory was cited was calculated as a percentage of the total relevant applications. RESULTS AND CONCLUSIONS 1. NCI cited author search; percentage of clinical applications. The NCI cited author searches identified a total of 109 references to eighty-four distinct citing papers. Of these, twenty-seven, or 32 percent, qualified as relevant applications of one of the nursing models, while 68 percent (fifty-seven) of the citations were not relevant (Figure 1). Therefore, a cited author search can prove fruitful for locating clinical applications of nursing theories. However, this type of search will provide a substantial number of irrelevant citations. 2. Analysis of MEDLINE and CINAHL access points in the clinical applications located with NCI. Although the NCI search was low in precision, it scored high on recall. In fact, the analysis of MEDLINE and CINAHL access points revealed that twenty, or fully 74 percent of the twenty-seven applications papers found in the NCI search, would not have been retrieved through a textword search of those files, since neither the name of the theorist nor the name of the model appeared in the title or abstract (Figure 2). Seven of the remaining clinical applications papers were retrievable by a NAHL textword search for either the theory or its creator. Three of these articles were not accessible in MEDLINE. They were retrieved in the NAHL search because an indexer had enriched the title by adding the name of the model in parentheses. Four of the twenty-seven applications contained access points making them retrievable in MEDLINE, but all of these references would also have appeared in searches of NAHL. Since 74 percent of the relevant 179

5 Johnson Figure 1 Proportion of relevant clinical applications papers retrieved in NCI search (n = 84). Figure 2 Accessibility of relevant NCI applications via textwords in MEDLINE & NAHL (n = 27). Clinical Applications (27 papers) Other (57 papers) citations would have gone undetected had an NCI cited author search not been conducted, NCI appears to offer much in locating clinical applications as compared to a traditional bibliographic textword approach. 3. Remaining unique NCI coverage after comparison with SSCI and SCI source journal lists. Next the journal titles containing applications papers were checked for coverage against the SSCI and SCI source journal lists. Six of the applications papers (22 percent) could have been retrieved through a SSCI cited author search. Of these, two papers contained access points that would also render them retrievable in MEDLINE and NAHL. The remaining four papers, representing 15 percent of the total applications, would have been missed using the conventional textword approach available in MEDLINE and NAHL. None of the relevant applications papers appeared in the SCI source journals. Figure 3 depicts the twenty-seven applications papers again; this time accounting for cited author access of those journals covered by SSCI. After SSCI, MEDLINE and NAHL accessibility was determined, there still remained sixteen papers, or 60 percent of the relevant applications located in NCI, which were not accessible by any other means. Therefore, it seems safe to conclude that the minimal coverage of nursing journals available via SSCI and SCI prior to the debut of NCI did not provide adequate cited author access to locate available clinical applications, and the NCI's enhanced source nursing journal coverage is helpful in identifying clinical applications which could not previously have been identified by either cited author or textword access. LEGEND -4I NCI & NAHL Only (3 papers) 11 % - NCI, NAHL, & MEDLINE (4 papers) 15% <C NCI Only (20 papers) 74% 4. Test for recall on NCI cited author search. The broad MeSH subject/textword search of MEDLINE yielded thirty-eight references which were also covered in the INI test issues. When these articles were examined, thirteen (34 percent) were deemed to be relevant applications papers (Table 1). A closer inspection of these relevant papers revealed that of Figure 3 Accessibility of relevant NCI applications via SSCI, MEDLINE, & NAHL (n = 27). LEGEND -II NCI Only (16 papers) 60% 4 NCI & SSCI Only (4 papers) 15% <C NCI, NAHL, MEDUNE & SSCI (2 papers) 7% -'4I NCI, NAHL, & MEDUNE (2 papers) 7% <U NCI & NAHL Only (3 papers) 11% 180

6 Nursing Citation Index Figure 4 Distribution of all relevant clinical applications by theory used (n = 32). Orem (18 Roy (6 papers) papers) Rogers (2 papers) 6% Neuman (3 papers) 9.5% Johnson (3 papers) 9.5% these, eight, or 61 percent had also been retrieved in the NCI cited author searches. Five, or 39 percent had not been identified with NCI. All five of these papers came from a single publication from the National League for Nursing (NLN). (NLN publications do not appear in NCI.) Since all of these papers came from a single publication excluded explicitly from NCI, these results indicate that clinical nurse researchers can indeed be relied on to cite the model which they are applying in their bibliographies, and an NCI cited author search is liable to yield virtually complete recall for clinical applications published in the nursing journal literature. It is less clear whether NCI should extend its indexing to the nursing association publications to ensure complete coverage, since all five unique relevant papers came from only one NLN publication. 5. Comparison of comprehensiveness and precision: cited author vs. conventional search methods. An analysis of the access points on the thirteen relevant clinical applications retrieved in MEDLINE revealed that nine, or 69 percent, mentioned explicitly the theorist or the model in the title or abstract. Meanwhile, four, or 31 percent, of the relevant applications, were retrievable only by broadening the search strategy to include such MeSH headings as "self care" and "nursing theory." However, the inclusion of MeSH headings accounted for the low precision on the MEDLINE search. Had the search been restricted to free-text occurrences of the names of the theorists or their models, precision would have been increased from 34 percent to 47 percent overall. It is interesting to note that the precision of the broad MEDLINE search, at 34 percent, was comparable to the 32 percent precision on the NCI search. However, there was a marked difference in comprehensiveness (Table 1). A total of thirty-two relevant applications papers were located: twenty-three unique to NCI and five unique to MEDLINE. Using these thirty-two papers as a yardstick to measure the comprehensiveness of each method, NCI's performance, with twenty-seven relevant applications papers, stood at 84 percent. Meanwhile, the broad MEDLINE search yielded 41 percent comprehensiveness with thirteen relevant applications papers (eight overlapping with NCI, plus five unique). These figures indicate that when using NCI to locate applications studies like these, the slightly lower precision achieved in a NCI cited author search is more than offset by a vast improvement in comprehensiveness. 6. Percentage of clinical applications using each theory. A breakdown showing how many times each theorist was cited in the thirty-two "applications" papers appears in Figure 4. With eighteen papers citing it, Orem's self care deficit theory was by far the most widely used framework. SUMMARY AND SUGGESTIONS FOR FUTURE RESEARCH Table 1 Comparison of precision and comprehensiveness for cited author search and broad and narrow (textword) conventional searches: application studies These results indicate that NCI is unique in the level of access it provides to clinical applications testing nursing theory. It identifies studies which were previously inaccessible either by textword approaches or by the limited cited author access to nursing literature provided by SCI and SSCI. The retrieval on MEDLINE suggests that a traditional subject/textword search for this type of appli- Relevant and Comprehenunique to siveness Total retrieval Relevant Not relevant source Precision (n = 32) Broad MEDLINE search % 41% Narrow MEDLINE search % 28% NCI cited author search % 84% 181

7 Johnson cation paper is at best an adjunct to a cited author search, even when precision is more important than comprehensiveness. However, before discarding the subject/textword approach entirely in this context, librarians will want to monitor a recent enhancement to the NAHL thesaurus. Beginning in 1988, the NAHL producers formalized the indexing of nursing models by adding descriptors for each. It remains to be seen how these headings will be used: will they be attached to theory application papers as well as to simple descriptions and critiques of the models? A recent study of depth of indexing in NAHL revealed that on the average, NAHL indexers assign only five descriptors per document, including checktags and geographics [32]. Is the depth of indexing in NAHL sufficient to permit assigning a descriptor for the model applied once descriptors have been assigned to cover the subject of the study? An exploration of now NAHL performs versus NCI once these descriptors are in place would answer these questions. Since this study was made to gauge the utility of NCI, it was by necessity restricted to applications studies published in the journal literature. Dissertation Abstracts International, with its long and informative abstracts, might warrant a closer look as an aid to locating this type of nursing research. Finally, there is much potential for investigating the utility of NCI for locating other types of elusive literature; for example, the use of various psychological tests in nursing research. For the time being, though, it can be concluded that NCI is providing a much-needed link between nursing theory and the research it generates. REFERENCES 1. OREM DE, TAYLOR SE. Orem's general theory of nursing. In: Winstead-Fry P, ed. Case studies in nursing theory. New York: National League for Nursing, 1986:58. (NLN Publication ) 2. DAHLEN R, BRUNDAGE CA, ROTH CB. Guidelines for doing a computerized literature search. Nurse Educ 1987 Mar/Apr;12(2): SILVA MC. Conceptual models of nursing. Annu Rev Nurs Res 1987;5:230, BECK CT. Theoretical frameworks cited in Nursing Research from January 1974-June Nurse Educ 1986 Nov-Dec;10(6): SILVA MC. Research testing nursing theory: state of the art. ANS 1986 Oct;9(1): SMITH LC. Citation analysis. Libr Trends 1981 Summer;30(1): GARFIELD E. Can citation indexing be automated? In: Stevens ME, Giuliano VE, Heilprin LB, eds. Statistical association methods for mechanized documentation. Washington DC: National Bureau of Standards, 1965:189. (NBS Misc. Pub 269) 8. GARFIELD E. Citation indexing: its theory and application in science, technology, and humanities. New York: Wiley, 1979:2. 9. SALTON G. Associative document retrieval techniques using bibliographic information. J Assoc Comput Machinery 1963 Oct;10: MORAVCSIK MJ, MURUGESAN P. Some results on the function and quality of citations. Soc Stud Sci 1975 Feb;5(1): SPIEGEL-RosING I. Science studies: bibliometric and content analysis. Soc Stud Sci 1977 Feb;7(1): GARFIELD E. Can citation indexing be automated? FAWCETT J. Analysis and evaluation of conceptual models of nursing. Philadelphia: Davis, 1984: LONG GE. Professional advancement: theory development through practice and research. Am Nephrol Nurs Assoc J 1985 Feb;12(1): RIEHL JP, Roy C. Nursing models in education, research, and service. In: Riehl JP, Roy C, eds. Conceptual models for nursing practice. 2d ed. New York: Appleton-Century-Crofts, 1980: HARRELL JS. Needed: nurse engineers to link theory and practice. Nurs Outlook 1986 Jul-Aug;34(4): SANTORA D. Conceptual frameworks used in baccalaureate and master's degree curricula. New York: National League for Nursing, 1980:6. (NLN Publication No ) (League Exchange No. 126). 18. WINSTEAD-FRY P. Case studies in nursing theory, iv. 19. SILVA MC. Research testing nursing theory, Ibid. 21. NIGHTINGALE F. Notes on nursing: what it is, and what it is not. London: Harrison, NURSING DEVELOPMENT CONFERENCE GROUP. Concept formalization in nursing: process and product. Boston: Little, Brown, FAWCETT J. Analysis and evaluation of conceptual models of nursing. Philadelphia: Davis, RIEHL JP. Nursing models in current use. In: Riehl JP, Roy C, eds. Conceptual models for nursing prac-

8 Nursing Citation Index tice. 2d ed. New York: Appleton-Century-Crofts, 1980: CAPERS CF. Some basic facts about models, nursing conceptualizations, and nursing theories. J Contin Educ Nurs 1986 Sep-Oct;17(5): Beck CT. Theoretical frameworks cited in Nursing Research from January 1974-June Nurse Educ 1986 Nov-Dec;10(6): RIEHL JP, Roy C. Nursing models in education, research, and service, FAWCETT J. Analysis and evaluation of conceptual models of nursing. 29. Ibid. 30. RIEHL JP, Roy C. Conceptual models for nursing practice. 31. SILVA MC. Conceptual models of nursing. Annu Rev Nurs Res 1987;5: BRENNER S. CINAHL and MEDLINE: A comparison of indexing practices (paper presented at Eightyseventh Annual Meeting, Medical Library Association, Portland, Oregon, May 1987). Received May 1988; accepted August 1988 APPENDIX 1 Cited references searched for each theorist Dorothy Johnson 1. JOHNSON DE. The nature of a science of nursing. Nurs Outlook 1959 Apr;7(4): JOHNSON DE. The significance of nursing care. Am J Nurs 1961 Nov;61(11): JOHNSON DE. One conceptual model of nursing (paper presented at Vanderbilt University, Nashville, Tennessee, April 1968). 4. JOHNSON DE. Application of theory in education and service. Chicago: Teach em Inc., Audiocassettes. Betty Neuman 5. NEUMAN B, YOUNG RJ. A model for teaching the total person approach to patient problems. Nurs Res 1972 May-Jun;21(3): NEUMAN B. The Betty Neuman health care systems model: a total patient approach to patient problems. In: Riehl JP, Roy C, eds. Conceptual models for nursing practice. New York: Appleton-Century Crofts, 1974: NEUMAN B. The Betty Neuman health care systems model: a total person approach to patient problems. In: Riehl JP, Roy C, eds. Conceptual models for nursing practice. 2d ed. New York: Appleton-Century- Crofts, 1980: NEUMAN B. The Neuman systems model: application to nursing education and practice. New York: Appleton-Century-Crofts, Dorothea Orem 9. OREM DE. Guides for developing curricula for the education of practical nurses. Washington DC: Government Printing Office, OREM DE. Nursing: concepts of practice. New York: McGraw-Hill, OREM DE. Nursing: concepts of practice. 2d ed. New York: McGraw Hill, Martha Rogers 12. ROGERS ME. An introduction to the theoretical basis of nursing. Philadelphia: Davis, ROGERS ME. Nursing: a science of unitary man. In: Riehl JP, Roy C, eds. Conceptual models for nursing practice. 2d ed. New York: Appleton-Century- Crofts, 1980: Callista Roy 14. Roy C. Adaptation: a conceptual framework basis for nursing. Nurs Outlook 1970 Mar;18(3): Roy C. Adaptation: a basis for nursing practice. Nurs Outlook 1971 Apr;19(4): Roy C. The Roy adaptation model. In: Riehl JP, Roy C, eds. Conceptual models for nursing practice. New York: Appleton-Century-Crofts, 1974: Roy C. Introduction to nursing: an adaptation model. Englewood Cliffs, NJ: Prentice-Hall, Roy C. The Roy adaptation model. In: Riehl JP, Roy C, eds. Conceptual models for nursing practice. 2d ed. New York: Appleton-Century-Crofts, 1980: Roy C, ROBERTS SL. Theory construction in nursing: an adaptation model. Englewood Cliffs, NJ: Prentice-Hall, Roy C, ED. Introduction to nursing: an adaptation model. 2d ed. Englewood Cliffs, NJ: Prentice-Hall, APPENDIX 2 Broad MEDLINE Search Strategy 1-: PHILOSOPHY-NURSING OR NURS- ING-THEORY 183

9 Johnson 2 : MODELS-THEORETICAL OR MODELS- PSYCHOLOGICAL OR MODEL$1.TI,AB. OR FRAMEWORK$1.TI,AB. OR THEOR$.TI,AB. 3-: N.LI. OR NURS$3.TI,AB,DE. 4 : (NEUMAN$1 OR OREM$1 OR JOHN- SON$1 OR ROGERS$1 OR ROY$1).TI,AB. 5-: 4AND (3 OR 1) 6-: 1 OR (2 AND 3) 7-: SELF ADJ CARE AND (DEFICIT$1 OR AGENCY OR OREM$1.TI,AB. OR 6) 8-: LIFE ADJ PROCESS AND (ROGER$1.TI,AB. OR 6) 9-: (BEHAVIORAL OR BEHAVIOURAL) ADJ SYSTEM$1 10-: 9 AND (JOHNSON$1.TI,AB. OR 6) 1 L: ADAPTATION AND (ROY$1.TI,AB. OR 6) 12-: (SCIENCE WITH UNITARY WITH (MAN OR BEINGS)).TI,AB. 13-: 5 OR 7 OR 8 OR 10 OR 11 OR : 13 AND LGEN 15:..L/ 14 IM> : L/15 IM<8611 FROM THE BULLETIN-25 YEARS AGO Mechanization of Library Procedures in the Medium-sized Medical Library: II. Records By Irwin H. Pizer, Associate Librarian and Research Associate in Machine Methods; Isabelle T. Anderson, Assistant Librarian for Technical Services; and Estelle Brodman, Ph.D., Librarian and Associate Professor of Medical History, Washington University School of Medicine Library, St. Louis, Missouri In deciding to do the work on a daily basis in order to get a better end result, we are reminded of the analogy of the peas and the potatoes. As any child at the conundrum age knows, it is possible to place a peck of peas in a bushel basket of potatoes without requiring a larger container, because the peas settle in the interstices between the potatoes. Similarly, a certain amount of extra work can be absorbed by a staff without dislocation, if it can be interspersed among other duties. Daily keypunching is an example of this; keypunching the entire file at one time would not be so. Bull Med Libr Assoc 1964 Apr;52(2):