Essential differences between evidence-based nursing and evidence-based medicine
Lavin, Mary Ann
Viewpoint provides a forum for scholarly discussion about the theoretical, philosophical, and practical issues related to nursing language and classification systems. The sometimes controversial views are intended to stimulate readers’ comments and perspectives. We welcome brief, informal responses to published viewpoints.
This Viewpoint is an expanded version of the introduction to “Online, Peer-Reviewed Bibliographic References on Sleep,” a presentation at the NNN conference “Developing, Linking, and Integrating Nursing Language and Informatics: Working Together for Quality Care,” April 10-13, Chicago.
In 1998, the NANDA Board of Directors conceived the idea of constructing a Web site to link nurses and nursing organizations involved in nursing language development and research. At its 2000 conference in Orlando, FL, the membership approved the idea. Between 2000 and 2002, contributors, using a small grant from St. Louis University’s Beaumont Faculty Development Fund, underwrote the development of the Network for Language in Nursing Knowledge Systems (NLINKS, www. nlinks.org).
NLINKS consists of a home page, an About NLUINKS page, a Concept Analysis Center, a Diagnostic Review Center, a Research Center, and a membership and donations page. For its first project, Research Center investigators decided to develop an online, peer-reviewed annotated bibliographic process. This was seen as a logical first step toward developing evidence-based systematic reviews analogous to those produced by the Cochrane Center.
Literature Review
The need for a high-quality, peer-reviewed annotated bibliographic process was determined by a literature review. As of March 2001, only five annotated bibliographies dealing with nursing diagnoses or potential diagnostic fields had been published in the preceding decade. Topics included spirituality research (Silva, 1999), fluid volume overload (Reed, 1996), pressure ulcers (Barnhart & Jarvis, 1993), anesthetic drug interactions (McAuliffe & Hartshorn, 1993), and family needs (Alpen & Halm, 1992).
While methods used in conducting literature reviews are well specified (Fink, Hays, Moore, & Beck, 1996; Hatsukami & Fischman, 1996; Yano, Fink, Hirsch, Robbins, & Rubenstein, 1995), annotated bibliographies often lack a systematic development and review process. Using the text words “annotated bibliography,” 153 citations were retrieved from a 1990-2000 Medline search. Of these, 104 were from medical journals, 19 from nursing journals, and 30 from journals in other healthcare professions. A CINAHL search for the same period and search term revealed 85 articles, with 18 representing medicine, 24 representing nursing, and 43 other healthcare professions. Six online full-text articles were found in both databases. These were selected as a sample to evaluate the quality of annotated bibliographic methods.
Of these six, two indicated the time period from which the articles were selected and the methods of organizing the topical headings (Dunn et al., 1994; Stevens & Thomas, 1996). Of these two, the methods employed by Stevens and Thomas were more comprehensive. They presented their annotated bibliography only after providing an overview of the topic (unintentional injury in the older adult), an explanation of the meaning of injury epidemiology, and a brief description of an appropriate model for studying the epidemiology of injury. The authors categorized subsequent references according to health education and epidemiologic concepts. Geber and Latts (1993) did not define selection criteria for the references chosen, nor did they categorize the references in a comprehensible manner. Other full-text annotated bibliographies presented both books and journal articles, but systematic approaches were lacking (Brown, 1999; Caputo & Gross, 1993; Mahon, 1997). This brief search indicated the need for a systematic method of selecting references for and evaluating the quality of annotated bibliographies.
Methods
Without predetermined rules to guide the construction of an online, peer-reviewed annotated bibliography, the authors chose an inductive approach to generate rules. The team formulated the initial steps in this process and reflected on the outcomes of any one step before proceeding to the next (Table 1). Tables 2 and 3 present the EvidenceBased Nursing (EBN) scale and the Contribution to Nursing Scale (CNS) developed during this inductive process.
Four St. Louis University investigators were each assigned to cover the time periods divided into the following categories: 1960s, 1970s, 1980s, and 1990 to present. Within categories, investigators were to select a convenience sample of three to five articles each for an initial review. Instructions were simple: Select from the time period to which you are assigned a minimum of three research articles, apply the scales, and write an annotated reference. From the 1,056 citations retrieved from a librarian-assisted Medline search,a,b 17 were selected for preliminary review.
As a result of this exercise, the expectation of actually writing and annotating was eliminated for two reasons. The work was too time-consuming even for motivated investigators, and no standards existed for the writing of free text annotated bibliographic. The scales did prove useful. To establish and test the functionality of posting, three were entered online (http://www.nlinks.org/ research_reviews.phtml).
Before sending the articles to reviewers at other universities for the purpose of evaluating interrater agreement, the St. Louis University team decided to evaluate interrater agreement among themselves. Two articles were selected and distributed to the five nursing faculty investigators. Each investigator independently applied the scales to the articles. Results indicated interrater agreement was greater than 80% on each of the 17 CNS items scaled, but was less than 40% on each of the 15 EBN scale items. These initial findings demonstrated that the investigators could independently agree on the evaluation of the articles in terms of the article’s contribution to nursing, but could not agree on the design used. An explanation was sought.
On reflection, the team decided the language used in the EBN scale was not applied by the investigator– reviewers in a standard fashion. For example, one reviewer might categorize one study as qualitative, whereas another reviewer might call it qualitative and quantitative, if any quantitative comparisons were made. In other words, the research language the team members used was more varied and not as standardized as the research language used by medicine.
The research team considered the possibility that better operationalization of terms might lead to better results. This thinking assumed there are agreed-on methods of categorizing nursing evidence and that the terms require only more explicit definition. On the other hand, the investigators entertained the notion that the language difficulties they themselves encountered with the EBN scale reflected the fact that they were trying to categorize nursing evidence within a model adapted from medicine’s evidence base. If so, perhaps there are basically different ways of categorizing the scientific evidence of the two professions. Perhaps the relatively small number of research design methods used by medicine represents a subset of a larger, as yet not fully identified, field of evidence used by nursing. Further, a larger field may be necessary to categorize the scientific basis not just of nursing but also of other healthcare professions. This discussion led to the formulation of the following viewpoint.
Viewpoint
While there are similarities, there are also essential differences between evidence-based medicine and evidence-based nursing. Differences include the focus of the research, the research methods used, and the subject matter (Table 4). First, consider the focus of nursing and medical research.
Focus of the research. Nursing focuses on the patient or client, and medicine focuses on treatment. This may seem just a semantic difference, at best, yet it reflects a philosophical subtlety that is real. Physicians focus their professional energy on preventing or treating disease. Nurses focus their professional energy on providing vigilance, reducing risks, anticipating patient needs, promoting client health, maintaining life processes, extending comfort, and at each and every encounter collecting and interpreting patient data. Of course, just as physicians do not treat or prevent disease outside the context of the patient, nurses do not care for patients outside the context of health or illness. Still, given the patient or client-centered focus of nursing and the treatment focus of medicine, it is logical to assume that the evidence that underlies each will reflect these distinctions.
Florence Nightingale, known for the biostatistical evidence she collected to support the care she recommended, was the first to make clear the distinction between medicine and nursing. “Medicine,” she said, “is the surgery of functions, as surgery proper is that of limbs and organs” (1859/1946, p. 74). In other words, Nightingale observed that medicine, by its nature, dissects functions much as a surgeon dissects limbs. She continues, “What [nursing does] … is to put the patient in the best condition possible for nature to act” (p. 75). We contend that while there is overlap in the evidence base of the two professions, there is enough distinctiveness to necessitate different evidence-based categories for nursing.
Research methods. Nursing’s research methods are more varied than medicine’s. It is logical to assume that if the two professions are philosophically distinct, there are going to be differences in how each defines its scientific levels of evidence. For example, medicine’s highest level of evidence categorizes its research methods on the basis of how subjects are selected and assigned to groups (University Library, 2002). These research methods indude the double-blind, randomized controlled trial; the randomized controlled trial; the controlled trial; and the cohort, case control, and case series. Except for the double-blind, randomized controlled trial, these research methods are appropriate for nursing as well. They allow nursing and medical scientists to study objectively treatment effects, incidence and prevalence of specific diagnoses, prognoses, and etiologic factors.
At the same time, nurse researchers use a greater variety of research methods. Therefore, nursing’s highest level of scientific evidence needs to be broad enough to include the studies medicine includes, plus the other kinds of research studies nursing conducts. The evidence-based medicine filters that are currently applied to the literature automatically reject many types of nursing research studies. These filters(c) accept articles that use medicine’s study designs (e.g., double-blind randomized controlled trials, randomized controlled trials, cohort studies, case-control studies, case series), but they reject articles using designs also used in nursing (e.g., qualitative research designs). To apply medicine’s scientific study language and its filters to capture nursing’s scientific studies artificially narrows nursing’s evidence base.
The University of Rochester Medical Center has addressed this issue by creating “Evidence-Based Filters for Ovid CINAHL,” posted on the Web page of the Edward G. Miner Library (http://www.urmc.rochester. edu/Miner/Educ/ebnfilt.htm). It posts filters for CINAHL diagnosis, prognosis, etiology (harm), therapy, meta-analysis, and qualitative studies. Although the work represents a step forward and supports the idea that nursing-sensitive filters are needed, the developers have not listed a nursing-sensitive filter for quantitative research methods. One suggestion might be to use the OVID-accessed (2002-2002) CINAHL subject heading of study design and its subheadings (e.g., crossover design, experimental studies, quasiexperimental studies) (http://gateway2.ovid.com/ovidweb. cgi). Using the definitions supplied by OVID, interobserver agreement among nursing experts on the manual assignment of articles to such categories could be tested.
Another way to address the breadth of nursing’s highest level of scientific evidence is to categorize all studies that collect data at the point of contact with the patient or client, call this level the primary data level, and devise a filter to sift the literature accordingly. This approach, as a first step, provides direct and comprehensive access to the human subject studies scholars need to evaluate the evidence base of nursing practice, conduct systematic reviews, and speak more directly to the full scope of nursing science.
Subject matter. Nursing’s subject matter is different from medicine’s. Medicine not only looks at how subjects are selected and assigned to groups, but it also categorizes these studies according to whether they apply to therapy, diagnosis, prognosis, etiology, prevention, or quality improvement (University Library, 2002). On the surface, these categories would seem to be appropriate for nursing, as we use the same terms. However, medicine defines diagnosis in terms of disease, whereas nursing defines diagnosis in terms of human responses to health and illness. Medicine defines therapy in terms of treatment interventions, most frequently drugs. Nursing defines therapy in terms of nursing treatments or interventions. Medicine defines etiology in terms of disease etiology. Nursing defines etiology in terms of factors related to human responses to health or illness. Here, too, the differences in the language of nursing and medicine mean the filters medicine uses to capture its scientific literature are not the same as those nursing would use to capture its literature. Resolution lies in nursing defining its own evidence-based nursing categories and filters, keeping in mind Florence Nightingale’s (1859/1946) admonition not to merely do all that medicine and other professions do but to do the best nursing can do.
a A CINAHL search was also conducted, but because of copyright issues, only the PUBMED search was used and documented.
b Name: sleep search
Date and Time search last updated: 11-Mar-2002 19:34:23 Database: PubMed
Search: (((((“sleep”[MeSH Terms] OR sleep[Text Word]) OR (“sleep disorders”[MeSH Terms] OR sleep disorders[Text Word])) AND (MM”nursing”[Subheading] OR “nursing”[MeSH Terms]) OR nursing[Text Word]) OR (“nursing process”[MeSH Terms] OR nursing pro cess[Text Word])) OR (((“nursing”[Subheading] OR “nursing”[MeSH Terms]) OR “nursing care”[MeSH Terms]) OR nursing care[Text Word])) OR (“nurse-patient relations”[MeSH Terms] OR nurse-patien relations[Text Word])) OR (“nurses”[MeSH Terms] OR nurses[Text Word])) OR (“nursing staff”[MeSH Terms] OR nursing staff[Text Word]))) AND English[Lang]) AND “human”[MeSH Terms]) Limits: ignored
c A filter is a sieve. Think of a coffee filter that allows the essence of the coffee to pass through the filter while retaining the coffee grounds. Relevant to this article, a filter is the searcher’s tool that separates or divides the literature, accepting some articles and rejecting others. Professionals, e.g., librarians working in conjunction with nurses or physicians, specify what these filters are.
References
Alpen, M.A., & Halm, M.A. (1992). Family needs: An annotated bibliography. Critical Care Nurse, 12(2), 32,41-50.
Barnhart, E.V, & Jarvis V (1993). Pressure ulcers: An annotated bibliography. Caet Journal, 12(4),11-22.
Brown, C.L. (1999). Ethics and health policy-annotated bibliography. Image, 31,394.
Caputo, G.M., & Gross, R.J. (1993). Medical consultation on surgical services: An annotated bibliography Annals of Internal Medicine, 118, 290-297.
Dunn, P.M., Gallagher, T.H., Godges, M.O., Prendergast, T.J., Rubenfeld, G.D., Tolle, S.W., & Lo, B. (1994). Medical ethics: An annotated bibliography. Annals of Internal Medicine, 121, 627-632.
Fink, A., Hays, R.D., Moore, A., & Beck, J.C. (1996). Alcohol-related problems in older persons: Determinants, consequences, and screening. Archives of Internal Medicine, 156, 1150-1156.
Geber, G., & Laos, E. (1993). Raising children with disabling conditions in a culturally diverse world: Race and ethnicity -Issues for adolescents with chronic illnesses and disabilities. An annotated bibliography. Pediatrics, 91(5), S1071 -$1081.
Hatsukami, D.IC, & Fischman, iW. (1996). Crack cocaine and cocaine hydrochloride: Are the differences myth or reality? JAMA, 276,1580 -1588. Last, J.M., & Abramson, J.H. (Eds.).(1995). A dictionary of epidemiology. New York: Oxford University Press.
Mahon, PY. (1997). Transcultural nursing: A source guide. Journal of Nursing Staff Development,13, 13,218-222.
McAuliffe, M.S., & Hartshorn, E.A. (1993). Anesthetic drug interactions. CRNA, 4, 207-212.
NANDA. (2001). Nursing Diagnosis: Definitions and Classification 2001-2002. Philadelphia: Author.
Nightingale, R (1859). Notes on nursing: What it is, and what it is not. London: Harrison (Reproduced by offset, Philadelphia, 1946).
Ovid Technologies (2000-2002). Version: re15.0.0, SourceID 1.6100.1.111, CINAHL Database. Retrieved July 11, 2002, from http://gateway2. ovid.com/ovidweb.cgi.
Reed, C. (1996). Care of postoperative patients with pulmonary edema. Journal of Perianesthesia Nursing, 11,164-169.
Silva, M.C. (1999). Spirituality and prayer research: A select annotated bibliography. Plastic Surgical Nursing, 19(1), 45-46.
Stevens, J.A., & Thomas, T.A. (1996). Major causes of unintentional injuries among older persons: An annotated bibliography. Atlanta: USPHS Centers for Disease Control and Prevention.
University Library, University of Illinois at Chicago (updated March 26, 2002). Levels of clinical evidence in primary literature, from Evidence Based Medicine: Is All Evidence Created Equal? Retrieved April 20, 2002, from http://www.uic.edu/depts/lib/lhsp/resources/ levels.shtml.
Yano, E M., Fink, A., Hirsch, S.H., Robbins, A., & Rubenstein, L.V. (1995). Helping practices reach primary care goals: Lessons from the literature. Archives of Internal Medicine, 155,1146-1156.
Mary Ann Lavin, ScD, RN, FAAN, Geralyn Meyer, PhD, RN, Mary Krieger, MLIS, RN, Patricia McNary, MALS, RN, Judith Carlson, MSN, RN, Anne Perry, EdD, RN, FAAN, Dorothy James, PhD, RN, and Tome Civitan
Mary Ann Lavin, ScD, RN, FAAN
Associate Professor
Geralyn Meyer, PhD, RN
Assistant Professor
Mary Krieger, MLIS, RN
Librarian
Anne Perry, EdD, RN, FAAN
Professor
Dorothy James, PhD, RN
Assistant Professor
Tome Civitan
Research Associate
Judith Carlson, MSN, RN
Associate Professor
Saint Louis University School of Nursing
St. Louis, MO
and Patricia McNary, MALS, RN
Assistant Professor
Massachusetts College of Pharmacy & Health Sciences
Boston, MA
Author contact: lavinma@edu, with a copy to the Editor: rose_mary@earthlink.net
Copyright Nursecom, Inc. Jul-Sep 2002
Provided by ProQuest Information and Learning Company. All rights Reserved