1. INTRODUCTION
In the context of the health information systems, two main categories of languages are described: Standardized vocabularies, like classifications and taxonomies and uncontrolled vocabularies, essentially the natural language [1,2].
Nursing controlled vocabularies generation started in the early 70s in order to describe nursing phenomena and since then, it has been associated to the theoretical development aiming to identify, define, and classify disciplinary concepts in order to improve nursing education, management and clinical practice [2].
Increasing pressure on healthcare systems to gain efficiency, quality and productivity is challenging nurses to demonstrate the impact of their professional services on health outcomes in individuals and communities, so standardized computer-compatible professional terminologies are becoming a requirement. In this sense, language systems development and evaluation have been included as priorities into the nursing international agendas [3-5].
Worldwide efforts in nursing terminological works are reflected in the American Nurses Association (ANA) recognition program for terminologies as supporting nursing practice, including several controlled vocabularies like the North American Nursing Diagnosis Association Taxonomy (NANDA-I), the Nursing Interventions Classification (NIC), the Clinical Care Classification (CCC), the Nursing Outcomes Classification (NOC) or the Omaha System among others [6]. In this same way, the International Council of Nurses (ICN) has been increasingly investing efforts in the development of a unified nursing language system, that is, the International Classification for Nursing Practice (ICNP) [7].
Institutions and researchers from different countries have joined this professional mandate designing, implementing and evaluating controlled vocabularies for nursing practice [4,8-12].
Despite the achievements, one of the reasons for the use of these terminologies in computer-based systems is not universal is that “in the clinical practice, nurses use terms other than those in standardized vocabularies” [3]. It has been suggested in the recent literature, that nurses refer standardized nursing languages reduce the “individualized approach” of nursing care and documentation, because these vocabularies are not able to reflect subtle changes in patient status and “they may foster inaccuracies in patients’ information in reporting clinical events” [13]. Similarly, the results of a survey conducted in the United States showed that most nurses had no experience with or knowledge of any nursing controlled vocabulary, being the NANDA framework “the most recognizable with over 1/3 of respondents reporting that they had used it in nursing school, but not since” [14].
This article focuses on a nursing interface terminology, termed ATIC. Interface terminologies are controlled vocabularies, based on close to natural language concepts, optimized for end-user data entry, aimed to ease a friendly use of the terminological system within the electronic health records (EHR) [15,16].
The ATIC terminology has been used for representing nursing phenomena in the electronic health records system implemented in 11 hospitals in Catalonia. This vocabulary is structured in three main axes: Assessment, diagnosis/outcomes and interventions. Details contained in the description of each concept are listed in Table 1.
Further information on the evolving status of the coverage and the general structure of this interface vocabulary, as well as its philosophical, theoretical and methodological background and other related studies are published elsewhere [17-23].
As any other clinical instrument, the use of a particular nursing terminology should be evaluated. Validity evaluation criteria for nursing controlled vocabularies described in the literature include, among others, that the terminology should be research-based and nursing phenomena-oriented [4].
The main goal of this study was to evaluate to what extent the diagnosis axis of the ATIC terminology is oriented to nursing phenomena, as a measure of its content validity.
2. METHOD
2.1. Design
This study applied an observational, descriptive design, using the technique of contrast with previous data, that is, a literature review strategy to identify the disciplinary scientific production on the diagnostic concepts included in the ATIC terminology.
2.2. Sample
The objects of the study were the concepts within the diagnosis axis of the ATIC terminology. Concepts under development or refinement at the time of starting the study were excluded. Sample size was calculated considering an 80% estimated proportion (P), with a 95% confidence level (α = 0.05) and a 0.05 precision (i). Sample size resulted in 246 objects of study. A correction formula was applied to the sample size (Na = N[1/(1–R)]) to keep statistical power in case of potential losses due to missed data or other reasons. Corrected sample size was estimated at 287 concepts to be included in the final analysis.
Concepts were randomly selected applying a random number list to the terms in the diagnosis axis of the terminology. Random numbers were obtained using the random generation function of Excel (Microsoft, Redmond, WA).
2.3. Data Collection
Nursing research papers related to the concepts randomly included in the study were searched in the following healthcare databases: Pubmed (http://www.ncbi.nlm.nih.gov/pubmed/), the Cochrane Library (http://www.cochrane.es/) and the Scientific Electronic Library Online (http://www.scielo.cl/).
Search limits were established in relation to language and time. Publication languages included were English, French, Portuguese, Catalan, Spanish and Italian. The review was performed considering a maximum of 20 years back in time. Redundancy was considered as an end point of the reviewing process before reaching the whole 20-year back period.
Quantitative and qualitative research designs were considered. Editorials, letters, news and historical articles were excluded. Located references in the databases with no abstract available were also excluded.
Research designs were classified into three main categories:
The first group included case studies, case series, reviews as well as concept analysis or concept development designs;
The second category included quantitative descriptive designs, validation studies and qualitative ethnographies, exploratory, grounded-theory and phenomenological studies;
The third category of studies included analytical designs, controlled trials, meta-analysis and qualitative metasynthesis.
Criteria were settled to systematically address the search for each concept: Keywords for the concept under study and its synonyms plus the word “nursing” (and/or midwifery if applicable) were used both for actual and risk diagnosis concepts. For risk diagnosis concepts, the search also included the keywords “prevention” (first search) and “risk factors” (second search) (Figure 1).
A short standardized data collection sheet was designed to document the research variables for each concept including:
1) Presence of the concept in the nursing literature2) Types of articles and3) Areas of disciplinary interest.
The first topic, the presence of the concepts in the disciplinary literature was defined as a dichotomyc variable: Presence of the concept, Yes or Not.
Types of designs were organized in three categories as previously described.
Areas of disciplinary interest were classified into five main domains: 1) Medical-Surgical Nursing, also including critical care and emergency nursing; 2) MaternalChild and Pediatric Nursing; 3) Family and Community Nursing; 4) Mental Health Nursing; and 5) other areas of disciplinary interest (including research papers on nursing ethics, politics, management, economics, education and theory development).
Geriatric nursing was considered in the Medical-Sur-
Table 1. Basic elements included in each concept construction (IA) = if applicable.
gical Nursing domain for elder hospital in-patients and in the Family and Community Nursing domain for elder outpatients, elder people living in the community and nursing home residents.
Home healthcare nursing was also considered in the Family and Community Nursing domain, except for hightech home healthcare nursing, which was included in the Medical-Surgical Nursing domain.
Data were systematically collected from July 2nd 2010 to January 31st 2012.
2.4. Data Analysis
Main outcome measures were the global score for the presence of the diagnostic concepts in the nursing literature and the distribution scores for the other two topics: Types of studies and areas of disciplinary interest.
Data were processed onto an Excel spreadsheet (Microsoft, Redmond, WA) and revised to identify potential processing errors or inconsistencies. Descriptive analysis of the main outcomes including frequencies in percentages, means and standard deviations when applicable were used. Confidence interval was calculated for a confidence level of 95%.
3. RESULTS
The final analysis included 287 concepts from the diagnosis axis of ATIC, 203 corresponding to actual nursing diagnosis concepts and 84 to risk nursing diagnosis (29.2%). The search process allowed the researcher to consider 7731 paper summaries. Mean number of references selected per concept was 25.5 (CI 2.09).
According to this analysis, the main results for the outcome measures indicate that 98.7% of the concepts included in the study were identified as “being present” in the nursing research literature. Four nursing diagnoses included in this analysis, did not match any nursing scientific production: Newborn physiological immaturity, abulia, situational emotional claudication, and risk for unintended self-exclusion.
Distribution results for the types of designs ranked analytical and meta-analytical studies at the top representing a 44.1%; followed by category 2 designs (quailtative and quantitative descriptive studies) obtaining a 35.6% and 19.7% for category 1 designs (case studies, case series and reviews). Sample diagnosis labels within each of these categories are listed in Table 2.
Rank distribution of concepts matching nursing research studies, organized by areas of disciplinary interest placed first Medical-Surgical Nursing (N = 248 concepts; 86.4%), followed by Maternal-Child and Pediatric Nursing (N = 232 concepts; 80.8%), Family and Community Nursing (N = 170 concepts; 59%), other areas of disciplinary interest (N = 131 concepts; 45.6%) and finally Mental Health Nursing (N = 108 concepts; 37.6%).
Nursing diagnosis concepts falling into two or more of these areas accounted for 87.7%. Figure 2 describes the basic distribution for nursing diagnosis concepts falling into one or more areas of disciplinary interest.
Table 3 shows a matrix with sample nursing diagnoses of the ATIC terminology falling into the different areas of disciplinary interest.