Generating Scientific Models of Knowledge Using arcs©

Background: Systematic approaches are needed to review literature on nutrition in heart failure for its scientific merit, relevance, and usefulness and identify directions for future research. Objectives: To evaluate the feasibility of arcs© (J.R.G., Indianapolis, Indiana), a computer program for managing data from literature and modeling knowledge, the objectives were to conduct an integrative review of 10 studies of nutrition in heart failure and generate scientific models of knowledge using arcs©. Methods: A unit of knowledge in arcs© is 2 variables linked by a statistical relationship. The computer program arcs© categorized variables and relationships found in the 10 explanatory observational studies. It also provided a scientific model for further empirical testing. Results: The computer program arcs© aggregated the following: 104 dependent and 93 independent operational variables and 60 associational, 16 predictive, 15 structural, 1 descriptive, and 85 difference relationships. A direct model produced by arcs© postulated a structural relationship between cachexia and 18-month mortality, independent of age or New York Heart Association classification, which can be tested as a path theoretical model. Conclusion: The computer program arcs© appeared to be feasible for conducting an integrative review of nutrition in heart failure. A larger, representative set of literature will enable generation of knowledge and identification of gaps and inconsistencies in findings.