Overcoming redundancies in bedside nursing assessments by validating a parsimonious meta-tool: findings from a methodological exercise study

Autor: Federica Canzan, Paola Zambiasi, Elisabetta Allegrini, Letizia Bazoli, Paolo Chiari, Annamaria Guarnier, Michele Picogna, Alvisa Palese, Luca Grassetti, Marisa Padovan, Paola Casson, Paolo Barelli, Eva Marini, Meri Marin, Luisa Saiani, Elisa Ambrosi, Daniele Salmaso, Oliva Marognolli, Patrizia Taddia
Rok vydání: 2016
Předmět:
Zdroj: Journal of Evaluation in Clinical Practice. 22:771-780
ISSN: 1356-1294
DOI: 10.1111/jep.12539
Popis: Rationale There is growing interest in validating tools aimed at supporting the clinical decision-making process and research. However, an increased bureaucratization of clinical practice and redundancies in the measures collected have been reported by clinicians. Redundancies in clinical assessments affect negatively both patients and nurses. Methods To validate a meta-tool measuring the risks/problems currently estimated by multiple tools used in daily practice. A secondary analysis of a database was performed, using a cross-validation and a longitudinal study designs. In total, 1464 patients admitted to 12 medical units in 2012 were assessed at admission with the Brass, Barthel, Conley and Braden tools. Pertinent outcomes such as the occurrence of post-discharge need for resources and functional decline at discharge, as well as falls and pressure sores, were measured. Explorative factor analysis of each tool, inter-tool correlations and a conceptual evaluation of the redundant/similar items across tools were performed. Therefore, the validation of the meta-tool was performed through explorative factor analysis, confirmatory factor analysis and the structural equation model to establish the ability of the meta-tool to predict the outcomes estimated by the original tools. Results High correlations between the tools have emerged (from r 0.428 to 0.867) with a common variance from 18.3% to 75.1%. Through a conceptual evaluation and explorative factor analysis, the items were reduced from 42 to 20, and the three factors that emerged were confirmed by confirmatory factor analysis. According to the structural equation model results, two out of three emerged factors predicted the outcomes. Conclusions From the initial 42 items, the meta-tool is composed of 20 items capable of predicting the outcomes as with the original tools.
Databáze: OpenAIRE
Externí odkaz:
Nepřihlášeným uživatelům se plný text nezobrazuje