RESUMO
OBJECTIVES: To determine the validity and reliability of the personalised nursing care index (PNCI). METHODS: A descriptive study was carried out in the Puerta del Mar Hospital (Cádiz) of the Andalusian Health Service, which included all patients admitted to an Internal Medicine ward from May 2007 to July 2008. A descriptive analysis was performed on the population included in the study. Furthermore, reliability was analysed with the 20-Kuder-Richardson coefficient and a correlation matrix between PNCI items; and validity via a maximum likelihood factorial analysis, and a ROC curve to determine sensitivity and specificity. RESULTS: A total of 219 people were included in the survey, of whom 74.4% were patients and 26.6% caregivers. The percentages of males and females were similar with a mean age of 62.5 years and 48.4% finished basic school education. The 20-Kuder-Richarson value was 0.835; and a positive correlation between items, which corresponded to 3 groups of items (factors). The maximum likelihood factorial analysis confirmed the 3 items-factor groups with a Promax rotation due to the high correlation between them. The maximum explained variance was 91.42%. The ROC curve area was 90.1% with a cut-off point of 8, for a sensitivity of 79.9% and a specificity of 90.6%. CONCLUSIONS: The PCNI is reliable, with the internal consistency coefficient value in between an optimum range. Furthermore, factors obtained from the matrix correlation inter-items were confirmed with a factorial analysis, resulting in a high explained variance. The curve ROC area is excellent compared to the gold standard, considered as the perception of confidence relationship.
