A new shortened version of the Boston Naming Test for those aged over 65: an approach from item response theory
*Correspondencia: Dr. Miguel Ángel Fernández Blázquez. Centro de Prevención del Deterioro Cognitivo. Montesa, 22. Edif. B. E-28006 Madrid.
Fax: +34 915 886 793.
E-mail: ma.fdez.blazquez@gmail.com
Introduction: One of the tests that is mostly widely used to evaluate aphasia in clinical practice is the Boston Naming Test (BNT), a classic test in which 60 black and white pictures are presented to subjects in order to evaluate their capacity to put a name to such pictures. Despite its psychometric goodness, the number of items in the test has to be reduced in order to lower the time required to apply it.
Subjects and methods: Researchers recruited a sample of 547 subjects over the age of 65, who were then administered a neuropsychological evaluation protocol, including the BNT, to determine their cognitive statuses. No relevant cognitive alterations were observed in 405 subjects versus 142 who were diagnosed with mild cognitive impairment.
Results: The number of items was reduced in accordance with the premises of the item response theory. Since both age and level of schooling were found to have a significant effect on performance in the test, the two variables were used to find the cut-off points of the shortened version. This new version presented an adequate degree of reliability (alpha = 0.765) and a high correlation with the original test (r = 0.876).
Conclusions: The new shortened version consists of 15 items that are ordered according to the degree of difficulty. It is a task with a high level of discriminating power that is useful in day-to-day clinical practice for detecting alterations in the language of the elderly.
Sujetos y métodos Se reclutó una muestra de 547 sujetos mayores de 65 años a los que se les aplicó un protocolo de evaluación neuropsicológica, que incluía el BNT, para determinar sus estados cognitivos; 405 sujetos no presentaron alteraciones cognitivas relevantes, frente a 142 sujetos que fueron diagnosticados de deterioro cognitivo leve.
Resultados La reducción del número de ítems se realizó de acuerdo con los supuestos de la teoría de respuesta al ítem. Puesto que tanto la edad como el nivel educativo mostraron un efecto significativo en el rendimiento de la prueba, ambas variables fueron utilizadas para hallar los puntos de corte de la nueva versión reducida. Ésta presentó una adecuada fiabilidad (alfa = 0,765) y una elevada correlación con la prueba original (r = 0,876).
Conclusiones La nueva versión reducida consta de 15 ítems ordenados en función de su dificultad. Se trata de una tarea con un elevado poder discriminativo de utilidad en la clínica diaria para la detección de alteraciones del lenguaje en personas mayores.