Objective: To estimate the prevalence of cognitive impairment in elderly people living in the community. To assess the diagnostic yield of the main cognitive screening tests in Primary Care: the Mini Mental State Examination (MMSE), the Clock Drawing Test (CDT), the Verbal Fluency Test (VFT) and the Informant Questionnaire test (IQ). We also calculated their correlation and validity.
Design: Descriptive study based on a personal interview by professionals using a standardised method to assess cognitive impairment.
Participants: A total of 293 men and women, who were older than 64 years old and residing in the province of Huesca. The sample was recruited by systematic random sampling from the database of personal health care card.
Main measurements: We analysed social and demographic variables (age, educational level, marital status) and the assessment of cognitive impairment by means of MMSE, CDT, VF, IQ tests.
Results: The prevalence of cognitive impairment with MMSE was 18.1% (95% CI: 15.3-24.4%). The correlation between MMSE and the other cognitive screening tests was moderate: CDT kappa=0.40; VF kappa=0.33; IQ kappa=0.36. Using the MMSE as a gold standard and the prior diagnosis of CI, the sensitivity of CDT was 78.3% and the specificity was 76.9%, 80%; there was 70% sensitivity and 76.7% specificity for the VF; and 71.4% for IQ respectively. The area under the VF ROC curve was significantly higher than the other tests.
Conclusions: The agreement between cognitive impairment screening tests was moderate. The VF yielded better diagnostic accuracy.
Objetivo: Estimar la prevalencia de deterioro cognitivo (DC) en ancianos de la comunidad. Evaluar el rendimiento diagnóstico de los principales test de cribado cognitivo usados en atención primaria: miniexamen cognoscitivo (MEC), test del reloj (TR), test de fluencia verbal (TFV) y test del informador (TIN). Calcular la concordancia entre ellos y su validez como pruebas de cribado.
Diseño: Estudio descriptivo transversal.
Participantes: Doscientos noventa y tres personas de ambos sexos, mayores de 64 años, residentes en la provincia de Huesca, seleccionados mediante muestreo aleatorizado sistemático. Marco muestral: base de datos de Tarjeta Sanitaria Individual. Ámbito provincial.
Mediciones: Variables sociodemográficas básicas, y exploración del DC con MEC, TR, TFV y TIN.
Resultados: En función del MEC y los diagnósticos previos, la prevalencia de DC es del 18,1% (IC del 95%: del 15,3 al 24,4%).
La concordancia del MEC con el resto de los test de cribado es moderada: con el TR kappa (κ)=0,40, con el de TFV: 0,33, con el TIN: 0,36.
La concordancia global del conjunto de los test cognitivos es moderada-baja: κ (Jackknife)=0,35 (IC del 95%: 0,269 a 0,444).
Considerando patrón de oro al MEC junto con el diagnóstico previo de DC, el TR presenta una sensibilidad del 78,3% y una especificidad del 76,9%; el TFV presenta el 80 y el 70%, respectivamente, y el TIN presenta el 76,7 y el 71,4%, respectivamente. El área bajo la curva ROC (receiver operating characteristics) del TFV es significativamente superior al resto.
Conclusión: La concordancia entre los distintos test de cribado cognitivo es moderada. El TFV es el que mejor rendimiento diagnóstico presenta.
Copyright 2009 Elsevier España, S.L. All rights reserved.