Objective. To evaluate the efficiency of the CAMDEX by exploring the contribution of its four main screening measures to the diagnosis of dementia at a memory clinic. Methods. A retrospective descriptive study was conducted on the diagnostic practice in a memory clinic regarding all consecutively referred patients who came for a first assessment. A standardised examination based on the CAMDEX included three cognitive tests (CAMCOG, MMSE, IMCT) and a test for functional competence (BDS). The predictive value of the tests was estimated by uni- and multivariate analysis with the clinical dementia diagnosis (yes/no) as dependent variable, taking into account the patients' age, education, gender and sensory ability. Results. Of the 180 patients consecutively referred, 150 completed the assessments and entered the study. Multivariate analysis revealed that the CAMCOG-score contributed most to the clinical diagnosis and was consistent with 84% of the clinical diagnoses using the conventional cutoff 79/80. The CAMCOG-score correlated best with the clinical diagnosis, however, at a cutoff score of 81/82. Forty-one patients (27%) scored closely around (± 5 points) the CAMCOG cutoff of 79/80. In this group 19 of the total of 23 deviations from the CAMCOG cutoff (83%) occurred. The patients' age, education, gender and visual ability explained some cases where the team's diagnosis deviated from the conventional cutoff score. Conclusion. Among four screening measures, after control, the CAMCOG was the only significant predictor for the clinical diagnosis of dementia. To gain efficiency, the screening measures of the CAMDEX protocol may be restricted to the CAMCOG. The interpretation of CAMCOG-scores around the cut-off is problematic. This indicates the need for reference values.
Hout et al. (Mon,) studied this question.