[Spanish version of the Memory Impairment Screen (MIS): normative data and discriminant validity]

Present data on discriminative and normative validity for the Memory Impairment Screen (MIS) in a cross-sectional validation study within a dementia-clinic setting using a Spanish adaptation of the MIS. Assess its usefulness as a screening instrument for memory problems related to primary dementing disorders, foremost Alzheimer's Disease (AD).

403 subjects, including 188 demented subjects according to DSM-IV criteria (119 of which presented AD according to NINCDS-ADRDA criteria) and 215 controls, over 50 years. The MIS, a four word memory test using a specific encoding technique and scoring on a 0-8 range, formed part of the neuropsychological workup, but it was not used for diagnosis. Sensitivity, specificity and predictive values positive (PPV) and negative (NPV) were calculated for the MIS against the gold-standard of clinical diagnosis by two blinded clinicians. ROC curves for the discrimination between demented subjects and controls, and AD subjects and controls were plotted.

Control subjects were significantly younger and better educated than the demented samples (p<0.001). No sex differences could be established. ROC curves demonstrated excellent discriminative validity of the MIS 402 for dementia on a whole (0.944) and even better for AD (0.978). The MIS presents satisfying results regarding sensitivity and specificity as well as PPV for different base rates, but the most effective cut-off score lies at < or = 4 points. For all demented subjects this cut-off represents a sensitivity of 0.74 and specificity of 0.96; for AD results are 0.86 and 0.96. More importantly, PPV values are above 0.65 for assumed base rates of 10% and more. Negative predictive values are almost 1 for all assumed base rates.

The Spanish version of the MIS reliably differentiates between normal and demented subjects. Nevertheless one has to consider the setting of the study when interpreting the results. The use of the S-MIS seems to be justified as first level dementia screening tool.